We Tested Meta's Incremental Attribution on $200K/Month: Here's Why We Turned It Off.

1
00:00:00,800 --> 00:00:03,319
Because just due to the nature of how one
2
00:00:03,319 --> 00:00:05,240
day click and incremental work, kind of
3
00:00:05,240 --> 00:00:06,759
like we said earlier, one day click is
4
00:00:06,759 --> 00:00:07,559
only giving credit to
5
00:00:07,559 --> 00:00:09,080
people who click within 24 hours.
6
00:00:09,320 --> 00:00:11,080
By default, it is substantially more
7
00:00:11,599 --> 00:00:12,279
conservative than
8
00:00:12,279 --> 00:00:13,240
incremental attribution.
9
00:00:13,679 --> 00:00:14,759
So we don't want to be
10
00:00:14,759 --> 00:00:15,880
misled by one day click.
11
00:00:16,000 --> 00:00:16,960
What should we do?
12
00:00:17,320 --> 00:00:18,760
That third data point that I'm alluding
13
00:00:18,760 --> 00:00:20,359
to is split it down to
14
00:00:20,359 --> 00:00:21,519
evergreen and promo ads.
15
00:00:22,039 --> 00:00:23,320
It still ended up being the case that
16
00:00:23,359 --> 00:00:24,440
evergreen consistently is
17
00:00:24,440 --> 00:00:25,519
stacking up month over month.
18
00:00:25,519 --> 00:00:26,719
And that's what compounds, right?
19
00:00:26,719 --> 00:00:29,559
It's like if you can make more ads every
20
00:00:29,559 --> 00:00:31,039
single month that earns spend in the ad
21
00:00:31,039 --> 00:00:33,079
account, then that's what compounds over
22
00:00:33,079 --> 00:00:35,679
time and allows you to increase spend.
23
00:00:35,840 --> 00:00:36,600
I think a lot of brands
24
00:00:36,600 --> 00:00:37,960
actually kind of forget about that.
25
00:00:37,960 --> 00:00:39,640
They're like, hey, they didn't buy.
26
00:00:39,640 --> 00:00:42,359
We did all these like whatever flows to
27
00:00:42,439 --> 00:00:43,439
get them to maybe get a
28
00:00:43,439 --> 00:00:44,679
bigger discount over time.
29
00:00:44,759 --> 00:00:46,200
But then there was probably one reason
30
00:00:46,200 --> 00:00:47,039
why they didn't buy.
31
00:00:47,079 --> 00:00:48,920
So talk to that customer about that one
32
00:00:48,920 --> 00:00:49,960
reason why they didn't buy and give them
33
00:00:49,960 --> 00:00:51,719
a strong discount to still close.
34
00:00:51,719 --> 00:00:53,920
Because like it's very likely better off
35
00:00:53,920 --> 00:00:55,479
when you still have contribution margin
36
00:00:55,479 --> 00:00:57,479
on that customer to be gained if you just
37
00:00:57,479 --> 00:00:59,439
give them even a slightly bigger discount
38
00:00:59,479 --> 00:01:00,880
to like bring in those dollars versus
39
00:01:00,880 --> 00:01:01,560
them just sitting on
40
00:01:01,560 --> 00:01:02,600
an email and SMS list.
41
00:01:09,280 --> 00:01:11,239
And now let's take a listen to the
42
00:01:11,239 --> 00:01:12,840
Scalability School podcast.
43
00:01:14,560 --> 00:01:16,239
Welcome back to another episode of the
44
00:01:16,239 --> 00:01:17,879
Scalability School podcast.
45
00:01:17,920 --> 00:01:20,640
Andrew is allowing me to have the
46
00:01:20,680 --> 00:01:22,640
absolute honor of introducing the
47
00:01:22,640 --> 00:01:24,640
podcast, opening it up for reasons that
48
00:01:24,640 --> 00:01:25,560
he will describe shortly.
49
00:01:26,079 --> 00:01:28,200
He sounds beautiful, but he's going to
50
00:01:28,200 --> 00:01:29,599
let me open this sucker up.
51
00:01:30,200 --> 00:01:30,680
We have a couple of
52
00:01:30,680 --> 00:01:32,280
topics to jump into today.
53
00:01:32,400 --> 00:01:33,120
We're going to talk about
54
00:01:33,159 --> 00:01:34,120
incremental attribution.
55
00:01:34,159 --> 00:01:34,560
We're going to talk
56
00:01:34,560 --> 00:01:35,640
about creative cohorts.
57
00:01:35,799 --> 00:01:38,879
And if we are blessed and lucky, Zach
58
00:01:38,879 --> 00:01:39,760
Stuck will be joining us
59
00:01:39,760 --> 00:01:40,840
for the back half of this.
60
00:01:41,319 --> 00:01:41,920
We're going to talk
61
00:01:41,920 --> 00:01:43,000
about scaling ads, man.
62
00:01:43,480 --> 00:01:44,239
It's going to be dope.
63
00:01:44,239 --> 00:01:46,799
But look, as you can hear, I'll sing you
64
00:01:46,799 --> 00:01:48,079
a little with my husky voice.
65
00:01:48,599 --> 00:01:50,879
I'm in the mood for love.
66
00:01:50,920 --> 00:01:51,840
That's where I'm at
67
00:01:51,840 --> 00:01:53,959
right now with my bad voice.
68
00:01:54,560 --> 00:01:55,799
I've lost my voice, but
69
00:01:55,840 --> 00:01:56,879
it's not totally gone.
70
00:01:57,719 --> 00:01:58,920
I just sound like I'm struggling to
71
00:01:58,920 --> 00:01:59,840
breathe, which is true.
72
00:02:00,840 --> 00:02:03,280
So and my laughs sound like I've been
73
00:02:03,280 --> 00:02:04,519
smoking for 40 years.
74
00:02:04,920 --> 00:02:06,359
Other than that, doing
75
00:02:06,400 --> 00:02:07,680
great, feeling great.
76
00:02:07,840 --> 00:02:08,879
These are aura points.
77
00:02:08,879 --> 00:02:09,319
You're just stacking
78
00:02:09,319 --> 00:02:10,240
aura points right now.
79
00:02:10,960 --> 00:02:11,719
So much aura.
80
00:02:12,080 --> 00:02:13,120
Yeah, people are going to actually
81
00:02:13,120 --> 00:02:15,120
require me to do podcast appearances with
82
00:02:15,120 --> 00:02:16,879
this voice from now on.
83
00:02:17,039 --> 00:02:18,400
So I'll have to 11 labs
84
00:02:18,400 --> 00:02:19,960
it and figure that out.
85
00:02:20,400 --> 00:02:21,960
I think that, you know, one of the things
86
00:02:21,960 --> 00:02:23,400
we love is tactical episodes.
87
00:02:23,879 --> 00:02:25,039
And that's what we're
88
00:02:25,039 --> 00:02:26,199
trying to get into here, right?
89
00:02:26,199 --> 00:02:29,319
Which is around incremental attribution
90
00:02:29,319 --> 00:02:31,120
and versus one day click.
91
00:02:32,439 --> 00:02:33,919
I'm really excited to talk about it.
92
00:02:34,960 --> 00:02:37,360
So let's go through this.
93
00:02:37,360 --> 00:02:40,199
Like, I mean, we have some data from the
94
00:02:40,240 --> 00:02:41,159
founders community as
95
00:02:41,159 --> 00:02:42,120
well, which we'll talk about.
96
00:02:42,759 --> 00:02:44,080
But Brett, frame it up for us.
97
00:02:44,639 --> 00:02:44,960
Yeah.
98
00:02:45,000 --> 00:02:47,120
So we've been thinking a lot about
99
00:02:47,120 --> 00:02:48,159
incremental attribution
100
00:02:48,159 --> 00:02:49,400
since meta rolled it out.
101
00:02:50,080 --> 00:02:54,199
And I have an N of one kind of sample of
102
00:02:54,439 --> 00:02:57,520
a recent experience that we've had where
103
00:02:57,520 --> 00:02:58,960
we went through kind of the
104
00:02:59,680 --> 00:03:02,199
disappointment, excited, back to
105
00:03:02,199 --> 00:03:03,560
disappointment and trying to figure out
106
00:03:03,560 --> 00:03:05,240
where we go with incremental attribution.
107
00:03:05,439 --> 00:03:06,879
So if it's something that you think, if
108
00:03:07,080 --> 00:03:08,240
it's something that you've been thinking
109
00:03:08,240 --> 00:03:10,280
about testing, maybe you've already
110
00:03:10,280 --> 00:03:11,360
tested and you've had a similar
111
00:03:11,360 --> 00:03:13,000
experience, hopefully you can lean on
112
00:03:13,000 --> 00:03:14,639
this to kind of decide, is it worth
113
00:03:14,639 --> 00:03:15,560
testing in our account?
114
00:03:15,560 --> 00:03:17,240
And should we expect that there's going
115
00:03:17,240 --> 00:03:20,159
to be meaningful performance swings?
116
00:03:21,520 --> 00:03:22,759
Because, you know, everybody's always
117
00:03:22,759 --> 00:03:24,680
looking for, I mean, I'm kind of the guy
118
00:03:24,680 --> 00:03:26,080
that always asks everybody for a hack at
119
00:03:26,080 --> 00:03:27,280
the end of the episode because I like to
120
00:03:27,280 --> 00:03:28,240
hear the tactical nuggets.
121
00:03:29,280 --> 00:03:30,960
So hopefully you walk away with a kind of
122
00:03:30,960 --> 00:03:33,000
an understanding of what to do with
123
00:03:33,039 --> 00:03:34,879
incremental attribution from here.
124
00:03:35,520 --> 00:03:37,000
But I can walk through, I actually have a
125
00:03:37,000 --> 00:03:38,719
bunch of notes in screen shares and I can
126
00:03:38,719 --> 00:03:41,159
just walk through a timeline of a test
127
00:03:41,159 --> 00:03:42,039
that we ran through with a
128
00:03:42,039 --> 00:03:43,520
client and their experience.
129
00:03:43,840 --> 00:03:45,280
And then Andrew, you can always interrupt
130
00:03:45,280 --> 00:03:47,000
me and we can ask questions
131
00:03:47,000 --> 00:03:49,000
and dig into anything deeper.
132
00:03:49,039 --> 00:03:49,599
Does that sound good?
133
00:03:50,400 --> 00:03:50,960
Yeah, definitely.
134
00:03:51,120 --> 00:03:52,759
I mean, I think I'll just mention the
135
00:03:52,759 --> 00:03:54,919
community stuff out of the gates, which
136
00:03:54,919 --> 00:03:57,719
is people had seen it
137
00:03:57,759 --> 00:04:01,120
underperform thus far utilizing IA.
138
00:04:02,639 --> 00:04:03,000
Right.
139
00:04:03,000 --> 00:04:05,159
So it depends.
140
00:04:05,159 --> 00:04:09,680
Like one member said, he saw that it was
141
00:04:09,840 --> 00:04:11,599
especially on remarketing review through
142
00:04:11,599 --> 00:04:12,479
attribution on past
143
00:04:12,479 --> 00:04:14,039
purchases was inflating numbers.
144
00:04:14,400 --> 00:04:16,079
Sees IA aligned well
145
00:04:16,079 --> 00:04:17,040
with seven day click.
146
00:04:18,519 --> 00:04:21,199
Another member said that incremental ROAS
147
00:04:21,199 --> 00:04:24,279
and ads manager often lands around 90% of
148
00:04:24,279 --> 00:04:25,399
seven day click ROAS
149
00:04:25,399 --> 00:04:26,279
when you break it down.
150
00:04:27,720 --> 00:04:32,040
And Barry Hot legendary said he thinks
151
00:04:32,040 --> 00:04:35,040
it's good 90% of the problems he sees
152
00:04:35,319 --> 00:04:36,839
because it solves those problems, 90% of
153
00:04:36,839 --> 00:04:37,600
the problems he sees,
154
00:04:37,720 --> 00:04:39,000
especially on bigger accounts.
155
00:04:39,360 --> 00:04:41,040
So it's interesting.
156
00:04:41,079 --> 00:04:42,000
It's like, it's interesting.
157
00:04:42,000 --> 00:04:44,240
It's kind of like people said hasn't been
158
00:04:44,240 --> 00:04:45,839
as good, but then like overall we're
159
00:04:45,839 --> 00:04:47,000
like, I can see where it's happening.
160
00:04:47,680 --> 00:04:48,600
And it's going to be better.
161
00:04:48,839 --> 00:04:50,560
So yeah, I'd love to walk through it.
162
00:04:51,879 --> 00:04:52,240
Sweet.
163
00:04:52,279 --> 00:04:53,600
I will flip this over.
164
00:04:54,480 --> 00:04:56,160
You're watching this on youtube.com,
165
00:04:56,600 --> 00:04:57,519
which is a website on the
166
00:04:57,519 --> 00:04:59,160
internet that shows videos.
167
00:04:59,720 --> 00:05:01,839
Brad is sharing a screen, but if you're
168
00:05:01,839 --> 00:05:03,920
not, you know, and you're just having it
169
00:05:03,920 --> 00:05:07,839
in your ears, then, you know, I'm sorry,
170
00:05:07,839 --> 00:05:08,800
but you can go and look at
171
00:05:08,800 --> 00:05:10,800
the visuals on youtube.com.
172
00:05:11,000 --> 00:05:12,040
Somebody the other day
173
00:05:12,079 --> 00:05:13,480
I met at a conference.
174
00:05:13,560 --> 00:05:14,160
This is true.
175
00:05:14,680 --> 00:05:17,079
And said, who is the other guy when he
176
00:05:17,079 --> 00:05:20,040
met me on the, on the podcast.
177
00:05:20,160 --> 00:05:21,399
And he goes, I love
178
00:05:21,399 --> 00:05:22,439
his, I love his voice.
179
00:05:22,439 --> 00:05:24,360
I've listened to every episode twice.
180
00:05:25,680 --> 00:05:27,680
So some people just come for the audio.
181
00:05:28,319 --> 00:05:30,120
Hey, if this helps you fall asleep at
182
00:05:30,120 --> 00:05:32,279
night, that is an acceptable, is an
183
00:05:32,279 --> 00:05:33,519
acceptable listening method.
184
00:05:33,600 --> 00:05:34,519
Kind of an honor really.
185
00:05:35,040 --> 00:05:36,120
Yeah, we're just trying to, you know,
186
00:05:36,120 --> 00:05:36,639
we're just trying to
187
00:05:36,639 --> 00:05:38,399
smash records around here.
188
00:05:39,639 --> 00:05:40,720
So, okay.
189
00:05:41,000 --> 00:05:42,759
I will also try to explain what's on the
190
00:05:42,759 --> 00:05:43,839
screen for everybody
191
00:05:43,839 --> 00:05:44,800
that's, that's listening.
192
00:05:44,800 --> 00:05:46,000
So let me, let me frame this up with a
193
00:05:46,000 --> 00:05:47,240
little bit of context ahead of time.
194
00:05:47,759 --> 00:05:48,920
First off, this brand
195
00:05:48,959 --> 00:05:50,240
is in the apparel space.
196
00:05:50,399 --> 00:05:52,680
Their AOV is over a hundred dollars, not
197
00:05:52,680 --> 00:05:53,639
quite one 50, but it's
198
00:05:53,639 --> 00:05:54,439
kind of in that range.
199
00:05:54,839 --> 00:05:56,000
Depends on the seasonality and what
200
00:05:56,000 --> 00:05:56,959
products are popular.
201
00:05:57,480 --> 00:05:58,959
And they have actually a
202
00:05:58,959 --> 00:06:01,519
fair amount of organic traffic.
203
00:06:01,560 --> 00:06:02,879
They influencer like
204
00:06:02,879 --> 00:06:03,759
posts, things like that.
205
00:06:03,959 --> 00:06:06,839
So if that's helpful for context for, for
206
00:06:06,839 --> 00:06:07,839
brands that are thinking about this.
207
00:06:08,120 --> 00:06:10,079
So the beginning of the year, we went
208
00:06:10,079 --> 00:06:11,480
through a bunch of different things with
209
00:06:11,480 --> 00:06:12,319
this, with this brand.
210
00:06:12,319 --> 00:06:14,639
And we started to test meta conversion
211
00:06:14,639 --> 00:06:16,519
live studies directly in the platform to
212
00:06:16,519 --> 00:06:18,319
try and tease out how incremental is
213
00:06:18,319 --> 00:06:19,600
meta, et cetera, et cetera.
214
00:06:19,680 --> 00:06:21,160
So here's what I can tell you.
215
00:06:21,160 --> 00:06:23,240
So their average one day click row is,
216
00:06:23,240 --> 00:06:25,240
which, which shows up, shows up here as a
217
00:06:25,240 --> 00:06:28,600
3.12 for, for the account or basically
218
00:06:28,600 --> 00:06:30,439
year to date up to this point is very,
219
00:06:30,439 --> 00:06:32,279
very close to the actual conversion live
220
00:06:32,279 --> 00:06:33,959
data results that we got back from meta.
221
00:06:34,120 --> 00:06:35,879
So there's like almost no difference
222
00:06:35,879 --> 00:06:38,360
between their one day click and the
223
00:06:38,399 --> 00:06:39,759
incremental number that we
224
00:06:39,759 --> 00:06:41,120
got back from, from meta.
225
00:06:41,120 --> 00:06:42,959
So that's encouraging because that means
226
00:06:42,959 --> 00:06:45,360
that we can use one day click as a, as
227
00:06:45,360 --> 00:06:46,240
this, I don't know if
228
00:06:46,240 --> 00:06:47,279
barometer is the right word.
229
00:06:47,399 --> 00:06:48,800
We can use it as a temperature check to
230
00:06:48,800 --> 00:06:51,079
say, Hey, if we're going to hold our
231
00:06:51,079 --> 00:06:52,399
performance accountable to something,
232
00:06:52,600 --> 00:06:53,439
this one day click
233
00:06:53,439 --> 00:06:54,879
number is reasonable for them.
234
00:06:54,959 --> 00:06:56,600
And one day click is probably, you know,
235
00:06:56,759 --> 00:06:58,600
like I said, for the organic traffic
236
00:06:58,600 --> 00:07:00,040
reasons that I mentioned, one day click
237
00:07:00,040 --> 00:07:02,240
is probably more applicable to them for
238
00:07:02,240 --> 00:07:03,160
them than the average
239
00:07:03,240 --> 00:07:04,680
account where seven day click.
240
00:07:05,120 --> 00:07:06,079
You know, I think there's this, this,
241
00:07:06,279 --> 00:07:07,199
this stat that's floating around
242
00:07:07,199 --> 00:07:09,079
recently, which is 70 click kind of under
243
00:07:09,079 --> 00:07:10,360
reports by 15 to 20%.
244
00:07:10,600 --> 00:07:11,560
We generally see that
245
00:07:11,639 --> 00:07:12,560
except for this account.
246
00:07:12,560 --> 00:07:14,040
So just, just kind of framing that up.
247
00:07:14,079 --> 00:07:15,680
So we ran a conversion of study to start
248
00:07:15,680 --> 00:07:17,279
the year, came back and basically said,
249
00:07:17,319 --> 00:07:18,800
Hey, your one day click row eyes number
250
00:07:18,839 --> 00:07:20,800
is basically your, your, your, um,
251
00:07:21,120 --> 00:07:22,680
incremental number, which is different
252
00:07:22,680 --> 00:07:24,920
than incremental, uh, attribution.
253
00:07:24,920 --> 00:07:25,720
Those are, those are different things.
254
00:07:25,720 --> 00:07:26,560
So just framing that up.
255
00:07:27,000 --> 00:07:28,480
So what we did is we ran through that
256
00:07:28,480 --> 00:07:29,800
conversional test and we started to get
257
00:07:30,319 --> 00:07:32,600
through the rest of, um, the, the, the Q1
258
00:07:32,600 --> 00:07:35,199
Q Q2 rolls around beginning of Q2 rolls
259
00:07:35,199 --> 00:07:36,120
around end of March.
260
00:07:36,120 --> 00:07:37,439
And we want to say, Hey,
261
00:07:38,079 --> 00:07:38,959
incremental attribution.
262
00:07:38,959 --> 00:07:39,879
We're starting to see some
263
00:07:39,879 --> 00:07:41,000
success in other accounts.
264
00:07:41,000 --> 00:07:41,800
We're hearing about it.
265
00:07:41,800 --> 00:07:42,319
Meta started to
266
00:07:42,319 --> 00:07:43,519
suggest it a little bit more.
267
00:07:43,800 --> 00:07:45,439
We want to actually test if this is going
268
00:07:45,439 --> 00:07:46,639
to be successful or not.
269
00:07:46,839 --> 00:07:49,879
So what we do is we launch, um, we take
270
00:07:49,879 --> 00:07:51,680
our top performing campaign on a one day
271
00:07:51,680 --> 00:07:53,959
click basis and we split out and have an
272
00:07:53,959 --> 00:07:55,639
additional cell for incremental
273
00:07:55,639 --> 00:07:57,439
attribution and throw them both into the
274
00:07:57,439 --> 00:07:58,240
conversion of study.
275
00:07:58,279 --> 00:07:58,920
So that's the test.
276
00:07:58,920 --> 00:08:01,319
It's one day click versus incremental
277
00:08:01,319 --> 00:08:02,240
attribution with a
278
00:08:02,240 --> 00:08:03,680
holdout for each of them.
279
00:08:03,680 --> 00:08:05,399
Um, so that's kind of test design.
280
00:08:05,800 --> 00:08:08,079
Uh, we go through the first week of
281
00:08:08,079 --> 00:08:10,240
performance and incremental attribution
282
00:08:10,519 --> 00:08:12,160
itself looks terrible.
283
00:08:12,240 --> 00:08:13,399
Well, I, I have a, I have
284
00:08:13,399 --> 00:08:14,399
a tweet thread about this.
285
00:08:14,560 --> 00:08:15,680
If you go to my Twitter, you can find it.
286
00:08:15,680 --> 00:08:16,600
And it just looks miserable.
287
00:08:16,600 --> 00:08:18,199
And I'm like, should we turn this off?
288
00:08:18,199 --> 00:08:18,879
Like, I'm really, I'm
289
00:08:18,879 --> 00:08:19,959
really like nervous.
290
00:08:19,959 --> 00:08:21,040
This is not going to pan out.
291
00:08:21,319 --> 00:08:22,879
Um, but no, we decided to commit to this
292
00:08:22,879 --> 00:08:23,480
and set up the test.
293
00:08:23,480 --> 00:08:25,160
So like, let's give it more time week
294
00:08:25,160 --> 00:08:27,879
two, three roll around and the conversion
295
00:08:27,879 --> 00:08:30,360
lift results inside of meta are reporting
296
00:08:30,360 --> 00:08:32,600
much, much better now in favor of
297
00:08:32,600 --> 00:08:33,519
incremental attribution.
298
00:08:33,519 --> 00:08:34,399
So I'm starting to feel good.
299
00:08:34,600 --> 00:08:35,960
Uh, I see, Hey, one day
300
00:08:35,960 --> 00:08:36,799
click on it looks good.
301
00:08:36,840 --> 00:08:38,159
Incremental attribution looks good.
302
00:08:38,320 --> 00:08:38,919
The commercial lift
303
00:08:38,919 --> 00:08:40,240
study results look good.
304
00:08:40,360 --> 00:08:41,759
They use MTA data as well.
305
00:08:41,759 --> 00:08:43,039
So they use, they use triple oil.
306
00:08:43,279 --> 00:08:44,679
Um, uh, but your, your North
307
00:08:44,679 --> 00:08:45,840
beam would be a comparable.
308
00:08:46,000 --> 00:08:47,840
It's like that also looked good.
309
00:08:47,840 --> 00:08:49,080
So everything was checking out that this
310
00:08:49,080 --> 00:08:49,759
was looking really good.
311
00:08:50,159 --> 00:08:51,679
And then, uh, we get to the end of the
312
00:08:51,679 --> 00:08:53,679
test and the results basically say
313
00:08:53,960 --> 00:08:56,399
incremental attribution is substantially
314
00:08:56,399 --> 00:08:58,360
more incremental than one day click.
315
00:08:58,399 --> 00:08:59,519
So there's a thrown around the word
316
00:08:59,519 --> 00:09:00,879
incremental, like a million times.
317
00:09:01,159 --> 00:09:02,320
Um, so if it's worth
318
00:09:02,320 --> 00:09:04,279
digging into that further, I can.
319
00:09:04,279 --> 00:09:06,799
But, um, basically, uh, incremental
320
00:09:06,840 --> 00:09:08,879
attribution was performing way better
321
00:09:08,919 --> 00:09:09,879
than one day click rise.
322
00:09:10,120 --> 00:09:12,320
So we say, okay, we feel
323
00:09:12,320 --> 00:09:13,639
pretty confident in these results.
324
00:09:13,720 --> 00:09:15,000
Um, now it's only one test.
325
00:09:15,000 --> 00:09:16,240
And so maybe it's worth repeating.
326
00:09:16,240 --> 00:09:17,519
It's on one skew and a
327
00:09:17,519 --> 00:09:18,960
subset of hundreds of skews.
328
00:09:19,120 --> 00:09:20,519
Uh, so before we roll this, you know,
329
00:09:20,519 --> 00:09:22,039
maybe before we roll it out aggressively
330
00:09:22,039 --> 00:09:23,840
or just like blanket account wide, we
331
00:09:23,840 --> 00:09:25,519
kind of want to go campaign by campaign.
332
00:09:25,960 --> 00:09:27,039
So we start to roll it out
333
00:09:27,120 --> 00:09:28,879
to several more campaigns.
334
00:09:29,720 --> 00:09:30,879
And what has ensued over the last couple
335
00:09:30,879 --> 00:09:32,759
of weeks is I think what's, what's really
336
00:09:32,759 --> 00:09:34,039
the visuals behind me.
337
00:09:34,320 --> 00:09:36,840
Um, and I can go into that in a second.
338
00:09:36,879 --> 00:09:38,240
Andrew, do you have any, any questions or
339
00:09:38,240 --> 00:09:40,120
like anything I can clarify before I dig
340
00:09:40,120 --> 00:09:40,799
into like what we've
341
00:09:40,799 --> 00:09:41,799
been seeing more recently?
342
00:09:42,399 --> 00:09:44,919
I mean, I think the big thing that you
343
00:09:44,919 --> 00:09:46,480
mentioned is basically like it was the
344
00:09:46,480 --> 00:09:48,279
patience that sort of paid off for you.
345
00:09:49,240 --> 00:09:51,440
Like, and that you're using
346
00:09:51,440 --> 00:09:52,639
incrementality as like a
347
00:09:52,639 --> 00:09:55,639
main measurement guide.
348
00:09:55,960 --> 00:09:56,799
You know what I'm saying?
349
00:09:57,600 --> 00:09:58,559
Like it's so much around
350
00:09:58,559 --> 00:10:00,080
like, is this truly incremental?
351
00:10:00,320 --> 00:10:02,399
And, and you waited, which is really hard
352
00:10:02,399 --> 00:10:02,960
for a lot of people to
353
00:10:02,960 --> 00:10:03,919
do both of those things.
354
00:10:04,360 --> 00:10:04,639
Yeah.
355
00:10:05,159 --> 00:10:05,559
Yes.
356
00:10:05,559 --> 00:10:05,759
Yeah.
357
00:10:05,799 --> 00:10:06,480
Very, very difficult.
358
00:10:06,480 --> 00:10:07,960
Especially after that first week started
359
00:10:07,960 --> 00:10:09,159
to look, look miserable.
360
00:10:09,159 --> 00:10:10,399
And when I say look miserable, it's like
361
00:10:10,960 --> 00:10:11,840
triple whale looked bad.
362
00:10:11,879 --> 00:10:13,120
One day clicked, looked really bad.
363
00:10:13,320 --> 00:10:14,840
Um, and it just seemed to take some time.
364
00:10:14,840 --> 00:10:16,000
And that's the sentiment that I've heard
365
00:10:16,039 --> 00:10:17,840
in the Foxhole group from people is like,
366
00:10:18,039 --> 00:10:20,000
Hey, either, uh, generally speaking at
367
00:10:20,000 --> 00:10:21,759
first, it's not going to look great, but
368
00:10:21,759 --> 00:10:22,799
over time it starts to get better.
369
00:10:22,799 --> 00:10:24,960
Uh, and then the sentiment I've been
370
00:10:24,960 --> 00:10:25,919
hearing lately is that
371
00:10:26,279 --> 00:10:27,360
it starts to fall off.
372
00:10:27,440 --> 00:10:28,519
Uh, and that's where these,
373
00:10:28,519 --> 00:10:29,320
these graphs start to come in.
374
00:10:29,320 --> 00:10:30,679
So I'm going to try to zoom in a little
375
00:10:30,679 --> 00:10:31,799
bit here and I'll kind
376
00:10:31,799 --> 00:10:32,600
of show you the change.
377
00:10:32,600 --> 00:10:34,799
So you can see kind of Q1 and this is the
378
00:10:34,799 --> 00:10:36,159
one day click versus incremental.
379
00:10:36,159 --> 00:10:37,279
The dotted line is incremental.
380
00:10:37,279 --> 00:10:38,720
So they're like very close in line.
381
00:10:38,759 --> 00:10:40,159
Like that's the thing to take away from
382
00:10:40,159 --> 00:10:42,639
this visual is that they are tightly
383
00:10:43,399 --> 00:10:45,200
correlated, not only correlated, but
384
00:10:45,200 --> 00:10:46,559
they're actually pretty close together.
385
00:10:46,559 --> 00:10:48,039
So one day click looks increment, looks
386
00:10:48,039 --> 00:10:49,919
like incremental attribution incremental,
387
00:10:49,919 --> 00:10:51,279
maybe parts a little bit higher, which I
388
00:10:51,279 --> 00:10:51,960
think you'd expect
389
00:10:51,960 --> 00:10:52,720
because one day click is
390
00:10:52,759 --> 00:10:55,120
very conservative, but then we start to
391
00:10:55,120 --> 00:10:57,200
have things trend off
392
00:10:57,200 --> 00:10:58,240
aggressively over time.
393
00:10:58,240 --> 00:11:00,039
So, uh, I'm highlighting on the graph,
394
00:11:00,240 --> 00:11:02,360
March 25th, we start to see this
395
00:11:02,360 --> 00:11:05,080
deviation from one day click and
396
00:11:05,080 --> 00:11:06,000
incremental attribution.
397
00:11:06,120 --> 00:11:08,720
They still follow each other from like a
398
00:11:08,759 --> 00:11:09,679
correlation standpoint.
399
00:11:09,679 --> 00:11:11,519
They're still pretty close, but the gap
400
00:11:11,519 --> 00:11:12,639
between those two lines
401
00:11:12,759 --> 00:11:14,600
starts to grow substantially.
402
00:11:14,879 --> 00:11:16,159
Um, and there's a couple of things that
403
00:11:16,159 --> 00:11:17,559
happened around this time.
404
00:11:17,600 --> 00:11:19,600
And that's where we're like, uh, from,
405
00:11:19,600 --> 00:11:21,039
from then to now is what we're trying to,
406
00:11:21,039 --> 00:11:22,279
trying to dig into and understand.
407
00:11:22,679 --> 00:11:24,480
So the first thing that happened, um,
408
00:11:24,480 --> 00:11:25,840
which was at kind of the beginning of
409
00:11:25,840 --> 00:11:27,960
March meta says, we are changing the way
410
00:11:27,960 --> 00:11:29,720
click attribution works.
411
00:11:29,759 --> 00:11:32,120
Um, we are going to say an outbound click
412
00:11:32,159 --> 00:11:34,919
is click attribution, but engagement, uh,
413
00:11:34,919 --> 00:11:36,279
the kind of engaged version
414
00:11:36,320 --> 00:11:37,960
of, of that is going to change.
415
00:11:38,000 --> 00:11:39,960
Even though a click to a profile, a click
416
00:11:39,960 --> 00:11:41,440
to comment, whatever things that used to
417
00:11:41,440 --> 00:11:42,799
historically you've been counted as, as
418
00:11:42,799 --> 00:11:44,039
click conversions, that's
419
00:11:44,039 --> 00:11:45,320
now moving to an engaged.
420
00:11:45,600 --> 00:11:48,679
Um, so that changed, um, and then, uh, so
421
00:11:48,720 --> 00:11:49,919
that changed and then started to roll out
422
00:11:49,960 --> 00:11:51,399
to other accounts that they announced at
423
00:11:51,399 --> 00:11:52,039
the beginning of March.
424
00:11:52,039 --> 00:11:53,039
I think later, later in
425
00:11:53,039 --> 00:11:54,080
March, it started to roll out.
426
00:11:54,320 --> 00:11:55,519
Um, that was one of the big things.
427
00:11:55,799 --> 00:11:57,799
Uh, we ended our test probably at the
428
00:11:57,799 --> 00:11:59,799
beginning of April, uh, and we started to
429
00:11:59,799 --> 00:12:01,759
roll out the test, uh, then as well.
430
00:12:02,080 --> 00:12:04,000
And that's kind of, since, since those
431
00:12:04,000 --> 00:12:05,679
things have happened, um, the kind of
432
00:12:05,679 --> 00:12:07,399
ensuing results have been, there's this
433
00:12:07,399 --> 00:12:09,399
deviation in one day click and
434
00:12:09,399 --> 00:12:10,279
incremental attribution.
435
00:12:10,279 --> 00:12:12,600
Uh, so that's an insight.
436
00:12:12,639 --> 00:12:14,840
Uh, and our question now becomes, uh,
437
00:12:15,000 --> 00:12:16,799
well, is there, is there a third data
438
00:12:16,799 --> 00:12:18,200
point that we can look at or a couple
439
00:12:18,200 --> 00:12:19,879
more data points that we can look at and
440
00:12:19,879 --> 00:12:20,679
ask the questions like,
441
00:12:20,679 --> 00:12:22,080
well, which one should we trust?
442
00:12:22,320 --> 00:12:24,720
Should we go back to one day click and
443
00:12:24,720 --> 00:12:26,200
use that as our guide?
444
00:12:26,399 --> 00:12:27,240
Or should we lean into
445
00:12:27,240 --> 00:12:28,200
incremental attribution?
446
00:12:29,000 --> 00:12:31,519
Because just due to the nature of how one
447
00:12:31,519 --> 00:12:33,360
day click and incremental work, kind of
448
00:12:33,440 --> 00:12:34,919
like we said earlier, one day click is
449
00:12:34,919 --> 00:12:35,960
only giving credit to people
450
00:12:35,960 --> 00:12:37,320
who clicked within 24 hours.
451
00:12:37,480 --> 00:12:39,799
By default, it is substantially more, uh,
452
00:12:39,799 --> 00:12:40,440
conservative than
453
00:12:40,440 --> 00:12:41,440
incremental attribution.
454
00:12:41,840 --> 00:12:42,879
So we don't want to be
455
00:12:42,919 --> 00:12:44,080
misled by one day click.
456
00:12:44,120 --> 00:12:45,159
What, what should we do?
457
00:12:45,480 --> 00:12:46,879
That third data point that I'm alluding
458
00:12:46,879 --> 00:12:50,279
to is triple whale, um, uh, MTA tool.
459
00:12:50,320 --> 00:12:52,879
Uh, historically triple has like largely
460
00:12:52,879 --> 00:12:54,480
followed, uh, the one day click.
461
00:12:54,519 --> 00:12:55,879
And so we've built up a lot of confidence
462
00:12:55,879 --> 00:12:57,559
in one day click in the platform because,
463
00:12:57,799 --> 00:12:58,960
um, we've spent three
464
00:12:58,960 --> 00:13:00,080
years using one day click.
465
00:13:00,120 --> 00:13:00,919
It's worked extremely
466
00:13:00,919 --> 00:13:01,759
well for those reasons.
467
00:13:02,080 --> 00:13:04,399
Um, and that is now also deviating very
468
00:13:04,399 --> 00:13:05,080
substantially from
469
00:13:05,080 --> 00:13:06,000
incremental attribution.
470
00:13:06,000 --> 00:13:08,759
Um, not to continue this yap, but I'll
471
00:13:08,759 --> 00:13:10,679
pause in a second, but, um, you would
472
00:13:10,679 --> 00:13:12,279
also expect that because triple is using
473
00:13:12,279 --> 00:13:13,840
click based it's, it's click based.
474
00:13:13,840 --> 00:13:15,960
So like it's not completely unexpected.
475
00:13:16,200 --> 00:13:17,960
Um, and at the same time, the deviation
476
00:13:18,200 --> 00:13:19,639
leaves us a little bit concerned.
477
00:13:20,080 --> 00:13:22,679
The final data point that we used, um, to
478
00:13:22,679 --> 00:13:24,159
kind of look at this more recently is,
479
00:13:24,279 --> 00:13:25,960
well, how did the actual new customer
480
00:13:25,960 --> 00:13:28,720
sales for these products look over
481
00:13:28,720 --> 00:13:30,159
particularly the last couple of weeks?
482
00:13:30,320 --> 00:13:31,720
We've had Memorial Day recently, so that,
483
00:13:31,720 --> 00:13:33,679
that certainly throws it off, but we've
484
00:13:33,679 --> 00:13:36,120
definitely noticed a downward trend in
485
00:13:36,120 --> 00:13:37,360
the overall efficiency.
486
00:13:37,360 --> 00:13:38,559
So something that we like to do for
487
00:13:38,559 --> 00:13:40,960
brands that have multiple products or
488
00:13:40,960 --> 00:13:43,120
product categories is take, um, as much
489
00:13:43,120 --> 00:13:45,440
as we can look at product level sales
490
00:13:45,440 --> 00:13:46,600
with product level spend.
491
00:13:46,639 --> 00:13:47,879
It's not perfect, but it allows you to
492
00:13:47,879 --> 00:13:48,720
see trends over time.
493
00:13:49,360 --> 00:13:50,879
Um, so if I were to take new customer
494
00:13:50,879 --> 00:13:54,000
sales of this, of SKU a divided by spend
495
00:13:54,000 --> 00:13:55,279
for SKU a, what is that?
496
00:13:55,320 --> 00:13:57,240
What is that product level MER that has
497
00:13:57,240 --> 00:13:58,600
gotten less efficient over
498
00:13:58,600 --> 00:13:59,440
the last couple of weeks?
499
00:13:59,919 --> 00:14:02,159
So what that's leading me to say is,
500
00:14:02,360 --> 00:14:03,159
okay, well, where do we go
501
00:14:03,159 --> 00:14:04,759
from here with this information?
502
00:14:05,000 --> 00:14:07,000
We are actually going to start deviating
503
00:14:07,000 --> 00:14:09,200
back to one day click in the short term
504
00:14:09,240 --> 00:14:10,360
until we feel more confident in
505
00:14:10,360 --> 00:14:11,120
incremental attribution.
506
00:14:11,480 --> 00:14:12,200
Maybe we need to run a
507
00:14:12,200 --> 00:14:13,200
new conversion of study.
508
00:14:13,440 --> 00:14:15,679
Um, but all of that to basically say, I
509
00:14:15,679 --> 00:14:16,879
don't really feel confident in
510
00:14:16,879 --> 00:14:18,440
incremental attribution for this account
511
00:14:18,440 --> 00:14:20,080
at this point in time, because it just
512
00:14:20,080 --> 00:14:22,120
seems to be heavy, heavily deviating
513
00:14:22,120 --> 00:14:23,559
from, um, the
514
00:14:23,559 --> 00:14:24,879
performance that we did see.
515
00:14:25,279 --> 00:14:28,519
Even though we started to validate it
516
00:14:28,519 --> 00:14:30,519
with an actual conversion
517
00:14:30,559 --> 00:14:32,240
lift test inside of meta.
518
00:14:32,360 --> 00:14:33,799
So, um, that's, that's,
519
00:14:33,919 --> 00:14:35,120
that's my personal anecdote.
520
00:14:35,120 --> 00:14:36,600
One other anecdote on incremental
521
00:14:36,600 --> 00:14:38,600
attribution, another account, same thing.
522
00:14:38,679 --> 00:14:39,879
Like it started off really slow.
523
00:14:40,039 --> 00:14:41,600
It got really strong for two weeks.
524
00:14:42,039 --> 00:14:42,679
Last couple of weeks,
525
00:14:42,679 --> 00:14:43,279
it's been miserable.
526
00:14:43,279 --> 00:14:44,600
So we're just turning it back off, going
527
00:14:44,600 --> 00:14:45,080
straight back to the
528
00:14:45,080 --> 00:14:45,919
seven day click over there.
529
00:14:46,080 --> 00:14:47,440
How much are they spending by the way?
530
00:14:47,440 --> 00:14:49,759
They'll spend 150, 200 K a month.
531
00:14:49,840 --> 00:14:52,720
And what's your hypothesis?
532
00:14:52,759 --> 00:14:55,480
So this is what it is now, but what's
533
00:14:55,480 --> 00:14:57,360
your hypothesis about like how this is
534
00:14:57,360 --> 00:15:01,240
going to change in the next six months?
535
00:15:01,600 --> 00:15:04,639
Yeah, I have this, this brand because of
536
00:15:04,639 --> 00:15:06,240
the organic trend of this brand.
537
00:15:06,399 --> 00:15:09,200
Um, that's why I've thought that one day
538
00:15:09,200 --> 00:15:11,000
click is, and we've, we've kind of
539
00:15:11,000 --> 00:15:12,519
experimented in one day click has been
540
00:15:12,519 --> 00:15:14,679
the more reliable thing to use.
541
00:15:15,000 --> 00:15:16,279
Seven day click for them.
542
00:15:16,600 --> 00:15:18,399
It's a little bit too wild because if
543
00:15:18,399 --> 00:15:19,879
they have an organic moment or something
544
00:15:19,879 --> 00:15:21,919
pops off, meta does not react quickly
545
00:15:21,919 --> 00:15:23,360
enough to bid back down.
546
00:15:23,360 --> 00:15:24,600
They use cost controls as well, pretty
547
00:15:24,600 --> 00:15:26,960
heavy on cost controls, cost cap and min
548
00:15:26,960 --> 00:15:29,720
row as, um, meta doesn't throttle back
549
00:15:29,720 --> 00:15:30,960
down quickly enough
550
00:15:30,960 --> 00:15:32,039
using seven day click.
551
00:15:32,240 --> 00:15:33,600
It still doesn't with one day click
552
00:15:33,600 --> 00:15:34,679
either, but we can manually
553
00:15:34,679 --> 00:15:35,639
make some adjustments there.
554
00:15:36,000 --> 00:15:37,879
And so, um, one day click
555
00:15:37,879 --> 00:15:38,919
for that reason feels better.
556
00:15:38,919 --> 00:15:40,080
The reason we'd like to incremental
557
00:15:40,080 --> 00:15:42,720
because it opened up, it was the theory
558
00:15:42,720 --> 00:15:45,159
was it would open up maybe the delivery a
559
00:15:45,159 --> 00:15:46,960
little bit more and be a little less
560
00:15:46,960 --> 00:15:48,360
conservative on who we're bidding for.
561
00:15:48,399 --> 00:15:49,600
Now, obviously we could change our bids.
562
00:15:49,600 --> 00:15:50,320
We can increase the budget.
563
00:15:50,320 --> 00:15:51,039
We could just force more
564
00:15:51,039 --> 00:15:52,080
spend into one day click.
565
00:15:52,399 --> 00:15:53,639
Um, but the idea, and I
566
00:15:53,639 --> 00:15:54,720
think this is why Barry likes it.
567
00:15:54,720 --> 00:15:56,080
Barry had a great explanation on Twitter.
568
00:15:56,080 --> 00:15:58,720
It's like in theory, it should solve all
569
00:15:58,720 --> 00:16:01,279
of the one day view weird
570
00:16:01,279 --> 00:16:02,519
issues that, that you get.
571
00:16:02,879 --> 00:16:05,480
Um, but I think the only hypothesis that
572
00:16:05,480 --> 00:16:06,679
I can really think of for this brand
573
00:16:06,679 --> 00:16:09,120
right now is that incremental attribution
574
00:16:09,120 --> 00:16:11,159
is not, it's not doing a great job for
575
00:16:11,159 --> 00:16:12,559
this account at teasing out what's
576
00:16:12,559 --> 00:16:13,919
actually incremental.
577
00:16:13,960 --> 00:16:15,399
And it's still conflating that with some
578
00:16:15,399 --> 00:16:17,039
of the view based conversion, some of the
579
00:16:17,039 --> 00:16:18,799
engaged conversions, uh, and
580
00:16:18,799 --> 00:16:20,240
some of that are organic demand.
581
00:16:20,240 --> 00:16:22,279
Uh, and so what it thinks is incremental
582
00:16:22,279 --> 00:16:24,799
is actually purchases that, um, I guess
583
00:16:24,799 --> 00:16:25,360
would have happened
584
00:16:25,399 --> 00:16:27,080
anyways, which are not incremental.
585
00:16:27,120 --> 00:16:28,559
What I think will happen, what I hope
586
00:16:28,559 --> 00:16:31,360
will happen is, I mean, credit to meta.
587
00:16:31,360 --> 00:16:32,759
Um, they, they allow you to do conversion
588
00:16:32,759 --> 00:16:34,320
of studies and don't make you spend
589
00:16:34,360 --> 00:16:35,240
absurd amounts of money.
590
00:16:35,480 --> 00:16:36,960
And they want this product to solve the
591
00:16:36,960 --> 00:16:38,039
problems that people are having.
592
00:16:38,279 --> 00:16:39,399
Like they're, they're, they're the ones
593
00:16:39,399 --> 00:16:41,600
that are saying, um, they're pushing
594
00:16:41,600 --> 00:16:42,799
people to test these things.
595
00:16:42,960 --> 00:16:44,440
And so I think as more conversion, uh,
596
00:16:45,039 --> 00:16:47,799
studies are run, um, and as, uh, more
597
00:16:47,799 --> 00:16:49,000
brands are doing this, I think
598
00:16:49,000 --> 00:16:50,720
incremental, incremental attribution
599
00:16:50,919 --> 00:16:52,320
should only get better.
600
00:16:52,320 --> 00:16:54,279
So I think three months, six months from
601
00:16:54,279 --> 00:16:55,080
now, we should be
602
00:16:55,080 --> 00:16:57,039
retesting it, um, for this brand.
603
00:16:57,519 --> 00:16:58,360
It's going to be a wait though.
604
00:16:58,399 --> 00:17:01,440
It's that with IA, they just don't have
605
00:17:01,440 --> 00:17:03,679
the data now, like, or as
606
00:17:03,679 --> 00:17:06,720
much data as they could have.
607
00:17:06,799 --> 00:17:07,960
It's like, it's still too limiting.
608
00:17:08,920 --> 00:17:11,119
Or is it sort of one of those things that
609
00:17:11,119 --> 00:17:13,519
we see on meta where it's good initially.
610
00:17:13,519 --> 00:17:15,039
And then like, as more people adopted, it
611
00:17:15,039 --> 00:17:17,759
goes down like, or, you know, like, or do
612
00:17:17,759 --> 00:17:19,240
you see that happening in the future?
613
00:17:19,640 --> 00:17:22,039
Um, or like, what is it, do you think
614
00:17:22,319 --> 00:17:24,400
mechanically that was holding it from not
615
00:17:24,400 --> 00:17:26,039
being not performing the same way?
616
00:17:26,319 --> 00:17:27,240
Or is it too generous?
617
00:17:27,279 --> 00:17:27,960
Is it the opposite?
618
00:17:27,960 --> 00:17:30,160
I mean, I, I kind of, I'm reading what
619
00:17:30,160 --> 00:17:30,599
you're saying, but
620
00:17:30,599 --> 00:17:31,440
there's a little bit of both.
621
00:17:31,440 --> 00:17:33,200
It's like some of it's limiting, but some
622
00:17:33,200 --> 00:17:34,880
of it's also just like, seems like it's
623
00:17:34,880 --> 00:17:36,480
kind of taken, putting
624
00:17:36,480 --> 00:17:37,480
it out there too widely.
625
00:17:37,480 --> 00:17:38,319
And that's what you thought was going to
626
00:17:38,319 --> 00:17:38,960
happen, but it didn't.
627
00:17:40,240 --> 00:17:40,559
Yeah.
628
00:17:40,720 --> 00:17:41,839
It's a, it's a good question.
629
00:17:41,839 --> 00:17:43,599
I, it would be interesting if somebody
630
00:17:43,599 --> 00:17:45,400
from meta, um, unfortunately they just
631
00:17:45,400 --> 00:17:46,240
had a bunch of layoffs.
632
00:17:46,279 --> 00:17:47,720
So hopefully none, uh, no people in the
633
00:17:47,720 --> 00:17:48,920
incremental attribution department,
634
00:17:49,319 --> 00:17:50,920
because, uh, it would be very interesting
635
00:17:50,920 --> 00:17:52,880
to hear from meta how.
636
00:17:53,640 --> 00:17:54,440
It actually works.
637
00:17:54,440 --> 00:17:55,759
So my understanding of how incremental
638
00:17:55,759 --> 00:17:57,440
attribution works is that there's tons of
639
00:17:57,440 --> 00:17:59,559
brands at any given time that are running
640
00:17:59,759 --> 00:18:01,440
conversion of studies on, on, on.
641
00:18:01,599 --> 00:18:05,559
Meta and they meta use those to inform
642
00:18:05,680 --> 00:18:08,240
building this new optimization algorithm.
643
00:18:08,240 --> 00:18:09,359
I don't, I don't know exactly how they
644
00:18:09,359 --> 00:18:10,240
would describe it, but
645
00:18:10,480 --> 00:18:12,200
conversion of studies were used.
646
00:18:12,440 --> 00:18:13,680
Historical conversion of studies were
647
00:18:13,680 --> 00:18:15,079
used to build this model.
648
00:18:15,319 --> 00:18:16,400
And that's how it optimizes.
649
00:18:16,680 --> 00:18:18,079
I think the tricky thing about that is
650
00:18:18,079 --> 00:18:18,480
like, there's,
651
00:18:18,480 --> 00:18:19,599
there's so much difference.
652
00:18:20,079 --> 00:18:21,799
And this is true of one day click and
653
00:18:21,799 --> 00:18:22,880
seven day click as well, but like,
654
00:18:22,880 --> 00:18:25,039
there's so much difference and uniqueness
655
00:18:25,039 --> 00:18:27,119
to each individual brand, depending on
656
00:18:27,440 --> 00:18:29,279
how much organic traffic they have, how
657
00:18:29,279 --> 00:18:30,000
much spend they're
658
00:18:30,000 --> 00:18:31,200
getting from other channels.
659
00:18:31,440 --> 00:18:33,079
Um, the industry, like there's, there's
660
00:18:33,079 --> 00:18:35,200
so much nuance in between those that
661
00:18:35,279 --> 00:18:38,400
maybe for this brand, the, the mean of
662
00:18:38,519 --> 00:18:40,039
what they built conversion, uh,
663
00:18:40,039 --> 00:18:41,079
incremental attribution on
664
00:18:41,119 --> 00:18:42,799
top of just doesn't apply yet.
665
00:18:43,200 --> 00:18:44,599
Um, because the other, here's the other
666
00:18:44,599 --> 00:18:46,079
thing that I didn't, I didn't say like
667
00:18:46,079 --> 00:18:47,720
we've continued particularly over the
668
00:18:47,720 --> 00:18:48,759
last, before we made the
669
00:18:48,759 --> 00:18:50,640
decision to shut off, shut this off.
670
00:18:50,759 --> 00:18:53,039
We've continued to dial back the bids and
671
00:18:53,039 --> 00:18:54,119
tell meta, Hey, we don't,
672
00:18:54,160 --> 00:18:55,319
we're not happy with performance.
673
00:18:55,319 --> 00:18:57,440
Like we're, we should dial this back and
674
00:18:57,440 --> 00:18:58,079
it's continued to
675
00:18:58,079 --> 00:18:59,759
basically spend regardless of that.
676
00:18:59,960 --> 00:19:02,160
It's not, it's not adjusting the way that
677
00:19:02,160 --> 00:19:03,079
one day click or seven day
678
00:19:03,079 --> 00:19:04,400
click does to bid changes.
679
00:19:04,680 --> 00:19:06,319
Um, so my, the only hypothesis that I
680
00:19:06,319 --> 00:19:07,799
really have at this point is that they
681
00:19:07,799 --> 00:19:09,960
built the model on kind of the, the
682
00:19:09,960 --> 00:19:11,279
average of the conversion of studies.
683
00:19:11,279 --> 00:19:12,880
And just, that's not working for this
684
00:19:12,880 --> 00:19:13,759
brand at this point in time.
685
00:19:14,200 --> 00:19:16,160
Um, I know a couple of brands that spend
686
00:19:16,160 --> 00:19:17,319
the majority of their budget through
687
00:19:17,640 --> 00:19:18,559
incremental attribution.
688
00:19:18,559 --> 00:19:19,799
So it's very possible that it works.
689
00:19:20,319 --> 00:19:22,200
You know, I'm curious if people are, uh,
690
00:19:22,200 --> 00:19:24,680
spending on multiple,
691
00:19:24,680 --> 00:19:27,359
uh, uh, attribution types.
692
00:19:27,400 --> 00:19:30,079
Um, and the, this, this needs to be one
693
00:19:30,079 --> 00:19:31,319
of the subsets of them.
694
00:19:31,559 --> 00:19:32,720
Um, and if they're reaching a different
695
00:19:32,720 --> 00:19:35,319
pocket of audience, but, uh, to the
696
00:19:35,319 --> 00:19:36,440
different pocket of audience, like we're
697
00:19:36,440 --> 00:19:37,200
not actually seeing
698
00:19:37,200 --> 00:19:38,200
really different delivery.
699
00:19:38,200 --> 00:19:39,880
There's a lot of overlap between one day
700
00:19:39,880 --> 00:19:41,039
click and incremental attribution.
701
00:19:41,039 --> 00:19:43,000
And the, um, so there's a
702
00:19:43,000 --> 00:19:44,000
lot of overlap and reach.
703
00:19:44,240 --> 00:19:45,880
And then the net new visit rates was
704
00:19:45,880 --> 00:19:46,960
basically the same between them.
705
00:19:46,960 --> 00:19:48,519
It's not like we were reaching a bunch of
706
00:19:48,519 --> 00:19:49,480
new people by using
707
00:19:49,480 --> 00:19:50,480
incremental attribution.
708
00:19:50,720 --> 00:19:52,640
Um, it was about the same, the same rate.
709
00:19:52,640 --> 00:19:54,839
And for this brand, that's like 70, 70%
710
00:19:55,359 --> 00:19:56,039
pretty consistently.
711
00:19:56,119 --> 00:19:56,279
Yeah.
712
00:19:56,279 --> 00:19:56,839
It's interesting.
713
00:19:56,839 --> 00:19:58,680
I mean, I think it's, it's, it's sort of
714
00:19:58,680 --> 00:20:01,279
in this weird, interesting phase where
715
00:20:01,880 --> 00:20:02,359
you hear a lot of
716
00:20:02,359 --> 00:20:03,359
people talking about it.
717
00:20:03,359 --> 00:20:04,440
There's a lot of hope in it.
718
00:20:04,519 --> 00:20:06,279
You know, do you think if you made this
719
00:20:06,279 --> 00:20:07,880
like one day click and you made it seven
720
00:20:07,880 --> 00:20:10,079
day click versus incremental attribution,
721
00:20:10,599 --> 00:20:11,480
what do you think would happen?
722
00:20:11,519 --> 00:20:11,880
That's it?
723
00:20:11,920 --> 00:20:13,039
Yeah, that's a good question.
724
00:20:13,119 --> 00:20:14,279
All right, friends, quick break.
725
00:20:14,279 --> 00:20:15,839
This episode of scalability school is
726
00:20:15,839 --> 00:20:17,680
brought to you by North beam
727
00:20:17,720 --> 00:20:18,960
incrementality, which
728
00:20:19,039 --> 00:20:20,839
honestly is so awesome.
729
00:20:21,039 --> 00:20:21,680
Let me just tell you
730
00:20:21,680 --> 00:20:22,279
a little bit about it.
731
00:20:22,519 --> 00:20:24,359
If you're a marketer, you probably agree.
732
00:20:24,400 --> 00:20:26,480
Incrementality testing is broken.
733
00:20:26,759 --> 00:20:27,359
It's insane.
734
00:20:27,359 --> 00:20:28,920
You got to do all this crazy stuff, but
735
00:20:28,920 --> 00:20:29,759
with North beam, you do
736
00:20:29,759 --> 00:20:30,960
it right in the platform.
737
00:20:31,599 --> 00:20:33,960
It builds tests using your own MTA data,
738
00:20:34,240 --> 00:20:36,559
launches them directly in the ad platform
739
00:20:36,559 --> 00:20:38,680
and automatically monitors them
740
00:20:38,720 --> 00:20:39,680
throughout the test.
741
00:20:40,039 --> 00:20:41,640
And if something threatens validity,
742
00:20:41,680 --> 00:20:43,839
you're alerted to it within North beam
743
00:20:44,359 --> 00:20:46,200
before it ruins the results.
744
00:20:46,839 --> 00:20:48,799
The really nice thing is it solves all
745
00:20:48,799 --> 00:20:50,039
these problems you hear about
746
00:20:50,039 --> 00:20:51,960
incrementality testing all the time, like
747
00:20:52,200 --> 00:20:53,319
labor, spreadsheets,
748
00:20:53,319 --> 00:20:54,480
statistical reference.
749
00:20:54,480 --> 00:20:55,440
Designing tests is
750
00:20:55,440 --> 00:20:56,799
normally so slow, right?
751
00:20:56,799 --> 00:20:57,599
And error prone.
752
00:20:57,920 --> 00:20:59,839
So with North beam incrementality, it's
753
00:20:59,880 --> 00:21:01,319
automated and accurate.
754
00:21:01,680 --> 00:21:02,759
And your team is able
755
00:21:02,759 --> 00:21:04,240
to spend with confidence.
756
00:21:04,920 --> 00:21:05,680
Give it a shot with
757
00:21:05,680 --> 00:21:07,440
North beam incrementality.
758
00:21:07,799 --> 00:21:09,039
Really, the future of
759
00:21:09,039 --> 00:21:10,119
incrementality is here.
760
00:21:10,119 --> 00:21:11,680
And North beam has figured out a way to
761
00:21:11,680 --> 00:21:12,759
make it easy for you.
762
00:21:13,079 --> 00:21:14,480
Do you think if you made this like one
763
00:21:14,480 --> 00:21:16,319
day click and you made it seven day click
764
00:21:16,440 --> 00:21:18,599
versus incremental attribution, what do
765
00:21:18,599 --> 00:21:19,279
you think would happen?
766
00:21:19,720 --> 00:21:21,400
Yeah, that's a good question.
767
00:21:21,480 --> 00:21:24,519
I mean, that's what I'm kind of like
768
00:21:24,799 --> 00:21:26,400
struggling with, generally speaking,
769
00:21:26,400 --> 00:21:29,240
right now is the difference between a one
770
00:21:29,240 --> 00:21:30,759
day and a seven day click obviously
771
00:21:30,759 --> 00:21:32,119
changes the way that meta delivers.
772
00:21:32,279 --> 00:21:34,680
Because we use cost controls, generally
773
00:21:34,680 --> 00:21:36,880
speaking, the question that I would go
774
00:21:36,880 --> 00:21:39,559
back to is, okay, if they're not reach,
775
00:21:39,599 --> 00:21:41,160
if there's, there's overlap between the
776
00:21:41,160 --> 00:21:42,559
different attribution methods and who
777
00:21:42,559 --> 00:21:45,559
you're going to reach, then there's like
778
00:21:45,559 --> 00:21:47,400
not, not a ton of value in that.
779
00:21:48,039 --> 00:21:51,839
Because if I wanted to use seven day
780
00:21:51,839 --> 00:21:53,200
click because it was going to bid more
781
00:21:53,240 --> 00:21:55,400
aggressively than a one day click, well,
782
00:21:55,480 --> 00:21:57,160
the alternative thing I could do instead
783
00:21:57,160 --> 00:21:58,440
of switching from one day click to seven
784
00:21:58,440 --> 00:21:59,279
day click is just open the
785
00:21:59,279 --> 00:22:00,839
bids or spend more forcefully.
786
00:22:00,920 --> 00:22:03,039
Like I could just be more aggressive.
787
00:22:03,359 --> 00:22:05,640
Um, now you can make the argument that,
788
00:22:05,640 --> 00:22:07,559
okay, more data in a seven day click,
789
00:22:07,759 --> 00:22:09,440
maybe makes Facebook optimized a bit
790
00:22:09,440 --> 00:22:11,279
better, meta optimized a bit better and
791
00:22:11,279 --> 00:22:12,480
allows them to be more efficient.
792
00:22:12,720 --> 00:22:14,440
Um, I've just not really seen that to be
793
00:22:14,440 --> 00:22:15,640
the case, especially when you have a
794
00:22:15,640 --> 00:22:16,680
brand that's consistently
795
00:22:16,680 --> 00:22:18,240
getting through learning.
796
00:22:18,759 --> 00:22:20,559
And in this case, it's just been, it's
797
00:22:20,559 --> 00:22:22,200
just been more effective to use one day
798
00:22:22,200 --> 00:22:24,799
click because it doesn't go for those,
799
00:22:24,799 --> 00:22:26,599
uh, especially with the way that they
800
00:22:26,599 --> 00:22:29,160
changed click based attribution to be
801
00:22:29,559 --> 00:22:31,640
actual link click, as opposed to on all
802
00:22:31,640 --> 00:22:34,359
kinds of clicks that are, uh, whatever
803
00:22:34,359 --> 00:22:35,400
you could qualify as a click.
804
00:22:35,599 --> 00:22:35,759
Yeah.
805
00:22:35,759 --> 00:22:36,519
I don't know if that answers your
806
00:22:36,519 --> 00:22:37,920
question, but, um, I have my
807
00:22:38,160 --> 00:22:39,200
hypotheses around all of it.
808
00:22:39,400 --> 00:22:40,960
I mean, what's your, what do you think
809
00:22:40,960 --> 00:22:42,960
about the overall change with like the
810
00:22:42,960 --> 00:22:44,920
definition of clicks and
811
00:22:44,920 --> 00:22:45,839
like link clicks and stuff?
812
00:22:45,880 --> 00:22:47,319
Like, do you think it's been a net
813
00:22:47,319 --> 00:22:48,160
positive or do you think
814
00:22:48,160 --> 00:22:49,319
it's a net positive overall?
815
00:22:49,720 --> 00:22:50,400
What's your thinking?
816
00:22:50,440 --> 00:22:52,200
Cause it seems like it's, I mean, I think
817
00:22:52,200 --> 00:22:53,799
it should be too, but it seems like it's
818
00:22:53,799 --> 00:22:55,319
in the short term, like really screwed up
819
00:22:55,319 --> 00:22:57,839
some results because people were, it was
820
00:22:57,839 --> 00:22:59,160
optimizing on one thing and it's like
821
00:22:59,160 --> 00:23:01,279
almost had to relearn, you know?
822
00:23:01,440 --> 00:23:05,440
Um, I would assume that through like gem
823
00:23:05,440 --> 00:23:06,839
and lattice and all this, like one of the
824
00:23:06,839 --> 00:23:07,559
signals is this
825
00:23:07,559 --> 00:23:10,279
definition and that's changed.
826
00:23:10,319 --> 00:23:11,519
Right.
827
00:23:11,519 --> 00:23:14,880
Um, that's, uh, that's screwed, screwed
828
00:23:14,880 --> 00:23:16,079
things up to a degree.
829
00:23:16,640 --> 00:23:17,079
I don't know.
830
00:23:17,079 --> 00:23:17,680
What do you think?
831
00:23:18,279 --> 00:23:18,960
Yeah, I think that's,
832
00:23:18,960 --> 00:23:19,680
I think that's right.
833
00:23:19,720 --> 00:23:21,039
Um, it's, uh, it's
834
00:23:21,039 --> 00:23:23,079
definitely, uh, it's definitely.
835
00:23:23,599 --> 00:23:26,079
Pausing my, I think more confusion in the
836
00:23:26,079 --> 00:23:28,119
short term and especially for, for brands
837
00:23:28,119 --> 00:23:29,160
that are on cost controls, right?
838
00:23:29,200 --> 00:23:31,039
Like you, you went from bidding on a very
839
00:23:31,039 --> 00:23:32,720
specific type of, of, of
840
00:23:32,720 --> 00:23:34,440
click and that changed.
841
00:23:34,440 --> 00:23:36,519
And if, if the amount of conversions that
842
00:23:36,519 --> 00:23:39,039
were coming through changed by 20%, you
843
00:23:39,039 --> 00:23:40,839
have a completely different bid that you
844
00:23:40,839 --> 00:23:42,039
need to adjust for now.
845
00:23:42,240 --> 00:23:43,720
Um, and so again, I think like short term
846
00:23:43,759 --> 00:23:46,079
volatility, but in the longterm, I think
847
00:23:46,079 --> 00:23:47,319
it's the right thing to do because one,
848
00:23:47,319 --> 00:23:49,519
it adds like way more clarity for an
849
00:23:49,519 --> 00:23:50,519
advertiser setting up a
850
00:23:50,519 --> 00:23:51,559
campaign from scratch.
851
00:23:51,680 --> 00:23:52,200
It's like what they're
852
00:23:52,200 --> 00:23:53,759
actually giving meta credit for.
853
00:23:54,200 --> 00:23:56,160
Uh, and then two, you know, I think
854
00:23:56,160 --> 00:23:57,359
people will, I think generally people
855
00:23:57,359 --> 00:24:00,240
will agree that if you had to go rank
856
00:24:00,240 --> 00:24:01,880
order, what's actually incremental.
857
00:24:02,039 --> 00:24:04,079
It's a click is probably the most
858
00:24:04,079 --> 00:24:06,799
incremental, most commonly, uh, of an
859
00:24:06,799 --> 00:24:08,400
engagement of some kind.
860
00:24:08,400 --> 00:24:09,960
So a click to a social profile.
861
00:24:10,079 --> 00:24:11,920
And so I should redefine that an outbound
862
00:24:11,920 --> 00:24:13,160
click is probably the most incremental.
863
00:24:13,440 --> 00:24:15,960
The next range of the, what they qualify
864
00:24:15,960 --> 00:24:17,400
as engagement, you know, it's,
865
00:24:17,440 --> 00:24:18,720
it's video views of some sort.
866
00:24:18,720 --> 00:24:19,960
It's clicks to profiles,
867
00:24:19,960 --> 00:24:21,000
it's likes, it's whatever.
868
00:24:21,279 --> 00:24:22,599
It's probably the next most incremental.
869
00:24:22,720 --> 00:24:23,680
And the view is the thing
870
00:24:23,680 --> 00:24:24,599
that's least incremental.
871
00:24:24,960 --> 00:24:26,680
Um, but you know, how's that changing
872
00:24:26,680 --> 00:24:28,079
with short form video content?
873
00:24:28,079 --> 00:24:30,640
And apparently, um, zoomers don't click
874
00:24:30,640 --> 00:24:31,640
ads or, you know, that
875
00:24:31,799 --> 00:24:32,759
that's the, that's the story.
876
00:24:32,759 --> 00:24:34,079
So, um, yeah, I mean, in
877
00:24:34,079 --> 00:24:35,039
theory, it all sounds great.
878
00:24:35,039 --> 00:24:36,240
It's just a TBD, how it all
879
00:24:36,240 --> 00:24:37,640
continues to, uh, to play out.
880
00:24:38,039 --> 00:24:38,920
Oh, Zach's here.
881
00:24:39,160 --> 00:24:39,880
Hey, what's up, dude?
882
00:24:41,119 --> 00:24:41,559
Hi guys.
883
00:24:43,920 --> 00:24:45,519
So let's talk about the creative cohort
884
00:24:45,519 --> 00:24:47,880
spend analysis, uh, that you pulled,
885
00:24:48,200 --> 00:24:50,359
like, what's the context of this, Brad?
886
00:24:50,359 --> 00:24:52,640
Um, I'm, I'm interested to hear about it.
887
00:24:53,440 --> 00:24:56,440
So creative hit rates, creative volume.
888
00:24:56,440 --> 00:24:57,960
That's been like a topic of discussion.
889
00:24:57,960 --> 00:24:59,720
Motion put out an awesome, I don't know
890
00:24:59,720 --> 00:25:00,880
if exactly, I don't remember what they
891
00:25:00,880 --> 00:25:02,440
called it, but it was like the creative.
892
00:25:02,559 --> 00:25:02,880
I don't know.
893
00:25:02,920 --> 00:25:05,319
They put out this 2026 creative PDF.
894
00:25:05,319 --> 00:25:06,559
It was like 80 pages long.
895
00:25:06,960 --> 00:25:08,200
And it was awesome because it showed
896
00:25:08,200 --> 00:25:09,839
like, here's what the biggest brands are
897
00:25:09,839 --> 00:25:10,960
doing from a creative volume
898
00:25:10,960 --> 00:25:12,720
perspective and by industry.
899
00:25:12,720 --> 00:25:14,559
And then also it said like, here, here
900
00:25:14,559 --> 00:25:15,359
are the types of ads
901
00:25:15,359 --> 00:25:16,079
that brands are running.
902
00:25:16,079 --> 00:25:17,119
Cause they have this AI tagging.
903
00:25:17,119 --> 00:25:17,599
That's awesome.
904
00:25:17,880 --> 00:25:19,519
Um, and so we've been
905
00:25:19,519 --> 00:25:20,359
thinking about that a lot.
906
00:25:20,400 --> 00:25:21,480
We've been looking at a lot and kind of
907
00:25:21,480 --> 00:25:23,400
using that as a benchmark for, okay, how
908
00:25:23,400 --> 00:25:24,720
much are our brands launching
909
00:25:25,200 --> 00:25:26,720
on creative on a monthly basis?
910
00:25:26,720 --> 00:25:28,039
And, um, what does that mean?
911
00:25:28,440 --> 00:25:30,160
But I think there's been, I think we
912
00:25:30,160 --> 00:25:31,279
talked about this in a recent episode.
913
00:25:31,279 --> 00:25:32,920
I feel like there's like this sentiment
914
00:25:32,920 --> 00:25:35,359
building in the Twitter sphere for DTC
915
00:25:35,359 --> 00:25:36,200
where it's like, people are kind of
916
00:25:36,200 --> 00:25:36,880
getting frustrated with
917
00:25:36,880 --> 00:25:37,960
the volume game a little bit.
918
00:25:38,400 --> 00:25:39,960
And they're, they're leaning back towards
919
00:25:40,359 --> 00:25:42,240
hit rates, making better ads, spending
920
00:25:42,240 --> 00:25:43,279
more time, being more thoughtful.
921
00:25:43,279 --> 00:25:44,519
And I think that's a fair exercise.
922
00:25:44,559 --> 00:25:46,119
I'm still definitely in like the creative
923
00:25:46,160 --> 00:25:47,240
volume is important camp
924
00:25:47,279 --> 00:25:48,319
because you have to take swings.
925
00:25:48,599 --> 00:25:50,839
Um, and at the same time you want those
926
00:25:50,839 --> 00:25:52,039
swings to be calculated, but I don't
927
00:25:52,039 --> 00:25:53,039
think anybody would say
928
00:25:53,039 --> 00:25:54,079
it shouldn't be calculated.
929
00:25:54,160 --> 00:25:55,960
Uh, but that's, that's a,
930
00:25:56,200 --> 00:25:56,880
that's a separate topic.
931
00:25:57,799 --> 00:25:59,960
In any case, the, the thing that we've
932
00:25:59,960 --> 00:26:01,720
been thinking about and trying to map
933
00:26:01,720 --> 00:26:04,160
back to is this idea of viewing our
934
00:26:04,160 --> 00:26:06,119
creative through a cohort lens.
935
00:26:06,319 --> 00:26:08,279
And I think I'm not sure if this is in,
936
00:26:08,519 --> 00:26:10,200
um, there might be like a creative kind
937
00:26:10,200 --> 00:26:11,400
of report in, uh, what
938
00:26:11,400 --> 00:26:12,880
you and Kurt built, Andrew.
939
00:26:12,960 --> 00:26:14,000
Um, but it might be more
940
00:26:14,000 --> 00:26:15,319
of a CPMR related thing.
941
00:26:15,319 --> 00:26:16,440
It's like the creative cohorts in their
942
00:26:16,440 --> 00:26:17,680
CPMR, which is sick.
943
00:26:17,920 --> 00:26:19,000
Um, it's a really cool view.
944
00:26:19,319 --> 00:26:21,319
Um, but what we basically want to know,
945
00:26:21,319 --> 00:26:22,200
the question we're trying to answer is
946
00:26:22,200 --> 00:26:23,039
like, where did our
947
00:26:23,039 --> 00:26:24,880
spend come from this month?
948
00:26:24,920 --> 00:26:27,240
So in May, was it ads that we launched in
949
00:26:27,240 --> 00:26:29,000
May or is it everything that we've had in
950
00:26:29,000 --> 00:26:30,039
the ad accounts historically?
951
00:26:30,319 --> 00:26:31,680
And then I can go back and I can look and
952
00:26:31,680 --> 00:26:33,759
I could say, okay, if our entire ad
953
00:26:33,759 --> 00:26:36,119
account is made up from spend,
954
00:26:36,200 --> 00:26:38,440
ads that were launched last year, all the
955
00:26:38,440 --> 00:26:39,880
new shit I'm making is going to waste.
956
00:26:39,880 --> 00:26:41,359
Like I, we are not doing a good job at
957
00:26:41,359 --> 00:26:43,200
making new ads that are resonating.
958
00:26:43,599 --> 00:26:44,519
And then you can start to
959
00:26:44,519 --> 00:26:45,440
dig into the reasons why.
960
00:26:45,480 --> 00:26:47,799
So this is kind of a, this is a bit of a
961
00:26:47,799 --> 00:26:50,240
surface view of like, okay, um, are we
962
00:26:50,240 --> 00:26:51,279
making the right things?
963
00:26:51,519 --> 00:26:52,559
Are we actually progressing?
964
00:26:52,559 --> 00:26:53,319
And I have an example
965
00:26:53,319 --> 00:26:53,880
of what this looks like.
966
00:26:53,880 --> 00:26:55,759
So I can, I'm sorry for the non
967
00:26:55,759 --> 00:26:57,200
YouTubers, but I'll flip my screen over
968
00:26:57,200 --> 00:26:58,680
again and we can look at this.
969
00:26:58,839 --> 00:27:00,680
Um, I realized that I didn't update the
970
00:27:00,680 --> 00:27:02,519
previous document properly.
971
00:27:02,519 --> 00:27:03,720
So I have a different looking one here.
972
00:27:03,759 --> 00:27:04,680
Sorry, you blame, blame,
973
00:27:04,680 --> 00:27:05,839
Claude, Claude for the visual.
974
00:27:06,079 --> 00:27:07,160
The data is accurate though.
975
00:27:07,559 --> 00:27:10,559
Um, so I'll try to describe what's here.
976
00:27:10,599 --> 00:27:11,519
Can you guys, guys see this?
977
00:27:11,519 --> 00:27:11,759
Okay.
978
00:27:12,240 --> 00:27:12,559
Yeah.
979
00:27:13,359 --> 00:27:14,880
So what you'll see is
980
00:27:14,880 --> 00:27:16,160
we've got spend by month.
981
00:27:16,160 --> 00:27:18,119
So January 300 grand, February 400, some
982
00:27:18,119 --> 00:27:19,400
grand, March three, 300
983
00:27:19,440 --> 00:27:21,759
grand, April 289, May 258.
984
00:27:21,759 --> 00:27:22,640
So far we've got some
985
00:27:22,640 --> 00:27:23,279
time left in the month.
986
00:27:23,720 --> 00:27:26,519
Um, so this cohort view is what we've
987
00:27:26,519 --> 00:27:27,519
been finding really valuable.
988
00:27:27,640 --> 00:27:29,400
So if we look at January, we can see how
989
00:27:29,400 --> 00:27:31,400
much spend in January was made up from
990
00:27:31,400 --> 00:27:33,319
ads that were launched last year in this
991
00:27:33,319 --> 00:27:35,839
case and ads launched in January.
992
00:27:35,880 --> 00:27:37,240
And in this case, about
993
00:27:37,240 --> 00:27:38,720
10% was from ads in January.
994
00:27:39,039 --> 00:27:40,240
And I guess what you're looking for when
995
00:27:40,319 --> 00:27:42,400
you're looking at this is that the month
996
00:27:42,400 --> 00:27:44,039
that you're currently in and previous
997
00:27:44,039 --> 00:27:46,000
months are making up like a reasonable
998
00:27:46,039 --> 00:27:47,519
percentage because there's, there's two
999
00:27:47,519 --> 00:27:48,359
things that would be an issue.
1000
00:27:48,519 --> 00:27:50,640
If January here, which is 90%
1001
00:27:51,119 --> 00:27:53,559
pre-January, like if we're looking at May
1002
00:27:53,640 --> 00:27:55,759
and it was all ads launched last year, be
1003
00:27:55,759 --> 00:27:57,079
like, fuck, what have we been doing?
1004
00:27:57,519 --> 00:27:58,279
What have we been doing
1005
00:27:58,599 --> 00:27:59,440
this entire year so far?
1006
00:27:59,440 --> 00:28:00,839
We haven't made any ads that are working.
1007
00:28:01,119 --> 00:28:02,559
Um, and then that's, that's the trigger
1008
00:28:02,559 --> 00:28:03,400
to go look into that.
1009
00:28:03,400 --> 00:28:04,960
It's, it's, uh, it's using this.
1010
00:28:05,440 --> 00:28:06,759
You shouldn't just look at this and not
1011
00:28:06,799 --> 00:28:08,319
do something with it, but you're trying
1012
00:28:08,319 --> 00:28:11,039
to find insights to actually go and do
1013
00:28:11,039 --> 00:28:11,680
something with this.
1014
00:28:11,720 --> 00:28:13,480
But the other way, the other thing to
1015
00:28:13,480 --> 00:28:15,880
consider is, well, if all of your spend
1016
00:28:15,880 --> 00:28:17,640
is from ads that you launched in this
1017
00:28:17,640 --> 00:28:19,279
month or in the previous month, I mean,
1018
00:28:19,599 --> 00:28:21,000
there's some caveats to this, but you
1019
00:28:21,000 --> 00:28:22,720
might also be concerned that like, okay,
1020
00:28:22,839 --> 00:28:25,200
you're really over, over, over indexed on
1021
00:28:25,200 --> 00:28:26,440
stuff you've launched recently.
1022
00:28:26,519 --> 00:28:28,200
Uh, and you might want to continue to get
1023
00:28:28,200 --> 00:28:29,319
new things in the pipeline because you
1024
00:28:29,319 --> 00:28:31,599
don't want to become overly reliant on
1025
00:28:31,799 --> 00:28:33,960
just a couple of, uh, outlier ads.
1026
00:28:34,119 --> 00:28:35,559
I mean, all our ads are good, but at the
1027
00:28:35,559 --> 00:28:36,599
same time, you don't want to become
1028
00:28:36,640 --> 00:28:37,920
overly reliant on this.
1029
00:28:38,039 --> 00:28:39,799
Uh, and so this is just a view that we've
1030
00:28:39,799 --> 00:28:41,000
been looking at to make sure, okay.
1031
00:28:41,279 --> 00:28:42,839
Um, are we, are we
1032
00:28:43,079 --> 00:28:44,559
launching new ads successfully?
1033
00:28:44,839 --> 00:28:45,960
Um, are there any pockets
1034
00:28:45,960 --> 00:28:47,079
that we can look into, right?
1035
00:28:47,079 --> 00:28:49,680
Like may, this is the biggest, this looks
1036
00:28:49,680 --> 00:28:51,440
like one of the biggest months, um,
1037
00:28:51,480 --> 00:28:54,160
where, uh, in the same month that those
1038
00:28:54,160 --> 00:28:55,920
ads are getting to spend almost 20% and
1039
00:28:55,920 --> 00:28:56,799
it's not even over yet.
1040
00:28:56,960 --> 00:28:58,559
So like, did we do something recently
1041
00:28:58,559 --> 00:28:59,839
that's hitting really well that we should
1042
00:28:59,839 --> 00:29:00,720
be leaning more into?
1043
00:29:01,000 --> 00:29:02,079
Um, so just kind of sparks that.
1044
00:29:02,319 --> 00:29:03,839
And if you wanted to run this, like
1045
00:29:03,839 --> 00:29:05,079
motions, not sponsoring this, but I can
1046
00:29:05,079 --> 00:29:05,920
give you the prompt to go.
1047
00:29:06,000 --> 00:29:07,759
You can just drop this in motion and, uh,
1048
00:29:07,920 --> 00:29:09,400
in their little runneth chat and it'll
1049
00:29:09,400 --> 00:29:10,319
produce this for you.
1050
00:29:10,319 --> 00:29:12,000
Um, HTML graph and everything.
1051
00:29:12,279 --> 00:29:14,799
Um, just kind of a cool way of, um,
1052
00:29:14,839 --> 00:29:15,920
looking at this and breaking it down.
1053
00:29:17,160 --> 00:29:19,559
So basically what you're saying is
1054
00:29:19,559 --> 00:29:21,799
instead of doing cohort or creative sort
1055
00:29:21,799 --> 00:29:24,000
of like by batches or
1056
00:29:24,000 --> 00:29:25,519
like, Hey, this was this batch.
1057
00:29:26,160 --> 00:29:28,200
You're doing it by a monthly cohort.
1058
00:29:28,720 --> 00:29:30,680
And that's how you're looking at them and
1059
00:29:30,680 --> 00:29:33,039
categorizing them in order to make sense
1060
00:29:33,039 --> 00:29:35,680
of the performance by each
1061
00:29:35,680 --> 00:29:37,039
of those cohorts by month.
1062
00:29:37,079 --> 00:29:39,240
Like it's, it's a more linear package.
1063
00:29:39,480 --> 00:29:42,000
Makes it easier to understand what we
1064
00:29:42,000 --> 00:29:43,240
did, what we shipped in
1065
00:29:43,240 --> 00:29:45,440
February is still good.
1066
00:29:45,920 --> 00:29:47,759
And what did we, and then you were able
1067
00:29:47,759 --> 00:29:49,559
to dig in and look at February and say,
1068
00:29:50,039 --> 00:29:52,599
these are the things that gripped that
1069
00:29:52,599 --> 00:29:53,759
are helped to scale.
1070
00:29:54,039 --> 00:29:55,160
How do we go back and
1071
00:29:55,160 --> 00:29:56,640
replicate some of the unlocks we had?
1072
00:29:56,680 --> 00:29:57,799
Is that, that's essentially what you're
1073
00:29:57,799 --> 00:29:58,440
getting at with this.
1074
00:29:58,599 --> 00:29:59,519
Yeah, exactly.
1075
00:29:59,519 --> 00:30:01,000
And there's, there's, this is, this is a
1076
00:30:01,000 --> 00:30:02,079
view in a certain way
1077
00:30:02,079 --> 00:30:02,960
that you can look at it.
1078
00:30:03,039 --> 00:30:04,759
Uh, if you're a motion user, you can go
1079
00:30:04,759 --> 00:30:06,400
to their new launches report and that'll
1080
00:30:06,400 --> 00:30:07,640
give you a weekly version of it.
1081
00:30:07,680 --> 00:30:09,400
It doesn't look exactly like this, but
1082
00:30:09,400 --> 00:30:10,799
it'll give you a weekly version where you
1083
00:30:10,799 --> 00:30:11,920
can expand and contract.
1084
00:30:12,200 --> 00:30:13,319
Uh, and what you want to do, I mean,
1085
00:30:13,319 --> 00:30:15,759
we've talked about, um, I feel like, I
1086
00:30:15,759 --> 00:30:17,200
feel like scalability school needs to
1087
00:30:17,200 --> 00:30:18,519
take a lot of credit for like spend
1088
00:30:18,519 --> 00:30:20,519
velocity as a, as something to lean into.
1089
00:30:20,559 --> 00:30:21,599
I know Zach, that was something that you
1090
00:30:21,599 --> 00:30:23,279
mentioned a ton last year, like spend
1091
00:30:23,279 --> 00:30:24,880
velocity and things picking up quickly.
1092
00:30:25,200 --> 00:30:27,400
But like, if you launched ads in the last
1093
00:30:27,400 --> 00:30:29,119
week that became the top spender, that's
1094
00:30:29,119 --> 00:30:30,759
the signal to lean into is like, okay,
1095
00:30:30,960 --> 00:30:32,039
um, did we hit on a
1096
00:30:32,039 --> 00:30:33,200
seasonal moment that's relevant?
1097
00:30:33,240 --> 00:30:34,559
Or is there something lasting in here
1098
00:30:34,559 --> 00:30:37,200
that we can go recreate, um, and lean
1099
00:30:37,200 --> 00:30:38,240
into and do more of.
1100
00:30:38,400 --> 00:30:39,279
And this is just another
1101
00:30:39,279 --> 00:30:40,440
way of, of looking at that.
1102
00:30:40,480 --> 00:30:41,720
So there's the weekly view of it.
1103
00:30:42,000 --> 00:30:43,319
Um, but it's, it's basically just like
1104
00:30:43,319 --> 00:30:44,599
forcing you to, to reconcile.
1105
00:30:44,599 --> 00:30:45,880
If you're, if you're a brand owner and
1106
00:30:45,880 --> 00:30:47,119
you are hiring create, like you, let's
1107
00:30:47,119 --> 00:30:48,519
say you have your brand owner and you
1108
00:30:48,519 --> 00:30:50,319
have a creative agency, uh, and you hired
1109
00:30:50,319 --> 00:30:52,960
them in March and you pull this report
1110
00:30:53,000 --> 00:30:54,559
and all of the spend is going through
1111
00:30:54,559 --> 00:30:55,319
shit you've created.
1112
00:30:55,559 --> 00:30:56,880
Like that's, you know, there's obviously
1113
00:30:56,880 --> 00:30:57,920
different ways of looking at it, but
1114
00:30:57,920 --> 00:30:59,200
like, that's the, that's the question to
1115
00:30:59,200 --> 00:31:01,039
go ask and say, it's like, Hey, is
1116
00:31:01,039 --> 00:31:02,720
anything you're making actually, um,
1117
00:31:02,720 --> 00:31:05,599
producing like net new net new spend in
1118
00:31:05,599 --> 00:31:08,400
the, in the, um, this one has seasonal,
1119
00:31:08,519 --> 00:31:09,519
um, seasonal swings.
1120
00:31:09,519 --> 00:31:10,240
And I feel like that's why this is
1121
00:31:10,240 --> 00:31:11,440
changing so much, especially recently.
1122
00:31:11,799 --> 00:31:12,240
I love it.
1123
00:31:12,640 --> 00:31:14,400
I don't have any questions on this.
1124
00:31:14,440 --> 00:31:15,720
Zach, do you have questions on this?
1125
00:31:15,759 --> 00:31:16,920
No, it makes a ton of sense.
1126
00:31:16,960 --> 00:31:20,279
Um, we do this kind of like end of month
1127
00:31:20,279 --> 00:31:23,400
report every month that, um, our like
1128
00:31:23,480 --> 00:31:25,480
creative ops manager and our VP of
1129
00:31:25,480 --> 00:31:26,519
performance creative put together.
1130
00:31:26,559 --> 00:31:28,200
It's similar to this and then just goes a
1131
00:31:28,200 --> 00:31:30,400
bit deeper where we see the difference
1132
00:31:30,400 --> 00:31:31,640
between what we call like breakthrough
1133
00:31:31,640 --> 00:31:33,319
ads versus scaled ads.
1134
00:31:33,400 --> 00:31:35,440
Um, which breakthrough is like, does it
1135
00:31:35,440 --> 00:31:37,400
hit $2,000 in spend within a seven day
1136
00:31:37,400 --> 00:31:40,640
period at the North beam
1137
00:31:40,640 --> 00:31:41,680
when they click target goal.
1138
00:31:42,200 --> 00:31:43,400
And so we kind of look at month over
1139
00:31:43,400 --> 00:31:44,599
month over month where we're saying like,
1140
00:31:44,599 --> 00:31:46,640
okay, how, how are we performing?
1141
00:31:46,680 --> 00:31:48,839
How many ads are actually breaking
1142
00:31:48,839 --> 00:31:50,119
through or scaling this
1143
00:31:50,119 --> 00:31:52,200
month versus previous months?
1144
00:31:52,200 --> 00:31:54,200
And it's, it's somewhat, it's almost like
1145
00:31:54,200 --> 00:31:55,279
a, maybe one step
1146
00:31:55,279 --> 00:31:56,519
deeper, maybe one step above.
1147
00:31:56,519 --> 00:31:57,519
I don't really know like the best way to
1148
00:31:57,759 --> 00:31:59,319
describe that, but yeah, I
1149
00:31:59,319 --> 00:32:01,480
mean, this is super important.
1150
00:32:01,480 --> 00:32:02,599
It's like, what do we miss?
1151
00:32:02,640 --> 00:32:03,920
And like, what is the longevity?
1152
00:32:03,920 --> 00:32:04,640
It'd be really interesting to
1153
00:32:04,640 --> 00:32:07,480
look at it, um, over time too.
1154
00:32:07,519 --> 00:32:08,720
So you're like, this is what it looks
1155
00:32:08,720 --> 00:32:09,839
like now, but then what does
1156
00:32:09,839 --> 00:32:11,480
it look like going forward?
1157
00:32:11,519 --> 00:32:11,799
I don't know.
1158
00:32:12,160 --> 00:32:13,160
I don't know if I'm like explaining that
1159
00:32:13,160 --> 00:32:15,119
right either, but the progressive like
1160
00:32:15,119 --> 00:32:17,519
swing, like does the, does each month end
1161
00:32:17,519 --> 00:32:18,799
up turning into 10% of a spend?
1162
00:32:18,799 --> 00:32:19,960
Or does it actually eat up
1163
00:32:19,960 --> 00:32:22,039
like, ideally, is it a quarter?
1164
00:32:22,039 --> 00:32:23,559
Is it 25% of spend for the first month
1165
00:32:23,559 --> 00:32:25,519
and then kind of degradates down to like
1166
00:32:25,519 --> 00:32:26,680
10 and then holds at 10?
1167
00:32:27,240 --> 00:32:28,599
That'd also be interesting too, to see.
1168
00:32:29,359 --> 00:32:30,440
Yeah, that's a great point.
1169
00:32:30,519 --> 00:32:32,200
I mean, the reason that like the, the
1170
00:32:32,200 --> 00:32:33,519
thing that inspired me to do this in the
1171
00:32:33,519 --> 00:32:34,839
first place is we had a brand who was
1172
00:32:34,839 --> 00:32:37,240
like, they were like generally bought
1173
00:32:37,240 --> 00:32:38,000
into the idea of
1174
00:32:38,000 --> 00:32:39,400
creative volume is important.
1175
00:32:39,640 --> 00:32:42,240
Um, but they wanted to like, see the
1176
00:32:42,240 --> 00:32:43,319
proof that it was working.
1177
00:32:43,599 --> 00:32:45,599
Um, and so they started, they, they
1178
00:32:45,599 --> 00:32:47,960
probably quadrupled their, their output,
1179
00:32:48,279 --> 00:32:50,039
um, through a bunch of like very
1180
00:32:50,039 --> 00:32:51,480
thoughtful adjustments on the way that
1181
00:32:51,480 --> 00:32:52,680
their internal team operated.
1182
00:32:52,799 --> 00:32:54,680
So they took inspiration from kind of
1183
00:32:54,680 --> 00:32:56,000
like ideas that we were sharing with them
1184
00:32:56,200 --> 00:32:57,079
and then they converted
1185
00:32:57,079 --> 00:32:58,079
it into actual output.
1186
00:32:58,319 --> 00:32:59,279
And it was just important for them to
1187
00:32:59,279 --> 00:33:01,400
make sure like, Hey, did what we, we, we
1188
00:33:01,400 --> 00:33:03,920
produced, uh, a hundred ads this month.
1189
00:33:03,960 --> 00:33:06,319
And you know, that's up from the 20 we
1190
00:33:06,319 --> 00:33:07,480
were previously locked into.
1191
00:33:07,599 --> 00:33:08,559
Did those do anything?
1192
00:33:08,680 --> 00:33:09,880
And so it allows us to look
1193
00:33:09,880 --> 00:33:11,440
back and say, yes, they worked.
1194
00:33:11,440 --> 00:33:13,559
But here's another way to split it.
1195
00:33:13,559 --> 00:33:15,160
It's like 50 of the ads that you gave us
1196
00:33:15,160 --> 00:33:16,480
for, were for a promo
1197
00:33:16,640 --> 00:33:18,119
that lasted three days.
1198
00:33:18,279 --> 00:33:20,480
And so like, that's, that's useful, but
1199
00:33:20,480 --> 00:33:22,200
like, and it makes sense because promo,
1200
00:33:22,359 --> 00:33:24,279
promo spends harder, uh, and over a
1201
00:33:24,279 --> 00:33:25,880
shorter period of time, but you want to
1202
00:33:25,880 --> 00:33:27,279
make sure that those evergreen ads that
1203
00:33:27,279 --> 00:33:28,799
are still live in the ad account continue
1204
00:33:28,799 --> 00:33:29,960
to actually produce.
1205
00:33:30,079 --> 00:33:30,960
Because then, you know, you've actually
1206
00:33:30,960 --> 00:33:32,400
hit something and it's not just like a
1207
00:33:32,680 --> 00:33:33,839
one hit wonder promo ad that
1208
00:33:33,839 --> 00:33:34,920
you can actually repurpose.
1209
00:33:35,160 --> 00:33:36,440
And so it was just like validation for
1210
00:33:36,440 --> 00:33:37,400
them that what they
1211
00:33:37,400 --> 00:33:38,880
were making was working.
1212
00:33:39,119 --> 00:33:40,759
And when you split it down to evergreen
1213
00:33:40,759 --> 00:33:42,799
and promo ads, it still ended up being
1214
00:33:42,799 --> 00:33:44,400
the case that evergreen consistently is
1215
00:33:44,400 --> 00:33:45,480
stacking up month over month.
1216
00:33:45,720 --> 00:33:46,680
And that's what compounds, right?
1217
00:33:46,680 --> 00:33:48,759
It's like, if you can, if you can make
1218
00:33:48,799 --> 00:33:50,400
more ads every single month that earns
1219
00:33:50,400 --> 00:33:52,440
spend in the ad account, then that's what
1220
00:33:52,440 --> 00:33:53,400
compounds over time.
1221
00:33:54,200 --> 00:33:56,000
And allows you to increase spend, uh,
1222
00:33:56,279 --> 00:33:57,240
over the, over the year.
1223
00:33:57,319 --> 00:33:59,160
Talk about, there's this guy, Zach Stuck,
1224
00:33:59,160 --> 00:34:00,599
who has this brand called Marsman.
1225
00:34:00,960 --> 00:34:04,920
Um, and it says, we're going to go into
1226
00:34:04,920 --> 00:34:06,480
now how Marsman scales on
1227
00:34:06,480 --> 00:34:08,360
next month's repeat revenue.
1228
00:34:09,719 --> 00:34:11,320
Uh, which is kind of a hot topic.
1229
00:34:11,960 --> 00:34:13,440
I feel like everybody wants to know.
1230
00:34:14,159 --> 00:34:15,599
I mean, headline, I mean,
1231
00:34:15,639 --> 00:34:16,840
Zach, I want to go into this.
1232
00:34:17,360 --> 00:34:17,719
Sure.
1233
00:34:17,760 --> 00:34:18,119
Sure.
1234
00:34:18,199 --> 00:34:19,159
I mean, Brad, do you want
1235
00:34:19,159 --> 00:34:20,239
to like tee this up anymore?
1236
00:34:20,239 --> 00:34:21,039
Do you, is that, is
1237
00:34:21,039 --> 00:34:21,880
that enough for me to.
1238
00:34:22,679 --> 00:34:22,920
Yeah.
1239
00:34:22,960 --> 00:34:25,119
I mean, the, the, the frame, the, I
1240
00:34:25,119 --> 00:34:25,719
remember we were having
1241
00:34:25,719 --> 00:34:26,760
a conversation one time.
1242
00:34:26,800 --> 00:34:28,039
And one of the things that you mentioned
1243
00:34:28,039 --> 00:34:29,320
is like, one of the things that we think
1244
00:34:29,320 --> 00:34:32,119
about for pushing spend aggressively.
1245
00:34:32,159 --> 00:34:33,480
Now you guys are, you guys have hit an
1246
00:34:33,480 --> 00:34:35,000
insane, um, growth rate
1247
00:34:35,000 --> 00:34:37,559
and, and, um, uh, MRR.
1248
00:34:37,559 --> 00:34:39,039
And so, and you're pushing
1249
00:34:39,039 --> 00:34:40,480
growth, growth aggressively.
1250
00:34:40,800 --> 00:34:43,320
And I look in brands that have solid
1251
00:34:43,320 --> 00:34:44,559
repeat rate all the time.
1252
00:34:44,800 --> 00:34:46,039
And maybe they have different
1253
00:34:46,039 --> 00:34:46,920
profitability goals.
1254
00:34:46,920 --> 00:34:47,320
They have different
1255
00:34:47,320 --> 00:34:48,119
growth goals, whatever.
1256
00:34:48,239 --> 00:34:51,000
Like often it's uncommon that the brand
1257
00:34:51,000 --> 00:34:51,440
is not being
1258
00:34:51,440 --> 00:34:52,679
aggressive enough in scaling.
1259
00:34:52,719 --> 00:34:54,440
Cause they haven't done the financial,
1260
00:34:54,440 --> 00:34:55,800
uh, homework that you guys have done to
1261
00:34:55,800 --> 00:34:56,440
understand, like, how
1262
00:34:56,440 --> 00:34:57,639
much can we spend into this?
1263
00:34:57,920 --> 00:34:58,920
Um, and one of the things that you
1264
00:34:58,920 --> 00:35:00,800
mentioned that you use is like a guide
1265
00:35:00,800 --> 00:35:02,199
for this is like, well, okay, if we know,
1266
00:35:02,559 --> 00:35:04,400
and I might not be framing this perfectly
1267
00:35:04,400 --> 00:35:05,960
exactly the way you phrased it, but, um,
1268
00:35:06,000 --> 00:35:07,599
you said, Hey, if we have this, an idea
1269
00:35:07,599 --> 00:35:09,159
that our EP customer revenue next month
1270
00:35:09,159 --> 00:35:09,800
is going to be, I'm just
1271
00:35:09,800 --> 00:35:10,440
going to pick a number.
1272
00:35:10,519 --> 00:35:11,840
It's going to be 500 grand.
1273
00:35:11,880 --> 00:35:12,559
I'm probably selling
1274
00:35:12,559 --> 00:35:13,559
that very short for Mars.
1275
00:35:13,920 --> 00:35:15,079
Like it's gonna be 500 grand.
1276
00:35:15,079 --> 00:35:16,199
It's like, well, we know we're
1277
00:35:16,199 --> 00:35:17,599
gonna have 500 grand in cash.
1278
00:35:17,840 --> 00:35:19,880
We can spend into that, like at least
1279
00:35:19,880 --> 00:35:20,880
that much next month.
1280
00:35:21,079 --> 00:35:22,119
Um, because we want to
1281
00:35:22,199 --> 00:35:23,199
be, want to be aggressive.
1282
00:35:23,440 --> 00:35:24,360
Um, so I'm sure there's a lot more
1283
00:35:24,360 --> 00:35:25,599
context and things that go into making
1284
00:35:25,599 --> 00:35:27,599
the decision, but it's, it's, it's, it's.
1285
00:35:27,639 --> 00:35:29,519
It's at least, uh, maybe safe.
1286
00:35:29,559 --> 00:35:31,039
Cause you have the cash to be able to
1287
00:35:31,039 --> 00:35:32,760
spend into, uh, into that.
1288
00:35:32,760 --> 00:35:33,639
So that's the framing.
1289
00:35:34,519 --> 00:35:34,760
Yeah.
1290
00:35:34,760 --> 00:35:36,280
So at the end of the day, like this is
1291
00:35:36,280 --> 00:35:38,119
where these MRR businesses or
1292
00:35:38,119 --> 00:35:40,280
subscription businesses scale so quickly.
1293
00:35:40,440 --> 00:35:42,480
It just comes down to this simple model,
1294
00:35:42,480 --> 00:35:45,119
which is if I can forecast for this
1295
00:35:45,119 --> 00:35:47,280
example, 500 K in reoccurring revenue.
1296
00:35:47,480 --> 00:35:50,559
And I know that my OPEX is going to be a
1297
00:35:50,559 --> 00:35:51,840
hundred thousand for the month.
1298
00:35:52,320 --> 00:35:54,960
Um, I know that I need to maybe run at
1299
00:35:54,960 --> 00:35:57,400
like a 1.8 or two, two, let's just call
1300
00:35:57,400 --> 00:36:00,440
it a two, uh, MER, uh, or two blended
1301
00:36:00,440 --> 00:36:01,639
rows, whatever term you want to use.
1302
00:36:02,119 --> 00:36:04,639
Um, to still maintain
1303
00:36:04,679 --> 00:36:06,679
profitability with a hundred K OPEX.
1304
00:36:07,400 --> 00:36:09,239
I then know I can spend
1305
00:36:09,280 --> 00:36:11,239
that full MRR into it.
1306
00:36:11,320 --> 00:36:13,039
And as long as I know if I'm trying to
1307
00:36:13,039 --> 00:36:14,639
achieve a two MER, I have to
1308
00:36:14,639 --> 00:36:16,639
be at a one new customer ROAS.
1309
00:36:17,079 --> 00:36:18,400
And so if I can stay at a one new
1310
00:36:18,440 --> 00:36:19,880
customer ROAS, I can still hit my target
1311
00:36:19,880 --> 00:36:20,679
and I can keep going.
1312
00:36:21,079 --> 00:36:22,719
The nice thing about this is like with
1313
00:36:22,719 --> 00:36:24,320
time and with more scale
1314
00:36:24,400 --> 00:36:26,719
that MER can actually go down.
1315
00:36:27,199 --> 00:36:29,000
Um, as long as your OPEX doesn't offset
1316
00:36:29,000 --> 00:36:30,960
it too much, meaning that like, let's
1317
00:36:30,960 --> 00:36:32,639
just like ratched up those numbers a ton.
1318
00:36:32,920 --> 00:36:35,760
So let's say, um, we're, our MRR is 5
1319
00:36:35,800 --> 00:36:39,039
million a month now, for example, and we
1320
00:36:39,079 --> 00:36:39,960
know what our OPEX is.
1321
00:36:39,960 --> 00:36:42,079
Maybe our OPEX is a million a month now.
1322
00:36:42,079 --> 00:36:43,119
I don't know, for example.
1323
00:36:43,440 --> 00:36:47,079
And so with that, because we have such
1324
00:36:47,079 --> 00:36:49,079
good margin and we know we're going to do
1325
00:36:49,079 --> 00:36:50,960
5 million in returning, we might be able
1326
00:36:50,960 --> 00:36:51,840
to say, Hey, we can run
1327
00:36:51,840 --> 00:36:53,880
a 1.5 MER now in total.
1328
00:36:54,000 --> 00:36:56,360
So I can still spend that full 5 million
1329
00:36:56,360 --> 00:36:58,880
in revenue and do only 2.5 million in
1330
00:36:58,880 --> 00:36:59,960
revenue to average
1331
00:36:59,960 --> 00:37:01,440
out to a 1.5 total MER.
1332
00:37:01,880 --> 00:37:04,760
So again, like this is where with more
1333
00:37:04,760 --> 00:37:06,800
scale plus the returning customer
1334
00:37:06,800 --> 00:37:09,440
revenue, when you can forecast, um, that
1335
00:37:09,440 --> 00:37:11,239
returning it like just, that's where
1336
00:37:11,239 --> 00:37:12,519
things really start to compound.
1337
00:37:12,880 --> 00:37:14,760
I think the thing that's been really
1338
00:37:14,760 --> 00:37:17,840
interesting for us is, is candidly like,
1339
00:37:19,360 --> 00:37:20,760
there's so many things that you do in
1340
00:37:20,960 --> 00:37:22,199
these subscription businesses that will
1341
00:37:22,199 --> 00:37:25,119
have like a, uh, fundamental changes on
1342
00:37:25,119 --> 00:37:26,719
like forecasting that returning customer
1343
00:37:26,719 --> 00:37:28,400
revenue that could have big impacts on
1344
00:37:28,400 --> 00:37:29,599
like how you think about the business.
1345
00:37:29,639 --> 00:37:31,519
So let's say you run a sale at the end of
1346
00:37:31,519 --> 00:37:34,079
a month, one month, and your cat goes
1347
00:37:34,079 --> 00:37:35,280
down and you scale into it.
1348
00:37:35,760 --> 00:37:37,159
The next month you could assume,
1349
00:37:37,199 --> 00:37:39,000
depending on how many percentage of those
1350
00:37:39,000 --> 00:37:40,840
customers are 30 day customers versus 90
1351
00:37:40,920 --> 00:37:41,880
day customers, you might see
1352
00:37:41,880 --> 00:37:43,639
a big spike at the end of it.
1353
00:37:43,800 --> 00:37:45,599
But let's say like you run an offer where
1354
00:37:45,599 --> 00:37:47,400
your 90 day subscription versus 30 day
1355
00:37:47,400 --> 00:37:49,039
subscription is a bigger discount.
1356
00:37:49,039 --> 00:37:50,920
So out of nowhere, that becomes a larger
1357
00:37:50,920 --> 00:37:51,559
percentage of your
1358
00:37:51,559 --> 00:37:52,639
customers at the end of that month.
1359
00:37:53,119 --> 00:37:54,800
You now are deferring all that revenue
1360
00:37:54,800 --> 00:37:57,039
for 90 days on the back of the month
1361
00:37:57,280 --> 00:37:59,039
where then you're like, Oh shit, I better
1362
00:37:59,039 --> 00:38:00,239
plan my spend
1363
00:38:00,280 --> 00:38:02,159
accordingly around those pockets.
1364
00:38:02,519 --> 00:38:04,119
Also, like if you're testing different
1365
00:38:04,119 --> 00:38:05,280
bundles, you're testing all sorts of
1366
00:38:05,280 --> 00:38:06,280
stuff, like this can all
1367
00:38:06,280 --> 00:38:07,440
have huge impacts on it.
1368
00:38:07,440 --> 00:38:10,159
So I think like as much as reoccurring
1369
00:38:10,159 --> 00:38:11,639
revenue is like a blessing, it can be a
1370
00:38:11,639 --> 00:38:13,159
curse too, because there's so many
1371
00:38:13,159 --> 00:38:14,079
different variables.
1372
00:38:14,119 --> 00:38:15,440
If you are really doing a bunch of
1373
00:38:15,440 --> 00:38:16,880
testing, like we have been with Mars,
1374
00:38:17,159 --> 00:38:18,440
where we're trying like different offers,
1375
00:38:18,440 --> 00:38:20,719
different price points, you know, 30 day,
1376
00:38:20,760 --> 00:38:23,880
90 day, you know, 180 day subscriptions
1377
00:38:23,880 --> 00:38:24,559
to kind of see like,
1378
00:38:24,599 --> 00:38:25,719
what is the sweet spot?
1379
00:38:26,199 --> 00:38:27,840
All of those changes will then have an
1380
00:38:27,880 --> 00:38:30,599
impact down 30, 60, 90, 180 days from now
1381
00:38:30,599 --> 00:38:31,519
that you really have to
1382
00:38:31,519 --> 00:38:32,559
like, be careful with.
1383
00:38:33,119 --> 00:38:35,000
And there's a lot of like assumptions
1384
00:38:35,000 --> 00:38:35,880
that go into the model
1385
00:38:35,880 --> 00:38:37,440
until you have like longevity.
1386
00:38:37,440 --> 00:38:39,760
So, I mean, the smartest thing you can
1387
00:38:39,760 --> 00:38:41,039
really do is like, just be like, cool,
1388
00:38:41,039 --> 00:38:41,519
we're going to sell a
1389
00:38:41,559 --> 00:38:42,599
30 day subscription.
1390
00:38:42,599 --> 00:38:43,599
That's all we're going to sell.
1391
00:38:43,599 --> 00:38:45,159
We're going to scale into that because
1392
00:38:45,199 --> 00:38:45,960
then there's no other
1393
00:38:46,000 --> 00:38:47,639
crazy nuance that happens.
1394
00:38:48,360 --> 00:38:49,840
The second you had a discount, additional
1395
00:38:49,840 --> 00:38:51,239
discount or whatever, like that could
1396
00:38:51,239 --> 00:38:53,280
actually impact your cohorts, which then
1397
00:38:53,280 --> 00:38:55,239
resets your, all of your learnings on
1398
00:38:55,239 --> 00:38:55,880
what's expected for
1399
00:38:55,880 --> 00:38:56,880
returning customer revenue.
1400
00:38:56,920 --> 00:38:59,719
But the simple math on it is when you
1401
00:38:59,719 --> 00:39:00,960
know what your MRR is going to be and you
1402
00:39:00,960 --> 00:39:02,639
have a good, good model built around it,
1403
00:39:02,639 --> 00:39:03,400
like you can usually
1404
00:39:03,400 --> 00:39:04,639
spend into that fully.
1405
00:39:04,920 --> 00:39:06,280
And most of these brands that are, that
1406
00:39:06,280 --> 00:39:08,719
are scaling the greens of the world, uh,
1407
00:39:09,119 --> 00:39:11,199
IM8s, whatever, like they're looking at
1408
00:39:11,199 --> 00:39:12,360
the returning customer revenue and
1409
00:39:12,360 --> 00:39:13,519
saying, Hey, if we're going to do 5
1410
00:39:13,599 --> 00:39:14,880
million this month in MRR, we're
1411
00:39:14,920 --> 00:39:15,519
definitely going to spend
1412
00:39:15,519 --> 00:39:16,760
at least 5 million in ads.
1413
00:39:18,039 --> 00:39:18,360
Yeah.
1414
00:39:18,559 --> 00:39:19,840
And the, in the issue, the point that you
1415
00:39:19,840 --> 00:39:21,280
made that's, you have to be careful with
1416
00:39:21,280 --> 00:39:24,119
is when you, you add the 90 day as you've
1417
00:39:24,159 --> 00:39:26,599
maybe, maybe you rip a bunch of 90 day.
1418
00:39:26,760 --> 00:39:27,719
And all of a sudden you went from a
1419
00:39:27,719 --> 00:39:29,599
hundred percent is 30 days, a hundred
1420
00:39:29,599 --> 00:39:30,840
percent of your subscriptions are 30 days
1421
00:39:30,840 --> 00:39:33,360
to 50% of them dramatically are 90 days.
1422
00:39:33,719 --> 00:39:35,039
Well, if you're, if you don't change your
1423
00:39:35,039 --> 00:39:36,840
cohort math and your four year cashflow
1424
00:39:36,840 --> 00:39:38,800
forecast to account for that, you're
1425
00:39:38,800 --> 00:39:41,039
going to spend into all of that, but
1426
00:39:41,039 --> 00:39:41,960
you're not, you're actually not going to
1427
00:39:41,960 --> 00:39:42,880
realize that cash for
1428
00:39:42,880 --> 00:39:44,320
another 60 days beyond that.
1429
00:39:44,519 --> 00:39:45,880
And that's where you put yourself into a
1430
00:39:45,920 --> 00:39:46,880
funky cash position.
1431
00:39:46,880 --> 00:39:47,960
So that's what you're saying about
1432
00:39:47,960 --> 00:39:49,079
keeping it simple to avoid.
1433
00:39:49,079 --> 00:39:50,760
The cash positions, plus then like
1434
00:39:50,800 --> 00:39:52,760
running different discounts throughout
1435
00:39:52,760 --> 00:39:54,440
different times of the month too.
1436
00:39:54,480 --> 00:39:56,039
So like maybe you run a discount, like
1437
00:39:56,039 --> 00:39:58,079
4th of July is in the front half of July.
1438
00:39:58,440 --> 00:40:00,360
Father's day is on like the back half of
1439
00:40:00,360 --> 00:40:01,400
the month or like mother's
1440
00:40:01,400 --> 00:40:02,119
days in the back of the month.
1441
00:40:02,280 --> 00:40:03,800
So like, depending on when you're running
1442
00:40:03,800 --> 00:40:05,440
those, those sales, it's also going to
1443
00:40:05,440 --> 00:40:06,960
have fluctuations on the returning, which
1444
00:40:06,960 --> 00:40:09,360
then ideally you're looking at, okay, I
1445
00:40:09,360 --> 00:40:10,960
did, you know, 500 K of MRR.
1446
00:40:10,960 --> 00:40:11,960
I can spend 500 K.
1447
00:40:12,000 --> 00:40:13,639
You have to really be cautious of when
1448
00:40:13,639 --> 00:40:15,840
you're spending into that 500 K around
1449
00:40:15,840 --> 00:40:17,119
all of those moments that
1450
00:40:17,119 --> 00:40:18,079
you've made historically.
1451
00:40:18,079 --> 00:40:19,480
Otherwise you can overspend in the front
1452
00:40:19,480 --> 00:40:20,960
half of the month and then you don't have
1453
00:40:20,960 --> 00:40:22,719
enough budget on the back half when CAC
1454
00:40:22,719 --> 00:40:23,480
is actually really good.
1455
00:40:25,239 --> 00:40:27,239
Can we talk about the forecasting and how
1456
00:40:27,239 --> 00:40:28,920
many times you're, how
1457
00:40:28,920 --> 00:40:29,760
often you're reforecasting?
1458
00:40:29,760 --> 00:40:32,719
Like, is it built on subscription base,
1459
00:40:32,760 --> 00:40:34,679
repeat purchase cohort curve?
1460
00:40:34,880 --> 00:40:35,119
Yeah.
1461
00:40:35,159 --> 00:40:37,039
Like, is it a simple percentage?
1462
00:40:37,039 --> 00:40:37,559
Like what is it?
1463
00:40:37,840 --> 00:40:38,079
Yeah.
1464
00:40:38,119 --> 00:40:39,800
So we just rebuilt our entire model this
1465
00:40:39,800 --> 00:40:41,519
week because we had some issues candidly
1466
00:40:41,519 --> 00:40:44,880
for May of us like missing, but all
1467
00:40:44,880 --> 00:40:46,079
learning opportunities for us.
1468
00:40:46,079 --> 00:40:49,800
But we, we just rebuilt it where it
1469
00:40:49,800 --> 00:40:51,320
actually goes down to the individual
1470
00:40:51,320 --> 00:40:53,280
customer level and then is now starting
1471
00:40:53,280 --> 00:40:55,480
to pull in all sorts of moments around
1472
00:40:55,480 --> 00:40:56,320
that individual customer.
1473
00:40:56,320 --> 00:40:58,719
So we basically look at who that customer
1474
00:40:58,719 --> 00:41:00,719
is, what they paid that first day, if
1475
00:41:00,719 --> 00:41:02,039
there was a discount or not.
1476
00:41:02,199 --> 00:41:04,239
We factor that into the retention curve
1477
00:41:04,239 --> 00:41:05,880
of that individual based on us knowing,
1478
00:41:05,880 --> 00:41:07,920
Hey, sometimes when they use a discount,
1479
00:41:07,920 --> 00:41:09,320
they're much more of like a, you know,
1480
00:41:09,320 --> 00:41:11,440
they're not as the longevity of that
1481
00:41:11,440 --> 00:41:12,800
customer is less because they're more of
1482
00:41:12,800 --> 00:41:13,679
a discount shopper and
1483
00:41:13,679 --> 00:41:14,679
they're more likely to cancel.
1484
00:41:14,920 --> 00:41:16,800
So we're like, we're applying retention
1485
00:41:16,800 --> 00:41:18,519
curves based on the decisions they made
1486
00:41:18,519 --> 00:41:20,239
and the sales price point or whatever
1487
00:41:20,239 --> 00:41:21,159
they did on the front end.
1488
00:41:21,480 --> 00:41:22,440
Plus then we're starting
1489
00:41:22,440 --> 00:41:23,440
to pull in other things.
1490
00:41:23,440 --> 00:41:25,880
Like we're pulling in, we're going to
1491
00:41:25,880 --> 00:41:27,280
start planning things from like gorgeous
1492
00:41:27,280 --> 00:41:29,360
or a CX team to say like, did this
1493
00:41:29,360 --> 00:41:30,360
customer reach out
1494
00:41:30,400 --> 00:41:32,119
through our portal to cancel?
1495
00:41:32,119 --> 00:41:33,360
But then they stayed on because we gave
1496
00:41:33,360 --> 00:41:34,239
them an additional discount.
1497
00:41:34,440 --> 00:41:35,280
What does that mean for the
1498
00:41:35,280 --> 00:41:36,639
longevity of that customer?
1499
00:41:36,840 --> 00:41:38,280
So now we're getting down to the per
1500
00:41:38,280 --> 00:41:40,679
customer basis where candidly, like prior
1501
00:41:40,679 --> 00:41:41,480
to that, we were looking
1502
00:41:41,480 --> 00:41:42,400
at it more holistically.
1503
00:41:42,400 --> 00:41:43,719
We were saying, okay, how many customers
1504
00:41:43,719 --> 00:41:45,760
bought on this day, 30 days ago, look at
1505
00:41:45,760 --> 00:41:47,320
that monthly core, look at the churn for
1506
00:41:47,320 --> 00:41:48,280
that cohort and just kind
1507
00:41:48,280 --> 00:41:49,559
of make a lot of assumptions.
1508
00:41:49,599 --> 00:41:51,320
But at the scale that we're getting to
1509
00:41:51,320 --> 00:41:52,679
now, like we had to look at it as
1510
00:41:52,719 --> 00:41:54,039
granular as we possibly could.
1511
00:41:54,960 --> 00:41:56,320
The thing that's really interesting now
1512
00:41:56,320 --> 00:41:57,960
is we're taking this forecast.
1513
00:41:57,960 --> 00:42:00,159
We're basically saying, okay, again,
1514
00:42:00,159 --> 00:42:01,440
simplified numbers, 5
1515
00:42:01,440 --> 00:42:02,679
million in MR this month.
1516
00:42:02,920 --> 00:42:04,719
We want to do 10 million in revenue.
1517
00:42:04,760 --> 00:42:07,519
We can run in a two blended ROAS, which
1518
00:42:07,519 --> 00:42:08,719
would mean that we'd have to run.
1519
00:42:09,559 --> 00:42:11,719
We'd have to, you know, run at a certain
1520
00:42:11,719 --> 00:42:13,360
degree where we would still maintain
1521
00:42:13,400 --> 00:42:14,559
profitability, officer
1522
00:42:14,559 --> 00:42:15,440
OPEX, all that stuff.
1523
00:42:16,199 --> 00:42:17,559
We're then saying, looking at our
1524
00:42:17,559 --> 00:42:19,239
forecast, we're then saying, okay, new
1525
00:42:19,239 --> 00:42:21,280
customer target is 5 million in revenue.
1526
00:42:21,519 --> 00:42:23,079
How do we go hit 5 million in revenue?
1527
00:42:23,199 --> 00:42:26,039
We're breaking down our targets by
1528
00:42:26,039 --> 00:42:27,760
channel and estimated orders
1529
00:42:27,760 --> 00:42:29,719
by channel every single day.
1530
00:42:29,760 --> 00:42:31,960
So now we're looking at like not just
1531
00:42:32,920 --> 00:42:35,639
orders or NCROAS or CPA
1532
00:42:35,639 --> 00:42:36,880
from Facebook or one day click.
1533
00:42:36,880 --> 00:42:38,599
Like we're literally looking at did
1534
00:42:38,599 --> 00:42:41,119
today, did email SMS drive the correct
1535
00:42:41,119 --> 00:42:42,039
amount of customers?
1536
00:42:42,239 --> 00:42:43,639
Did organic drive the right amount of
1537
00:42:43,639 --> 00:42:45,199
customers like by channel?
1538
00:42:45,440 --> 00:42:47,000
Because again, like at the scale that
1539
00:42:47,000 --> 00:42:49,039
we're getting to, if like one of those
1540
00:42:49,039 --> 00:42:50,559
channels misses, but everything else is
1541
00:42:50,559 --> 00:42:52,599
hitting, we could miss our target by, you
1542
00:42:52,599 --> 00:42:53,480
know, six figures,
1543
00:42:53,480 --> 00:42:54,920
potentially even seven figures a month.
1544
00:42:54,960 --> 00:42:57,440
So it's been really, it's been really
1545
00:42:57,440 --> 00:42:59,400
cool to actually like be able to start to
1546
00:42:59,400 --> 00:43:00,840
diagnose it that way and actually look at
1547
00:43:00,840 --> 00:43:03,039
it on a per day per channel basis.
1548
00:43:03,039 --> 00:43:04,800
Because I don't think many brands get to
1549
00:43:04,800 --> 00:43:06,039
the point where they're doing enough
1550
00:43:06,039 --> 00:43:07,679
volume where then it actually validates
1551
00:43:07,679 --> 00:43:09,039
like did email actually
1552
00:43:09,039 --> 00:43:10,719
drive 300 customers today?
1553
00:43:10,719 --> 00:43:11,400
Yes or no?
1554
00:43:12,719 --> 00:43:14,199
Because it's just like too granular, too
1555
00:43:14,199 --> 00:43:15,960
segmented based on size.
1556
00:43:15,960 --> 00:43:17,360
But now that the numbers have gotten big
1557
00:43:17,360 --> 00:43:18,719
enough that we can actually like keep an
1558
00:43:18,719 --> 00:43:19,559
eye on the trends,
1559
00:43:19,559 --> 00:43:20,320
like it's really helpful.
1560
00:43:20,800 --> 00:43:22,159
Also, then we can plan, hey, if we're
1561
00:43:22,159 --> 00:43:24,039
going to run a promotion, email can be
1562
00:43:24,039 --> 00:43:25,159
flat for the first half of the month.
1563
00:43:25,159 --> 00:43:26,400
But then during that sale period, we
1564
00:43:26,400 --> 00:43:27,760
expect it to actually pop up.
1565
00:43:27,760 --> 00:43:29,840
We also, you know, have the indication
1566
00:43:29,840 --> 00:43:31,400
that like, you know, new customer cac is
1567
00:43:31,400 --> 00:43:32,559
going to drop on these channels, but
1568
00:43:32,559 --> 00:43:34,400
maybe it doesn't move on podcasts because
1569
00:43:34,400 --> 00:43:35,239
that's linear throughout
1570
00:43:35,239 --> 00:43:36,679
the month, things like that.
1571
00:43:36,679 --> 00:43:40,239
So, yeah, I mean, forecasting is
1572
00:43:40,239 --> 00:43:41,320
important for all businesses, but
1573
00:43:41,320 --> 00:43:43,000
especially for subscription, like you
1574
00:43:43,000 --> 00:43:45,039
have to really dive into it.
1575
00:43:45,039 --> 00:43:47,960
Otherwise, like you will either not grow
1576
00:43:47,960 --> 00:43:49,800
as quickly as what you possibly can or
1577
00:43:50,880 --> 00:43:52,800
you'll actually inverse, which is like
1578
00:43:52,800 --> 00:43:54,199
you're not thinking enough of how many
1579
00:43:54,199 --> 00:43:55,920
net new subscribers you need to add and
1580
00:43:55,920 --> 00:43:57,320
what performance needs to be on every
1581
00:43:57,320 --> 00:43:59,119
single channel for you to actually add
1582
00:43:59,119 --> 00:43:59,920
net new subscribers.
1583
00:43:59,920 --> 00:44:01,280
Because for the most part, most of these
1584
00:44:01,280 --> 00:44:02,960
subscription brands are only growing
1585
00:44:03,320 --> 00:44:04,719
because they're adding net new because
1586
00:44:04,719 --> 00:44:06,440
likely most of them are churning the same
1587
00:44:06,440 --> 00:44:07,400
amount as they're adding in.
1588
00:44:07,960 --> 00:44:09,519
So everything above and beyond that is
1589
00:44:09,519 --> 00:44:11,519
like the real growth path.
1590
00:44:11,599 --> 00:44:13,599
And the podcast point is interesting
1591
00:44:13,599 --> 00:44:14,280
because it's like,
1592
00:44:14,400 --> 00:44:15,320
those are the things you do.
1593
00:44:15,480 --> 00:44:16,559
And like, I think a lot of people, it's
1594
00:44:16,559 --> 00:44:18,039
like maybe they put together a monthly
1595
00:44:18,039 --> 00:44:20,400
view, but the day to day can shift so
1596
00:44:20,400 --> 00:44:22,039
much to your point for a promo.
1597
00:44:22,599 --> 00:44:24,559
Even like week or day to day trends, like
1598
00:44:24,559 --> 00:44:25,880
maybe your weekends are more aggressive,
1599
00:44:25,880 --> 00:44:27,320
maybe they're slower, depending on the
1600
00:44:27,320 --> 00:44:28,119
category that you're in.
1601
00:44:28,679 --> 00:44:29,880
I have two follow questions on that.
1602
00:44:30,840 --> 00:44:33,199
When you get when you see that channel
1603
00:44:33,199 --> 00:44:34,559
level performance, like
1604
00:44:34,719 --> 00:44:35,880
maybe you can use two examples.
1605
00:44:35,880 --> 00:44:37,239
You could use a meta example
1606
00:44:37,280 --> 00:44:39,559
or the email example or both.
1607
00:44:39,639 --> 00:44:41,719
It's like if those are off, are you
1608
00:44:41,719 --> 00:44:43,960
reallocating in the platform?
1609
00:44:44,039 --> 00:44:44,840
Is it kind of like that's
1610
00:44:44,840 --> 00:44:45,559
where the deep dive starts?
1611
00:44:45,559 --> 00:44:46,559
Like, OK, meta is off.
1612
00:44:47,039 --> 00:44:47,840
OK, well, is there somewhere
1613
00:44:47,840 --> 00:44:49,079
in meta that we can reallocate?
1614
00:44:49,280 --> 00:44:50,599
And if it's like, nah, probably not.
1615
00:44:50,599 --> 00:44:51,639
This is all shit.
1616
00:44:51,880 --> 00:44:53,400
We got to go to somewhere else.
1617
00:44:53,400 --> 00:44:53,880
It's like, are you
1618
00:44:53,880 --> 00:44:55,480
making that daily decisions?
1619
00:44:55,480 --> 00:44:56,280
Are you zooming out over
1620
00:44:56,280 --> 00:44:57,000
a longer period of time?
1621
00:44:57,480 --> 00:44:59,400
And the mini follow to that is like, what
1622
00:44:59,400 --> 00:45:01,039
are you using as the, I
1623
00:45:01,039 --> 00:45:02,119
guess, like source of truth?
1624
00:45:02,119 --> 00:45:03,599
Are you using just like MTA
1625
00:45:03,880 --> 00:45:05,320
data or using platform data?
1626
00:45:05,320 --> 00:45:06,320
Just kind of normalized.
1627
00:45:06,480 --> 00:45:06,719
Yeah.
1628
00:45:06,719 --> 00:45:08,559
So for the CAC by channel, we're using
1629
00:45:08,559 --> 00:45:11,039
last click to dictate that.
1630
00:45:11,079 --> 00:45:12,800
And then we kind of back into a last
1631
00:45:12,800 --> 00:45:14,280
click and then an MTA to kind of like
1632
00:45:14,280 --> 00:45:15,840
figure out for our paid channels, like
1633
00:45:15,840 --> 00:45:17,599
what is the happy medium there?
1634
00:45:18,159 --> 00:45:19,719
So that's kind of like the blend.
1635
00:45:19,880 --> 00:45:21,920
At the end of the day, it's like, you
1636
00:45:21,920 --> 00:45:23,920
know, we've had, so for example, like
1637
00:45:23,960 --> 00:45:25,480
email and SMS has been falling short
1638
00:45:25,480 --> 00:45:26,639
based on our daily targets.
1639
00:45:26,639 --> 00:45:27,800
And we're sort of, we're trying to figure
1640
00:45:27,800 --> 00:45:29,159
out, OK, like what's happening on email
1641
00:45:29,159 --> 00:45:30,920
and SMS that's not allowing us to
1642
00:45:31,679 --> 00:45:33,480
continue to like perform at the levels
1643
00:45:33,480 --> 00:45:34,840
and the growth that we need to see month
1644
00:45:34,840 --> 00:45:36,480
over month from those channels.
1645
00:45:36,800 --> 00:45:39,159
And so for us, it was like, oh, well, we
1646
00:45:39,159 --> 00:45:41,119
run a subscription offer, like a discount
1647
00:45:41,119 --> 00:45:43,440
50% off your first purchase and like for
1648
00:45:43,440 --> 00:45:45,159
life on that subscription.
1649
00:45:45,599 --> 00:45:47,719
Um, but we don't really have that many
1650
00:45:47,719 --> 00:45:49,880
like unique offers to add into email and
1651
00:45:49,920 --> 00:45:51,880
SMS to get those extra pops or get the
1652
00:45:51,880 --> 00:45:53,079
people that weren't willing to purchase.
1653
00:45:53,079 --> 00:45:55,280
So there's, there's one like tactic that
1654
00:45:55,280 --> 00:45:56,599
we're about to start testing where
1655
00:45:56,880 --> 00:45:59,480
basically if a customer is sitting in our
1656
00:45:59,519 --> 00:46:02,360
ecosystem, in our own channel list, and
1657
00:46:02,360 --> 00:46:03,960
they haven't converted over the course of
1658
00:46:03,960 --> 00:46:05,159
like a 60 day period, we're actually
1659
00:46:05,159 --> 00:46:07,079
going to start discounting single
1660
00:46:07,079 --> 00:46:08,400
one-time purchase offerings.
1661
00:46:08,599 --> 00:46:10,280
So normally we're not a one-time purchase
1662
00:46:10,280 --> 00:46:11,199
brand. We're not trying to get any
1663
00:46:11,199 --> 00:46:13,079
one-time purchases, but we have, we have
1664
00:46:13,079 --> 00:46:15,239
such big lists now that like there's all
1665
00:46:15,239 --> 00:46:16,920
these people that just, just don't buy
1666
00:46:16,920 --> 00:46:18,000
simply for the fact that they don't want
1667
00:46:18,000 --> 00:46:19,599
to subscribe or they haven't gone on
1668
00:46:19,599 --> 00:46:21,320
Amazon yet to buy on one-time purchase.
1669
00:46:21,320 --> 00:46:23,079
So that's just like a way for us to be
1670
00:46:23,079 --> 00:46:24,719
like, Hey, we have this like dead list
1671
00:46:24,920 --> 00:46:26,760
essentially of just people that have come
1672
00:46:26,760 --> 00:46:28,159
into the funnel, but haven't bought for
1673
00:46:28,159 --> 00:46:29,719
some reason that I think a lot of brands
1674
00:46:29,719 --> 00:46:31,079
actually like kind of forget about that.
1675
00:46:31,079 --> 00:46:32,960
They're like, Hey, they didn't buy. We
1676
00:46:33,000 --> 00:46:35,400
did all these like whatever, uh, flows to
1677
00:46:35,400 --> 00:46:36,440
like get them to maybe get
1678
00:46:36,440 --> 00:46:37,639
a bigger discount over time.
1679
00:46:37,719 --> 00:46:39,320
But then there was probably one reason
1680
00:46:39,320 --> 00:46:41,119
why they didn't buy. So like talk to that
1681
00:46:41,119 --> 00:46:42,440
customer about that one reason why they
1682
00:46:42,440 --> 00:46:43,400
didn't buy and give them a strong
1683
00:46:43,400 --> 00:46:44,760
discount to still close.
1684
00:46:44,800 --> 00:46:47,000
Cause like it's very likely better off
1685
00:46:47,000 --> 00:46:48,679
and you still have contribution margin on
1686
00:46:48,679 --> 00:46:50,239
that customer to be gained.
1687
00:46:50,239 --> 00:46:51,840
If you just give them even a slightly
1688
00:46:51,880 --> 00:46:53,519
bigger discount to like bring in those
1689
00:46:53,519 --> 00:46:54,800
dollars versus them just sitting on an
1690
00:46:54,800 --> 00:46:56,960
email and SMS list and then just like
1691
00:46:56,960 --> 00:46:58,599
dying off because you, you pull them from
1692
00:46:58,599 --> 00:46:59,239
the list because they
1693
00:46:59,239 --> 00:47:00,159
don't, don't interact.
1694
00:47:00,159 --> 00:47:01,679
So yeah, there's like little things like
1695
00:47:01,679 --> 00:47:04,360
that, that we found by digging into, Hey,
1696
00:47:04,760 --> 00:47:06,519
email and SMS is off for that day.
1697
00:47:06,639 --> 00:47:08,079
What, what can we do to actually get
1698
00:47:08,079 --> 00:47:09,599
those extra pops throughout the month?
1699
00:47:10,119 --> 00:47:11,519
Um, and yeah, same
1700
00:47:11,519 --> 00:47:12,159
thing kind of with paid.
1701
00:47:12,159 --> 00:47:13,119
I mean, paid is more of
1702
00:47:13,119 --> 00:47:14,320
the typical analyzing.
1703
00:47:14,360 --> 00:47:16,679
It's like Apple ovens off this week.
1704
00:47:16,719 --> 00:47:17,000
Okay.
1705
00:47:17,000 --> 00:47:17,840
What are we, what are we doing?
1706
00:47:17,960 --> 00:47:19,400
Have we not, you know, have we tested a
1707
00:47:19,400 --> 00:47:20,880
quiz funnel to Apple oven or are we just
1708
00:47:20,880 --> 00:47:22,119
running listicles or whatever?
1709
00:47:22,159 --> 00:47:22,400
Right.
1710
00:47:22,400 --> 00:47:23,800
So it's like landing page creative, the
1711
00:47:23,800 --> 00:47:25,320
whole gambit of normal
1712
00:47:25,320 --> 00:47:26,360
funnel optimization stuff.
1713
00:47:27,360 --> 00:47:29,119
Have you, so the thing you just said
1714
00:47:29,119 --> 00:47:30,639
about like the retention, you guys
1715
00:47:30,639 --> 00:47:32,440
running the same playbook then for people
1716
00:47:32,440 --> 00:47:33,960
who have like lapsed as customers.
1717
00:47:33,960 --> 00:47:35,400
Like they, they tried it for two months
1718
00:47:35,400 --> 00:47:37,559
and they, they unsubscribed and now you
1719
00:47:37,559 --> 00:47:38,400
have to do this whole win
1720
00:47:38,400 --> 00:47:39,480
back to get them back on board.
1721
00:47:39,559 --> 00:47:40,199
Are you guys good?
1722
00:47:40,440 --> 00:47:40,880
Same thing.
1723
00:47:40,920 --> 00:47:42,719
One time purchase or push them to Amazon.
1724
00:47:43,440 --> 00:47:43,800
Yeah.
1725
00:47:43,840 --> 00:47:44,079
Yeah.
1726
00:47:44,119 --> 00:47:44,559
Same thing.
1727
00:47:44,639 --> 00:47:45,119
Same thing.
1728
00:47:45,159 --> 00:47:45,320
Yeah.
1729
00:47:45,320 --> 00:47:46,760
I think the biggest thing this week was
1730
00:47:46,760 --> 00:47:48,320
just like, Hey, we're not talking to
1731
00:47:48,320 --> 00:47:49,400
these people that just don't want to buy
1732
00:47:49,400 --> 00:47:50,800
it because it's just a subscription.
1733
00:47:51,360 --> 00:47:51,960
So it's like, let's just
1734
00:47:51,960 --> 00:47:52,840
go clean that up first.
1735
00:47:52,880 --> 00:47:54,119
Cause that's probably hundreds of
1736
00:47:54,119 --> 00:47:55,000
thousands of revenue,
1737
00:47:55,000 --> 00:47:55,880
like every single month.
1738
00:47:55,880 --> 00:47:57,719
If we just give them a proper discount to
1739
00:47:57,719 --> 00:47:59,199
come in one time purchase at like not
1740
00:47:59,199 --> 00:48:00,960
quite our subscription offering, but like
1741
00:48:01,199 --> 00:48:02,719
in, in between our one time purchase and
1742
00:48:02,719 --> 00:48:03,639
our subscription price.
1743
00:48:04,920 --> 00:48:05,480
So yeah.
1744
00:48:05,480 --> 00:48:06,079
And then next would
1745
00:48:06,079 --> 00:48:07,519
obviously be like lapsed customers.
1746
00:48:07,760 --> 00:48:09,079
The interesting thing about, about
1747
00:48:09,079 --> 00:48:10,679
supplements and lapsed customers is like
1748
00:48:10,760 --> 00:48:12,519
very, very low chance of their, we've
1749
00:48:12,519 --> 00:48:14,320
tried things before in the past, like
1750
00:48:14,320 --> 00:48:15,119
free gifts and shit.
1751
00:48:15,119 --> 00:48:16,079
It just like doesn't move the needle.
1752
00:48:16,119 --> 00:48:19,719
So, um, yeah, we've, we've, we've, we've
1753
00:48:19,719 --> 00:48:20,840
attempted it, but I would say like,
1754
00:48:20,840 --> 00:48:21,400
there's probably still
1755
00:48:21,400 --> 00:48:22,800
more, more room to fix there.
1756
00:48:22,880 --> 00:48:24,079
I mean, absolute banger.
1757
00:48:24,199 --> 00:48:25,760
I tried not to talk very much cause I
1758
00:48:25,760 --> 00:48:28,679
sound terrible, but Brad appreciate you
1759
00:48:28,679 --> 00:48:29,800
asking good questions.
1760
00:48:30,239 --> 00:48:31,719
Zach, thank you for sharing.
1761
00:48:31,719 --> 00:48:32,320
Is there anything else
1762
00:48:32,320 --> 00:48:33,159
you guys want to go into?
1763
00:48:33,280 --> 00:48:33,519
No.
1764
00:48:34,000 --> 00:48:34,239
Yeah.
1765
00:48:34,480 --> 00:48:34,719
Okay.
1766
00:48:34,800 --> 00:48:35,199
All set.
1767
00:48:35,639 --> 00:48:37,480
Poskey Foxwell signing off.
1768
00:48:37,639 --> 00:48:39,679
Thank you for listening and
1769
00:48:39,679 --> 00:48:40,639
we'll talk to you guys soon.
1770
00:48:46,280 --> 00:48:48,440
This episode is brought to you by the
1771
00:48:48,440 --> 00:48:50,679
Foxwell founders membership that Andrew
1772
00:48:50,679 --> 00:48:51,840
and his wife, Gracie run.
1773
00:48:51,880 --> 00:48:54,519
It has been absolutely pivotal for not
1774
00:48:54,519 --> 00:48:55,599
just the Homestead team, but
1775
00:48:55,599 --> 00:48:56,679
the easy street brands team.
1776
00:48:57,199 --> 00:48:58,480
We've had, I don't even know how many
1777
00:48:58,480 --> 00:49:00,159
members are currently in there that are a
1778
00:49:00,159 --> 00:49:01,199
part of our ecosystem.
1779
00:49:01,199 --> 00:49:02,960
But when it comes to anything from
1780
00:49:02,960 --> 00:49:04,599
learning ads to understanding what's
1781
00:49:04,599 --> 00:49:06,960
going on to building an agency to knowing
1782
00:49:06,960 --> 00:49:08,119
retention, it's been
1783
00:49:08,199 --> 00:49:10,320
absolutely useful for our team.
1784
00:49:10,320 --> 00:49:11,800
When they get stuck or they need help to
1785
00:49:11,800 --> 00:49:12,559
just go there and
1786
00:49:12,760 --> 00:49:14,199
resource all the other experts.
1787
00:49:14,199 --> 00:49:15,639
So definitely would recommend it for
1788
00:49:15,639 --> 00:49:17,519
anybody that's looking to, you know, take
1789
00:49:17,519 --> 00:49:19,280
it a step deeper, try to get a little bit
1790
00:49:19,280 --> 00:49:21,159
more knowledge on growth marketing and
1791
00:49:21,519 --> 00:49:22,599
all the world D2C is.
1792
00:49:22,679 --> 00:49:22,760
Yeah.
1793
00:49:22,760 --> 00:49:24,239
I think one of the most incredible things
1794
00:49:24,239 --> 00:49:25,960
about it is you can just like open up the
1795
00:49:25,960 --> 00:49:27,400
Slack group every single day.
1796
00:49:27,400 --> 00:49:28,840
You can pin your favorite channels for
1797
00:49:28,840 --> 00:49:30,320
the topics that you care most about.
1798
00:49:30,480 --> 00:49:31,920
And like every day there's going to be
1799
00:49:31,920 --> 00:49:33,880
somebody who just like, because they want
1800
00:49:33,880 --> 00:49:35,440
to contribute something valuable to the
1801
00:49:35,440 --> 00:49:36,800
group, you can go learn
1802
00:49:36,800 --> 00:49:37,880
something every single day.
1803
00:49:37,960 --> 00:49:39,559
And it's going to be extremely useful.
1804
00:49:40,119 --> 00:49:41,639
There's some ballers in there that you
1805
00:49:41,639 --> 00:49:42,159
just get like the
1806
00:49:42,159 --> 00:49:43,199
benefit of learning from that.
1807
00:49:43,199 --> 00:49:45,400
But like for the for the costs, like you
1808
00:49:45,400 --> 00:49:47,159
couldn't pay them that for their time.
1809
00:49:47,239 --> 00:49:48,400
But through the membership, like you get
1810
00:49:48,400 --> 00:49:49,639
access to some incredible
1811
00:49:49,639 --> 00:49:51,320
people and tons of resources.
1812
00:49:51,719 --> 00:49:51,920
Yeah.
1813
00:49:51,960 --> 00:49:53,280
I mean, I think the biggest resource to
1814
00:49:53,280 --> 00:49:55,360
me, too, is like the events that, you
1815
00:49:55,360 --> 00:49:56,320
know, Foxville Founders does.
1816
00:49:56,320 --> 00:49:57,639
They've been able to do some even in
1817
00:49:57,639 --> 00:49:59,920
Wisconsin, even in the boring state of
1818
00:49:59,920 --> 00:50:01,719
Wisconsin, which is pretty awesome.
1819
00:50:01,719 --> 00:50:03,440
Getting people together in person and
1820
00:50:03,920 --> 00:50:05,119
able to have really just like honest
1821
00:50:05,119 --> 00:50:06,679
conversations of what's going on, what's
1822
00:50:06,679 --> 00:50:08,559
working for them now, you know, where
1823
00:50:08,719 --> 00:50:09,800
they're at in their business.
1824
00:50:09,840 --> 00:50:11,000
Knowing that there's going to be, like
1825
00:50:11,000 --> 00:50:12,840
Brad said, some real killers in the space
1826
00:50:12,920 --> 00:50:15,199
in this in this membership that can that
1827
00:50:15,199 --> 00:50:16,000
can help and are willing
1828
00:50:16,000 --> 00:50:16,920
to take the time and help.
1829
00:50:17,159 --> 00:50:19,320
So that's been a huge part of why a lot
1830
00:50:19,320 --> 00:50:20,079
of our team have
1831
00:50:20,079 --> 00:50:20,920
really enjoyed it as well.
1832
00:50:20,960 --> 00:50:22,880
And the applications are now open if
1833
00:50:22,880 --> 00:50:23,519
you're looking to join.
1834
00:50:23,519 --> 00:50:25,519
So Foxville Founders.
1835
00:50:25,559 --> 00:50:25,920
Yeah.
1836
00:50:26,119 --> 00:50:27,119
Foxvillefounders.com.
1837
00:50:27,360 --> 00:50:27,880
Go check it out.
1838
00:50:27,880 --> 00:50:28,159
Go apply.
1839
00:50:28,159 --> 00:50:34,880
The only way that we grow this podcast is
1840
00:50:34,880 --> 00:50:36,840
by you sharing it with your friends.
1841
00:50:36,880 --> 00:50:38,400
Honestly, like reviews kind of don't
1842
00:50:38,400 --> 00:50:39,800
really mean anything too much anymore.
1843
00:50:39,800 --> 00:50:41,559
They're really meaningful, but they don't
1844
00:50:41,559 --> 00:50:43,320
do a lot for the growth of the podcast.
1845
00:50:43,760 --> 00:50:46,039
And so sharing YouTube links, sharing
1846
00:50:46,039 --> 00:50:49,159
Spotify links, sharing Apple, whatever we
1847
00:50:49,159 --> 00:50:50,920
call it under the podcast app now.
1848
00:50:51,199 --> 00:50:52,519
Anything you can share,
1849
00:50:52,639 --> 00:50:54,639
the better we're going to be.
1850
00:50:54,880 --> 00:50:55,639
Guys, anything else
1851
00:50:55,639 --> 00:50:56,519
you want to say on this?
1852
00:50:56,639 --> 00:50:56,960
Yeah.
1853
00:50:57,000 --> 00:50:58,559
Please go check us out on YouTube.
1854
00:50:58,599 --> 00:50:59,880
Rack up those views for us.
1855
00:50:59,880 --> 00:51:00,679
We'd love to see it.
1856
00:51:00,760 --> 00:51:01,440
And then subscribe.
1857
00:51:01,639 --> 00:51:02,000
Make sure to
1858
00:51:02,000 --> 00:51:03,039
subscribe on YouTube as well.
1859
00:51:03,159 --> 00:51:04,840
And I relentlessly refresh the YouTube
1860
00:51:04,840 --> 00:51:06,000
comments because it dictates
1861
00:51:06,000 --> 00:51:06,920
my mental health for the day.
1862
00:51:06,920 --> 00:51:07,440
So please say
1863
00:51:07,440 --> 00:51:08,519
something nice about all of us.
1864
00:51:10,000 --> 00:51:10,880
Thank you, everyone.
1865
00:51:11,000 --> 00:51:11,960
Thanks for listening.
1866
00:51:11,960 --> 00:51:12,480
Honestly.