전사 🔗
1
00:00:02,189 --> 00:00:11,825
Today we have a special guest, Brian Beachkofski, CTO of Rithmm, a cutting edge platform
that's revolutionizing sports betting with advanced data analytics.
2
00:00:11,825 --> 00:00:20,489
Rhythm provides personalized betting models for recreational sports bettors, leveraging AI
and complex data analysis to deliver actionable insights.
3
00:00:20,489 --> 00:00:28,888
Brian builds a wealth of experience to the, or brings about wealth of experience to the
table with a background that spans from jet engine reliability analysis,
4
00:00:28,888 --> 00:00:32,679
to leading large scale software development teams in the US Air Force.
5
00:00:32,679 --> 00:00:43,382
He holds a PhD in probabilistic analysis and has applied his expertise in data science
across various domains before venturing into the world of sports betting analytics.
6
00:00:43,382 --> 00:00:44,953
Brian, it's great to have you on the show.
7
00:00:44,953 --> 00:00:46,433
Thank you.
8
00:00:46,433 --> 00:00:49,504
It's great to be here looking forward to a great conversation.
9
00:00:49,504 --> 00:00:49,924
Yeah.
10
00:00:49,924 --> 00:00:54,005
Tell us a little bit about, you know, Rhythm and your role within the company.
11
00:00:54,725 --> 00:00:55,145
Yeah.
12
00:00:55,145 --> 00:00:57,349
So Rhythm was founded to solve
13
00:00:57,349 --> 00:01:10,809
problem for folks in recreational sports betting, which is we bet a lot the same way that
we did 100 years ago, which is either go with your gut or you ask your buddy who they like
14
00:01:10,809 --> 00:01:12,029
on tonight's game.
15
00:01:12,029 --> 00:01:18,981
But in the rest of the world, you have analytics and you have data that guide those
decisions and help you create
16
00:01:18,981 --> 00:01:19,661
create them.
17
00:01:19,661 --> 00:01:27,981
And in the bedding world, the people who are using those kinds of models are the ones that
are betting thousands of dollars and probably doing it professionally.
18
00:01:27,981 --> 00:01:34,126
We're trying to make that same level of sophistication accessible to the recreational
better that has primarily been
19
00:01:34,126 --> 00:01:35,967
available to the professionals.
20
00:01:36,108 --> 00:01:37,248
That's incredible.
21
00:01:37,248 --> 00:01:41,219
And I know, you you have an incredible and are a fascinating career journey.
22
00:01:41,219 --> 00:01:47,770
Tell us a little bit about your path from, you know, the Air Force Academy to Rhythm.
23
00:01:48,250 --> 00:01:55,632
Yeah, so I grew up in Wisconsin and went to the Air Force Academy where I studied
engineering mechanics.
24
00:01:56,172 --> 00:01:58,973
Did a lot on life prediction.
25
00:01:59,335 --> 00:02:02,136
looking at how materials age and break.
26
00:02:02,156 --> 00:02:11,101
They sent me to Georgia Tech to get my master's in aerospace engineering and from there
went to the Air Force Research Lab where I applied that reliability and material strength
27
00:02:11,101 --> 00:02:12,261
analysis.
28
00:02:12,321 --> 00:02:13,662
to jet engines.
29
00:02:13,662 --> 00:02:23,467
And looking at how we can make reliability predictions better, can we change the way we do
maintenance so that it's more like we have in our cars now.
30
00:02:23,467 --> 00:02:31,152
an oil change light comes on and it's based on how you're actually driving and not just
how many miles have clicked off the odometer.
31
00:02:31,152 --> 00:02:33,213
So we did a lot of analysis on that.
32
00:02:33,213 --> 00:02:39,987
And one of the funny things that I ran into there was that it was hard for me to get
actual real world engine
33
00:02:39,987 --> 00:02:43,579
data, but what I could get a lot of data from is sports.
34
00:02:43,579 --> 00:02:53,644
So when I was trying to get streaming data, it was easier for me to just go to
baseballreference .com and get a hundred years of at bats and treat that as streaming data
35
00:02:53,644 --> 00:02:56,101
and ask interesting questions and practice my
36
00:02:56,101 --> 00:03:04,661
So that's kind of where my love of sports came into combination with my love of stats and
doing that.
37
00:03:04,661 --> 00:03:09,951
It was how I taught myself how to use some of the advanced statistical methods.
38
00:03:10,051 --> 00:03:21,454
From there I went to the Pentagon, then I went into private industry, worked for a
nonprofit in Boston doing data analysis on how state and local governments can use data to
39
00:03:21,454 --> 00:03:25,113
understand the benefits of human service delivery in their community.
40
00:03:25,113 --> 00:03:33,820
And then from there, I got an opportunity to come back into the Air Force and lead Kessel
Run, which was the premier software development team within the Air Force doing the
41
00:03:33,820 --> 00:03:36,132
command and control software development.
42
00:03:36,132 --> 00:03:39,551
So learned a lot of software delivery practices while I was there.
43
00:03:39,551 --> 00:03:40,922
It's been, it's been a fun ride.
44
00:03:40,922 --> 00:03:41,753
Yeah.
45
00:03:41,753 --> 00:03:48,419
And was there kind of a pivotal moment in that career that kind of influenced your
current, you know, your, your path now?
46
00:03:48,907 --> 00:03:52,620
Yeah, I think there's been a couple of guiding principles through my career.
47
00:03:52,620 --> 00:04:00,025
One is just been how do we use data to make better decisions, whether that's how do you
change maintenance of jet engines.
48
00:04:00,025 --> 00:04:08,911
When I was in the Pentagon, it was about how do you make investment decisions for the
future, whether that's in human capital, like what kinds of people do we need to bring
49
00:04:08,911 --> 00:04:13,514
into the military or what kind of science and technology should we be investing in?
50
00:04:13,514 --> 00:04:17,236
And then at Third Sector, it was looking at
51
00:04:17,771 --> 00:04:27,847
Like I said, how human service delivery is working
Load full transcript
in the community and can we allocate
those funds to the people making the biggest difference in their community?
52
00:04:27,847 --> 00:04:33,849
And now it's how do you have the available data to help people make recreational bedding
decisions?
53
00:04:33,849 --> 00:04:35,811
But that's been one through line.
54
00:04:36,252 --> 00:04:38,624
And then another has just
55
00:04:38,624 --> 00:04:43,396
been innovation and how do we apply the new things into what we're doing?
56
00:04:43,396 --> 00:04:50,587
So I was in the research lab for the Air Force and in the Pentagon, of the offices I
worked in was the office of the
57
00:04:50,587 --> 00:04:55,280
Secretary of Defense for Research and Engineering, asking questions about what's next.
58
00:04:55,280 --> 00:05:08,188
So we did a lot of future studies, looked at things like 3D printing back in the day when
that was one of the big questions, or synthetic biology and how do you...
59
00:05:08,733 --> 00:05:13,747
How would manufactured DNA, RNA help influence things?
60
00:05:13,747 --> 00:05:18,480
And we saw that was a pretty big deal when it came to COVID vaccination.
61
00:05:19,101 --> 00:05:22,294
So it's always been like, what are the trends for the future?
62
00:05:22,294 --> 00:05:28,916
How can we change what we're doing and do things better by using more technology to make
those
63
00:05:28,916 --> 00:05:30,119
to make those decisions.
64
00:05:30,119 --> 00:05:31,800
Yeah, definitely.
65
00:05:32,481 --> 00:05:37,594
in terms of, know, I think you've been with a company almost coming up maybe on two and a
half years, is that right?
66
00:05:37,594 --> 00:05:39,525
Yeah, that's right.
67
00:05:39,525 --> 00:05:39,855
Okay.
68
00:05:39,855 --> 00:05:40,746
been a while.
69
00:05:40,746 --> 00:05:43,827
You know, with startups, it always depends when you call the start line.
70
00:05:43,827 --> 00:05:44,238
Yeah.
71
00:05:44,238 --> 00:05:47,610
So we were messing around with it a year or two before that.
72
00:05:47,610 --> 00:05:48,140
Okay.
73
00:05:48,140 --> 00:05:50,061
Officially company formed about...
74
00:05:50,061 --> 00:05:51,371
and a half years
75
00:05:51,479 --> 00:06:02,504
Okay, and tell me a little bit about how it's evolved, know the initial vision when you
joined versus, you know how it's evolving now Tell me a little bit about what how that's
76
00:06:03,019 --> 00:06:05,151
No, that's a fun journey.
77
00:06:05,151 --> 00:06:09,133
So in the beginning, like any entrepreneur, have an interesting idea.
78
00:06:09,133 --> 00:06:13,026
And ours was, can you use data to make a personalized model?
79
00:06:13,026 --> 00:06:18,280
So not just have a company that tells people what the model thinks is best.
80
00:06:18,280 --> 00:06:26,516
But if you want to have your own model based on your own intuition, how would we make
modeling easier for you?
81
00:06:26,516 --> 00:06:29,968
So if you don't have that time, if you don't have the...
82
00:06:31,591 --> 00:06:39,257
computational background to be able to build your own model, how can we still turn your
intuition and what makes a winning team into a model?
83
00:06:39,257 --> 00:06:41,098
will align with that more easily?
84
00:06:41,098 --> 00:06:44,760
So the first start was just like proof of concept, right?
85
00:06:44,760 --> 00:06:57,398
Like I built something pretty quick and started tracking data and saw that it was working,
that when we said there was a higher variance to the current market prices, things would
86
00:06:57,398 --> 00:07:00,229
win at a higher rate, which was really cool to see.
87
00:07:00,422 --> 00:07:07,615
Then we went out and built a very basic product that we could start getting some user
feedback.
88
00:07:07,615 --> 00:07:12,965
So we went out and outsourced to a development company to be able to make a web app
version.
89
00:07:12,965 --> 00:07:15,336
did that for a year and collected more data.
90
00:07:15,336 --> 00:07:24,109
And that's actually what allowed us to go to our initial funding round because we had two
years of data that showed, yeah, this works.
91
00:07:24,109 --> 00:07:25,729
You could have not only...
92
00:07:25,729 --> 00:07:34,793
the basic model work, but then we had the next year of individuals making their own models
and showing that that own modeling process still could create winning.
93
00:07:35,604 --> 00:07:36,735
predictions.
94
00:07:36,920 --> 00:07:46,650
I love that because I do that for my own stocks where I I mean in a really basic sense I
create my own models based on what my priorities are right so I have I'm a very risky
95
00:07:46,650 --> 00:07:56,590
investor investor so I have maybe Less than 10 stocks that I'm that I'm in and I really
prioritize, know company turnover things like this that some of these models don't
96
00:07:56,590 --> 00:08:06,047
prioritize as heavy as I currently weight it and I would love to have a tool like that for
my stock portfolio, know
97
00:08:06,047 --> 00:08:08,339
yeah, there's plenty of use cases for this.
98
00:08:08,339 --> 00:08:09,779
And we've talked about that as well.
99
00:08:09,779 --> 00:08:14,192
Where else can turning intuitions into a model help somebody?
100
00:08:14,192 --> 00:08:20,486
Because one of the big problems we've seen in AI are people don't know what's inside that
black box.
101
00:08:20,486 --> 00:08:31,105
So if someone just told you these are the stocks to pick, but you didn't know why, it
would be very hard for you to believe it and put your money behind
102
00:08:31,105 --> 00:08:34,810
But if you say it's because these things are emphasized, right?
103
00:08:34,810 --> 00:08:41,826
It's emphasizing last quarter growth or it's emphasizing, like you said, turnover, then
you know what's going in
104
00:08:41,826 --> 00:08:42,206
in there?
105
00:08:42,206 --> 00:08:43,026
Yeah, exactly.
106
00:08:43,026 --> 00:08:45,227
Or, you know, internal purchases, right?
107
00:08:45,227 --> 00:08:47,418
Like, that's another huge factor that I care about.
108
00:08:47,418 --> 00:08:47,709
Right.
109
00:08:47,709 --> 00:08:48,589
And I want to wait.
110
00:08:48,589 --> 00:08:50,210
I want to put that pie chart together.
111
00:08:50,210 --> 00:08:57,373
You know, I don't want to rely on their pie chart because I like to test my own theories
and figure out my own winning strategy.
112
00:08:57,373 --> 00:08:58,273
So that's incredible.
113
00:08:58,273 --> 00:09:01,468
And I can see how this technology could be leveraged across
114
00:09:01,468 --> 00:09:02,708
multiple industries.
115
00:09:02,708 --> 00:09:04,299
You mentioned AI a little bit.
116
00:09:04,299 --> 00:09:15,212
Tell me a little bit about how this rapidly evolving landscape of AI is affecting the
business at the moment.
117
00:09:16,032 --> 00:09:20,634
So AI is a wide research area.
118
00:09:20,634 --> 00:09:25,535
So there's lots in there that different methodologies that people can use.
119
00:09:25,535 --> 00:09:31,124
I think the ones that people are seeing at the very top of the headlines are the large
language
120
00:09:31,124 --> 00:09:38,188
Ours are not large language models, but they're using some of the same principles that
have been used in machine learning for quite some time.
121
00:09:39,871 --> 00:09:42,193
so we combine actually multiple models.
122
00:09:42,193 --> 00:09:46,518
have linear models, we have nonlinear models, each making their own prediction.
123
00:09:46,518 --> 00:09:51,851
And then we do what's called model ensembling, which is another characteristic of AI.
124
00:09:51,851 --> 00:10:03,744
where if you have independent models creating their own predictions, the wisdom of the
crowd or the wisdom of the crowd of models will be better than any one model individually.
125
00:10:03,744 --> 00:10:07,493
So that's what we've done is had, we've created a way to have
126
00:10:07,493 --> 00:10:10,774
have different models built underneath that each make their prediction.
127
00:10:10,774 --> 00:10:16,301
And then when we combine those, we get a more accurate prediction than either one
individually.
128
00:10:16,301 --> 00:10:19,312
And the cool part is we can tailor that to your own preference.
129
00:10:19,312 --> 00:10:31,027
So when you come in and say, I like offense and I like a high tempo team, it goes in and
uses those parameters to create your own bespoke training data.
130
00:10:31,027 --> 00:10:35,592
And that training data is what gets your model
131
00:10:35,592 --> 00:10:42,984
particularly sensitive to teams with those kinds of statistical traits and brings us to
the top.
132
00:10:43,215 --> 00:10:44,226
Interesting.
133
00:10:44,226 --> 00:10:51,911
And I know before we talked a little bit about, you know, periods of time where Rhythm has
experienced rapid growth.
134
00:10:51,911 --> 00:10:54,723
What has kind of growth looked like since you've been at the company?
135
00:10:54,723 --> 00:10:58,566
Yeah, it's been a really fun experience.
136
00:10:58,566 --> 00:11:06,171
And one of the things that any entrepreneur needs to do is use their own app.
137
00:11:06,171 --> 00:11:09,133
or as we say, eat your own dog food, right?
138
00:11:09,133 --> 00:11:13,175
Like you can't guess what the user experience is.
139
00:11:13,175 --> 00:11:14,476
You've got to be in there.
140
00:11:14,476 --> 00:11:19,009
And our first iteration was built to just to have something working, right?
141
00:11:19,009 --> 00:11:23,541
And then what we noticed is last football season,
142
00:11:24,179 --> 00:11:33,213
as we started to get more growth, the higher user count was creating a bigger demand on
the services and things were slowing down quite a bit.
143
00:11:33,213 --> 00:11:36,933
So there was a time there where it wasn't the greatest user experience.
144
00:11:36,933 --> 00:11:42,673
If you wanted to see your prediction, it was sometimes 30 seconds or so you had to wait
after logging in.
145
00:11:42,673 --> 00:11:48,293
And you know in an app, if it's more than five seconds, it might as well never show up
because somebody is out of there.
146
00:11:48,293 --> 00:11:51,173
So what we were able to do is reaffirm
147
00:11:51,839 --> 00:11:55,749
and re -architect our whole way that we were making.
148
00:11:55,749 --> 00:11:57,049
delivering those predictions.
149
00:11:57,049 --> 00:12:07,659
The core model stayed the same, but how we were making those predictions available to the
user to change, which brought down that time quite a lot.
150
00:12:07,659 --> 00:12:14,789
It went from 30 seconds down to one second, and you were now seeing your predictions.
151
00:12:14,789 --> 00:12:15,789
So like
152
00:12:16,319 --> 00:12:20,819
did a couple other things, but that was the main one of just improving the user
experience.
153
00:12:20,819 --> 00:12:32,269
We know that what we're doing is unique and different than other people out there and just
maximizing user experience so that people don't have to wait for what's there,
154
00:12:32,269 --> 00:12:34,444
everything's intuitive on how
155
00:12:34,444 --> 00:12:35,914
the information it's providing.
156
00:12:35,914 --> 00:12:41,957
And that change led to a pretty dramatic growth rate increase last January.
157
00:12:43,083 --> 00:12:51,867
it was an exciting time there where we started saying, good thing we made those changes
because the growth rate now, by making those changes the growth rates going higher which
158
00:12:51,867 --> 00:12:54,338
if we didn't make those changes would have absolutely crushed us.
159
00:12:54,338 --> 00:12:57,020
So it was good experience.
160
00:12:57,020 --> 00:13:03,613
So once you make those changes, what was the turnaround time to really start seeing that
growth happen and is that happening by word of mouth?
161
00:13:03,613 --> 00:13:04,393
Is that happening
162
00:13:04,393 --> 00:13:06,164
your own internal marketing and PR?
163
00:13:06,164 --> 00:13:07,855
What does that look like?
164
00:13:07,855 --> 00:13:12,527
Yeah, so what we saw was that our two things happened.
165
00:13:12,527 --> 00:13:17,300
One is as people were trying the app out, more were retaining.
166
00:13:17,300 --> 00:13:24,024
So we have a seven day trial period that's free and then it goes into a subscription if
you want to continue.
167
00:13:24,024 --> 00:13:28,646
That rate of people continuing there is really what made the big change.
168
00:13:28,646 --> 00:13:29,986
So people who
169
00:13:30,426 --> 00:13:35,133
before frustrated by that user experience and said, it's not worth me subscribing, we had
170
00:13:35,133 --> 00:13:38,874
a higher adoption rate going to subscriptions at that time.
171
00:13:38,874 --> 00:13:40,945
So that was immediately what you could see.
172
00:13:40,945 --> 00:13:49,138
And the time delay there was only the trial period because the people who came in when it
was a great experience, seven days later were saying, yeah, I want this.
173
00:13:49,138 --> 00:13:54,980
And then the second part of it is because it was a good experience, we saw a bigger
increase in word of mouth.
174
00:13:54,980 --> 00:13:59,039
So the marketing actually got more productive and we're able to get more people coming
175
00:13:59,039 --> 00:14:01,220
door and then more people retaining.
176
00:14:01,220 --> 00:14:05,073
So all of these things are sort of synergistic and build on.
177
00:14:05,073 --> 00:14:09,974
build on each other to make it more effective overall.
178
00:14:09,974 --> 00:14:12,895
How often are you in the platform as a user?
179
00:14:13,585 --> 00:14:14,616
multiple times a day.
180
00:14:14,616 --> 00:14:15,046
Really?
181
00:14:15,046 --> 00:14:18,417
Yeah, I'm probably addicted.
182
00:14:18,417 --> 00:14:20,438
That's another thing.
183
00:14:20,438 --> 00:14:21,358
never
184
00:14:21,639 --> 00:14:23,020
before doing this.
185
00:14:23,020 --> 00:14:27,813
And then when we built that first prototype, I was like, we need to make it happen for a
while.
186
00:14:27,813 --> 00:14:28,493
was difficult.
187
00:14:28,493 --> 00:14:36,929
I live in Massachusetts and we only got legalized gambling here about a year and a half
ago now.
188
00:14:36,929 --> 00:14:41,597
And before that, I'd have to drive down to Rhode Island and play some bets.
189
00:14:41,597 --> 00:14:49,286
I, you know, you can do so much on paper betting, but you really don't understand the
stakes until you're putting your own money behind it.
190
00:14:49,286 --> 00:14:51,629
So yeah, that was actually how I got into it.
191
00:14:51,629 --> 00:14:53,061
And now I love it.
192
00:14:53,061 --> 00:14:54,052
It's a great experience.
193
00:14:54,052 --> 00:14:54,876
It's super fun.
194
00:14:54,876 --> 00:14:56,637
And I'm in there multiple times
195
00:14:56,665 --> 00:15:06,845
Yeah, and I don't know much about like, I guess, regulatory, the
regulatory environment,
how does that affect you and how could that change over the coming years that would affect
196
00:15:06,845 --> 00:15:08,005
your business?
197
00:15:08,226 --> 00:15:14,102
Yeah, one of the big benefits that we have is that we don't take the bets directly.
198
00:15:14,102 --> 00:15:15,933
So we are not a clearing house.
199
00:15:15,933 --> 00:15:19,117
We don't have that kind of regulatory pressure.
200
00:15:19,117 --> 00:15:20,357
We are
201
00:15:20,936 --> 00:15:22,047
analytics company.
202
00:15:22,047 --> 00:15:30,323
So we are providing that information and looking at people who are letting people use that
however they want to.
203
00:15:30,323 --> 00:15:36,551
One of the great insights we got from our user data is that people were using it not just
204
00:15:36,551 --> 00:15:40,823
for not just for traditional sports betting.
205
00:15:40,823 --> 00:15:42,423
They were using it for daily fantasy.
206
00:15:42,423 --> 00:15:47,275
They were using it for these other ways that we weren't we weren't anticipating.
207
00:15:47,275 --> 00:15:51,887
So even in some states where it wasn't legal, were where it wasn't regulated.
208
00:15:51,887 --> 00:15:57,702
We were seeing that there was a bigger uptake and we asked some people what's going on and
they said,
209
00:15:57,702 --> 00:16:00,652
use it for daily fantasy because it's a great product for that as
210
00:16:00,652 --> 00:16:00,982
well.
211
00:16:00,982 --> 00:16:02,292
Yeah, definitely.
212
00:16:02,292 --> 00:16:10,392
Tell me a little bit about just your kind of leadership philosophy, you know, leading and
managing the team, the culture.
213
00:16:10,392 --> 00:16:12,412
Tell me a little bit about that.
214
00:16:12,412 --> 00:16:19,112
Yeah, I touched briefly on it before, but I was lucky enough to be in the Air Force for
quite a while.
215
00:16:19,112 --> 00:16:25,157
So active duty military, I was a reservist for 20 years about and then.
216
00:16:25,157 --> 00:16:29,017
a government civilian working within the Department of Defense.
217
00:16:29,017 --> 00:16:35,667
And one of the great things about that is there is a formal leadership training process.
218
00:16:35,667 --> 00:16:36,907
So I think I had
219
00:16:36,907 --> 00:16:42,870
a pretty strong advantage by going through formal leadership development process within
the military.
220
00:16:43,011 --> 00:16:48,023
And one of the things that you really see stressed there is servant leadership.
221
00:16:48,023 --> 00:16:54,473
So I see my job as the person who does two things first.
222
00:16:54,473 --> 00:17:02,706
provide the vision, make it very clear so that the individual contributor is making the
product know exactly the direction they're getting.
223
00:17:02,706 --> 00:17:10,701
There's nothing more frustrating than thinking you're going in the right way, only to be
at the end of the project, know, like, wait, no, that's not right, we gotta rework, we
224
00:17:10,701 --> 00:17:13,053
gotta go this way, it's super frustrating.
225
00:17:13,053 --> 00:17:18,356
So the one very simple thing is make the direction you're going very clear.
226
00:17:18,356 --> 00:17:19,276
The other
227
00:17:20,266 --> 00:17:23,106
my job is to remove blockers, right?
228
00:17:23,227 --> 00:17:26,217
It is to say, okay, what would make your job easier?
229
00:17:26,217 --> 00:17:29,058
What would, what do you, what information do you need?
230
00:17:29,058 --> 00:17:31,648
Who do you need to talk to and just make that stuff
231
00:17:31,648 --> 00:17:32,029
happen.
232
00:17:32,029 --> 00:17:38,895
So it is setting the direction and then just making getting in that direction as easy as
possible.
233
00:17:40,460 --> 00:17:47,453
Yeah, I say that to people all the time in their careers as a recruiter, you know, you
don't the last thing you want to do is climb all the way up the ladder to find out that
234
00:17:47,453 --> 00:17:48,794
you're on top of the wrong building.
235
00:17:48,794 --> 00:17:55,247
You get into the weeds of the executive culture and you realize, my gosh, this is not a
company I truly want to work for or believe in.
236
00:17:55,247 --> 00:17:59,979
But I see that so many times where people have spent their last five or so years working
for this company.
237
00:17:59,979 --> 00:18:03,450
And then I've come to the realization that they're on top of the wrong building.
238
00:18:04,211 --> 00:18:06,532
So that's really key.
239
00:18:07,113 --> 00:18:10,327
What about just in terms of your own personal philosophies
240
00:18:10,327 --> 00:18:20,927
maybe not leadership, but just in terms of scaling and growing and the right way to manage
some of those challenging times where you're starting to see this rapid growth and your
241
00:18:20,927 --> 00:18:24,687
role is evolving rapidly because the company starts to grow rapidly.
242
00:18:24,687 --> 00:18:26,387
Tell me a little bit about that.
243
00:18:26,647 --> 00:18:32,687
Yeah, leadership and leaders are made when there isn't an obvious answer.
244
00:18:32,687 --> 00:18:40,029
One of the big common quotes about the presidency is everything that comes to the
president is a 50 -50 decision, right?
245
00:18:40,029 --> 00:18:42,989
decision is really super hard.
246
00:18:43,049 --> 00:18:49,924
And that's the same, especially in a startup, when you're looking at every single decision
247
00:18:49,924 --> 00:18:55,265
is, we spend the time now to address this or are we willing to accumulate tech debt?
248
00:18:55,265 --> 00:19:06,007
we, you know, like, talked about those growth patterns, how much are we willing to improve
user experience now to test whether or not we're on the right direction versus putting
249
00:19:06,007 --> 00:19:09,975
infrastructure in place to ensure that scaling can happen appropriately?
250
00:19:09,975 --> 00:19:12,877
And there aren't obvious answers to those, right?
251
00:19:12,877 -->
00:19:18,041
Like, if you always say we should just accumulate tech debt,
252
00:19:18,757 --> 00:19:22,849
you'll put yourself in a position where you can't handle the growth you've done.
253
00:19:22,849 --> 00:19:29,192
And if you always say we need to make this perfect, you'll never deliver anything of value
to your customers.
254
00:19:29,192 --> 00:19:38,719
So there's a constant tension between working for today and making sure that you have a
sustainable platform.
255
00:19:38,719 --> 00:19:48,633
And that's kind of the leadership is understanding what the demands are at the moment,
what the context is, having an intuition for the forecasted growth and understanding both
256
00:19:48,633 --> 00:19:59,357
the risk and consequence of turning left or turning right and in owning that decision of
saying I'm turning left or turning right and sometimes saying, you know what?
257
00:19:59,551 --> 00:20:01,022
We built up a little tech debt.
258
00:20:01,022 --> 00:20:10,629
We're just going to have to pause for a quarter and come back and build that stuff in
place and in being truthful about that, not hiding it and not
259
00:20:10,818 --> 00:20:12,309
Yeah, definitely.
260
00:20:12,309 --> 00:20:22,368
And transitioning to the future of sports betting and rhythm, tell me a little bit about
where you see the future of the industry heading in the next five years, maybe emerging
261
00:20:22,368 --> 00:20:25,860
technologies that you think are gonna have a significant impact.
262
00:20:26,261 --> 00:20:29,263
We talked a little bit about regulation or the regulatory environment.
263
00:20:29,263 --> 00:20:31,105
Where do you see the industry heading?
264
00:20:31,465 --> 00:20:38,201
Yeah, it's a super interesting time because the last five years have largely been
dominated
265
00:20:38,201 --> 00:20:42,051
the sports books trying to grow into the industry and see where things are going.
266
00:20:42,051 --> 00:20:52,111
And that was based on marketing, which is why I don't know about you, but I think I've
seen 20 ,000 of the big sports book ads on TV over the last couple of years.
267
00:20:52,111 --> 00:20:52,421
Right.
268
00:20:52,421 --> 00:20:54,261
So it was all about growth.
269
00:20:54,261 --> 00:20:58,481
And what we're seeing now is a bigger change for user engagement.
270
00:20:58,481 --> 00:20:59,428
And how do you keep
271
00:20:59,428 --> 00:21:00,568
you keep people in?
272
00:21:00,568 --> 00:21:01,658
How do you make it more fun?
273
00:21:01,658 --> 00:21:03,788
How do you have a better experience?
274
00:21:04,728 --> 00:21:07,508
So you're seeing this pivot towards.
275
00:21:09,054 --> 00:21:10,795
more engaging bet types.
276
00:21:10,795 --> 00:21:17,061
So we see it in our user base where we thought that a lot of people were going to want
full game bets, right?
277
00:21:17,061 --> 00:21:18,343
Like who wins?
278
00:21:18,343 --> 00:21:21,505
Does the score go over under the point value?
279
00:21:21,505 --> 00:21:24,386
But what we're seeing is a huge focus on props.
280
00:21:24,386 --> 00:21:31,609
People who just want to be part of the game and pull for their particular player.
281
00:21:31,609 --> 00:21:36,811
There's also a macro trend of people caring less about teams and more following players.
282
00:21:36,811 --> 00:21:46,745
As you go from watching games on TV to watching highlights on TikTok and Reels, people are
following players more than they are teams.
283
00:21:46,745 --> 00:21:51,217
So I think that's a big industry trend that we see continuing is
284
00:21:51,277 --> 00:21:58,077
People wanting to be engaged with the event by having some small stake in it.
285
00:21:58,077 --> 00:22:05,997
People may not want to put $100 down, but they'll put five bucks down on a player to do
something so that they have that cheer.
286
00:22:06,697 --> 00:22:13,545
And also seeing them focus a lot more on player level engagement than team engagement.
287
00:22:13,640 --> 00:22:22,364
Yeah definitely I mean I could imagine that with like you know TikTok in your pocket just
constantly showing you ads of your favorite player not ads but showing you you know the
288
00:22:22,364 --> 00:22:30,171
most popular videos from your favorite player you watch the game once a week or a couple
times whatever but every time you open up your phone you're getting bombarded
289
00:22:30,171 --> 00:22:33,283
and that makes sense.
290
00:22:33,537 --> 00:22:34,758
And you see it...
291
00:22:35,059 --> 00:22:44,787
Sorry, one really good highlight of that is the NBA All -Star game now has a slam dunk
contest being won by a G League player, right?
292
00:22:44,787 --> 00:22:54,685
Like, it's not even an NBA player, but because his highlights are so great and that they
have this engagement with the fan base, they're bringing in G League players to the All
293
00:22:54,685 --> 00:22:56,837
-Star game for that kind of engagement.
294
00:22:56,837 --> 00:23:02,141
So I think even the leagues are seeing this trend where player -level engagement is
sometimes
295
00:23:02,469 --> 00:23:04,530
even stronger than teams.
296
00:23:04,530 --> 00:23:05,080
Got it.
297
00:23:05,080 --> 00:23:08,604
So beyond you being in, you you and probably the team being
298
00:23:08,604 --> 00:23:18,449
the platform, tell me a little bit about some of the user experience, you know, philosophy
and how you stay close to those, the user experience.
299
00:23:18,689 --> 00:23:23,212
Yeah, user experience is at the heart of any product, right?
300
00:23:23,212 --> 00:23:27,114
Like it is, if you're not, if you're not delivering value,
301
00:23:27,834 --> 00:23:29,855
going to be gone.
302
00:23:29,855 --> 00:23:34,476
And the first thing you need to know is what is valuable to the user.
303
00:23:34,837 --> 00:23:38,329
And you have a hypothesis about that, but you need to go and test it.
304
00:23:38,329 --> 00:23:42,892
So we do continual user interviews
where we talk to not only the people
305
00:23:42,892 --> 00:23:46,232
who have been the most active on our platform and
306
00:23:46,232 --> 00:23:49,763
know our power users and engage every day.
307
00:23:49,763 --> 00:23:54,965
But it's just as important to talk to the people who decided not to subscribe, right?
308
00:23:54,965 --> 00:23:56,646
And understand why.
309
00:23:56,766 --> 00:23:58,887
Why did they download?
310
00:23:58,887 --> 00:24:01,888
So they saw that there was value potential there.
311
00:24:01,888 --> 00:24:05,949
But then during the trial, they said, this didn't meet my expectation.
312
00:24:06,081 --> 00:24:16,025
That is one of the blind spots that founding teams and executive teams often have, is they
see the value that's in their app, because they're in it a lot.
313
00:24:16,025 --> 00:24:25,269
But having conversations with people who said no, who have these set of expectations you
did not meet, understanding that can make your product even better.
314
00:24:25,269 --> 00:24:26,850
Because one of two things can happen.
315
00:24:26,850 --> 00:24:28,571
One, you can say, okay,
316
00:24:28,649 --> 00:24:30,961
they were expecting something that we're not delivering.
317
00:24:30,961 --> 00:24:33,493
It's not our value proposition.
318
00:24:33,613 --> 00:24:36,624
The question is why did they think that through your marketing?
319
00:24:36,624 --> 00:24:44,676
So you can go back and you can have an update to your marketing campaign to be more
specifically targeted at your unique value proposition.
320
00:24:44,697 --> 00:24:47,478
Or you can say, man, I thought we were doing that.
321
00:24:47,478 --> 00:24:49,578
Why didn't this person see that value?
322
00:24:49,578 --> 00:24:53,140
And then you can update your product and actually deliver value.
323
00:24:53,140 --> 00:24:57,001
So those user interviews from your most dedicated
324
00:24:57,396 --> 00:25:02,720
best users all the way to the people who open the app and say, what's this and shut it
down.
325
00:25:02,720 --> 00:25:03,221
down.
326
00:25:03,221 --> 00:25:11,967
You need to hear all that feedback and then prioritize it against what actually helps you
deliver on your unique value proposition.
327
00:25:11,967 --> 00:25:12,727
Yeah.
328
00:25:12,727 --> 00:25:22,293
What user experience feedback has been, I guess, either the most surprising or has
generated just the biggest impact to the organization?
329
00:25:22,293 --> 00:25:25,235
Is it what you already mentioned about prop bets or what would that be?
330
00:25:26,293 --> 00:25:29,164
Yeah, so I think three come to mind right away.
331
00:25:29,164 --> 00:25:32,235
We've addressed two of them and we're working on the third here.
332
00:25:32,235 --> 00:25:34,525
The first one was that load time issue, right?
333
00:25:34,525 --> 00:25:42,497
It was just a bad experience to like open it, want to see your predictions and then have
to put your phone down, get a cup of
coffee and come back and see if they're there.
334
00:25:42,497 --> 00:25:44,338
That was just bad.
335
00:25:44,338 --> 00:25:51,150
The second was the clamor for prop and player prop bets was huge.
336
00:25:51,150 --> 00:25:55,841
So we re -prioritized things to be able to get those prop bets out for...
337
00:25:56,829 --> 00:25:59,830
for both NFL and NBA last year.
338
00:26:00,070 --> 00:26:09,774
And then the third is, a core part of what we do is let people track their bets in the app
so that they can say, okay, these are the ones that my model's recommending and you can
339
00:26:09,774 --> 00:26:12,075
see your model performance over time.
340
00:26:12,638 --> 00:26:20,240
And sometimes that is a pain point for users too, because there's a couple different
clicks you have to do there.
341
00:26:20,240 --> 00:26:22,763
Right now we can only track one bet at a time.
342
00:26:22,763 --> 00:26:28,833
but if somebody wants to track all 10 recommendations, how do we make that simpler?
343
00:26:28,833 --> 00:26:38,633
So that's a big part of what we're doing is looking at that user flow, what is the
behavior we want to do, and trying to make that easier for them in fewer clicks and just
344
00:26:38,633 --> 00:26:39,474
more intuitive.
345
00:26:39,474 --> 00:26:47,547
Yeah, I would imagine a challenge also is you're hiring, you know, a tech team, a software
team, and some of them might have never placed a bet before.
346
00:26:47,547 --> 00:26:53,289
so there's kind of a onboarding transition period to really understand the user experience
and what they desire.
347
00:26:53,289 --> 00:26:56,870
What does that look like when you're, when you're building out a tech team?
348
00:26:57,130 --> 00:27:02,472
Yeah, that was a question that we had internally too, is like, do we need people who have
349
00:27:02,714 --> 00:27:08,418
bet before that know the lingo, know what's going on already, or can we train that?
350
00:27:08,418 --> 00:27:12,041
And the answer is, yeah, we'll do a little of both.
351
00:27:12,041 --> 00:27:20,708
So we have some people that either have bet personally or are interested in sports and
kind of know the language of sports already.
352
00:27:20,708 --> 00:27:27,013
And we have people who have never done any kind of athletics in their history, both on the
team.
353
00:27:27,589 --> 00:27:34,434
But what Levels says is creating a culture of that, about having the culture of everyone
gets in that a couple times a day.
354
00:27:34,434 --> 00:27:45,801
We talk about whose model is winning more than other people's and have these kind of
contests that gamify being in there and knowing exactly what's going on in the game.
355
00:27:46,021 --> 00:27:52,105
And then you build out the language and you understand what...
356
00:27:52,530 --> 00:27:56,365
has to happen from a user perspective organically.
357
00:27:56,365 --> 00:28:00,530
comes out in some of the people we had that had never done.
358
00:28:01,405 --> 00:28:11,029
kind of sports before are now some of the people who are most fluid in talking about point
spreads and payouts and the difference between American payouts and decimal payouts and
359
00:28:11,029 --> 00:28:11,755
all
360
00:28:11,755 --> 00:28:13,557
Yeah, definitely.
361
00:28:13,858 --> 00:28:15,219
It's funny just listening to you.
362
00:28:15,219 --> 00:28:19,404
just so I'm so out of it every active every day in my in my stock portfolio.
363
00:28:19,404 --> 00:28:21,906
I just wish I had something like this for that.
364
00:28:22,591 --> 00:28:25,871
And hopefully something comes out like that.
365
00:28:26,191 --> 00:28:36,351
My kind of last question here, just kind of wrapping it up, and this is, like to ask this
to every guest at the end, towards the end, is if you were to go back to your 18 year old
366
00:28:36,351 --> 00:28:46,591
self and looking at the insane career journey you've had, what would be like three pieces
of advice that you would give your 18 year old self knowing what you know today?
367
00:28:46,971 --> 00:28:51,013
Yeah, it's interesting to look at where you
368
00:28:51,013 --> 00:28:52,193
in your journey.
369
00:28:52,193 --> 00:28:58,793
I think one is I kind of have learned both from the successes and the failures.
370
00:28:58,893 --> 00:29:09,373
one of the things about I have a 17 year old son, so I know talking to someone about that
age, the one thing I know is there I would not have listened to myself talking to me at
371
00:29:09,373 --> 00:29:09,773
18.
372
00:29:09,773 --> 00:29:10,153
Right.
373
00:29:10,153 --> 00:29:15,873
So so there's always that of how much advice do you accept when you're that age?
374
00:29:15,967 --> 00:29:22,037
But a couple things that I remember were pivotal moments in my own career growth.
375
00:29:22,037 --> 00:29:35,707
One was the recognition that I moved from thinking that everybody who had been in the
industry for a decade or longer had some sort of secret knowledge of everything.
376
00:29:35,707 --> 00:29:37,515
And then I remember
377
00:29:37,515 --> 00:29:41,248
day going, actually, I think I know more about this area than that guy, right?
378
00:29:41,248 --> 00:29:45,414
So this moment of one advice would
379
00:29:45,414 --> 00:29:48,656
be don't give into imposter syndrome, right?
380
00:29:48,656 --> 00:29:49,997
Don't give into it.
381
00:29:49,997 --> 00:29:55,680
You may know as much as other people and if you don't, they'll tell you and then you will,
right?
382
00:29:55,680 --> 00:30:05,755
Like, so you use the opposite of imposter syndrome to your advantage and make claims and
see, because people will always correct you if you're wrong.
383
00:30:05,755 --> 00:30:09,427
So come on there and learn from that and grow.
384
00:30:09,427 --> 00:30:14,638
And that's the other half of that is don't be bashful about staking a claim out there in
the
385
00:30:14,638 --> 00:30:20,275
And even if you're wrong, there's never a harm in being wrong.
386
00:30:20,275 --> 00:30:22,637
harm is if you don't grow from being wrong.
387
00:30:22,637 --> 00:30:30,925
So that's a big one is they can claim, put something out there, try by experience and grow
into the person you want to be.
388
00:30:30,925 --> 00:30:33,605
Yep, definitely, no, I love that.
389
00:30:33,625 --> 00:30:36,885
Perfect, well, it was a pleasure having you on the podcast today.
390
00:30:36,885 --> 00:30:40,097
We really thank you for sharing your experiences, your wisdom.
391
00:30:40,097 --> 00:30:45,468
I can't wait to see what the future of rhythm is and we'll talk soon, thanks.
392
00:30:45,668 --> 00:30:47,609
Sounds great, thanks for having me.