WEBVTT
1
00:00:00.120 --> 00:00:15.900
[upbeat music] Welcome back to the Playbook Podcast from Agency Focus. I'm John Suter. And I'm Bart Miraz. Bart, any, uh, opening thoughts today?
2
00:00:16.940 --> 00:00:30.680
A priest walks into the bar. [laughs] Ow. [laughs] All right, let's kick it off. Bart, I wanna read you something. Accenture has added tens of thousands of AI and data people, as we've been saying.
3
00:00:31.040 --> 00:00:44.100
EY put more than a billion dollars behind its AI platform. McKinsey has put AI in the center of how it delivers work, and the phrase all of them are using is AI-native delivery. D- [snorts] Yes. Okay.
4
00:00:44.220 --> 00:00:54.820
So that, [laughs] that laugh tells me you have, uh, some sort of opinion. [clears throat] I have very simple opinion. Take away the language and read what they are actually describing.
5
00:00:55.180 --> 00:01:07.220
Small teams embedded inside the client, building a working version fast and iterate it. Measure the outcome instead of [clears throat] handing over the document and leaving. And what do you call that?
6
00:01:08.360 --> 00:01:23.340
That is called working at an agency. [laughs] All right. Say more. [laughs] I ran one for 13 years. We did exactly that. Uh, we did not [clears throat] need a billion-dollar announcement to do it.
7
00:01:23.500 --> 00:01:32.220
So when the agency world reads these press releases, the right response is not to panic. It is, "Welcome." Welcome.
8
00:01:32.700 --> 00:01:41.040
That's what we're gonna get into today, why Big Consultancy just discovered the agency model, what it means for a small shop, and why Bart thinks this is good news.
9
00:01:41.400 --> 00:01:50.280
Let's start with the language, because I think a lot of agency owners read AI-native delivery and feel like they are already behind. What's actually inside that phrase? Uh, three things.
10
00:01:50.540 --> 00:02:03.780
Uh, rapid prototyping, practitioners embedded inside the client's team, and working towards measurable outcome. That is it. The AI part is real, but the model underneath it, it's not, like, new.
11
00:02:04.200 --> 00:02:16.400
All right, so the delivery model is the old part, and AI is the new part. Is that what you're saying? Correct. And here's a piece people miss. Small agencies did not build that model because they were ahead of the trend.
12
00:02:17.160 --> 00:02:31.520
Uh, they built it because they had no choice. What do you mean they had no choice? When you are a 10-person shop going after the same work as 10,000-person firm, you cannot win on size. It-- You cannot win on brand.
13
00:02:32.160 --> 00:02:49.720
You win on specificity and speed. So you embed, you iterate, you ship, you show up on Monday with working version instead of a slide deck. You do not have to have the budget for theater.
14
00:02:49.720 --> 00:02:59.760
[laughs] All right, theater kid, let me push on that. Isn't that a little defensive? These firms obviously have a lot more money, more people, more data than any agency listening to this.
15
00:02:59.920 --> 00:03:09.720
"We did it first" doesn't pay the bills. It is not defensive. It's an advantage, and it's a specific one. AI did not create this model.
16
00:03:09.880 --> 00:03:24.940
AI plugged into it, and it plugged in first at shops that were already in a room when the client's problem showed up. If you are embedded, you see the problem on Tuesday, and you test a fix on Wednesday.
17
00:03:25.580 --> 00:03:34.460
If you're writing a capability deck, you see the problem in quarter reviews. Right. So proximity's actually the head start. Proximity, proximity is the head start.
18
00:03:34.980 --> 00:03:44.860
The big firms are now spending billions to buy proximity. Small agencies already have it. Well, then why does it feel like the agencies are the ones playing catch-up? Why, because they're smaller?
19
00:03:45.720 --> 00:04:00.420
Eh, because the big firms are better at announcing things. [laughs] That is a real skill. They will win with enormous contracts. They will write the white papers. They will host events where CMO is not along.
20
00:04:00.940 --> 00:04:12.800
I'm not saying they will fail. I'm saying we should call rebranding what it is. All right, it's me nodding along. All right, so let's talk about the hiring numbers because they're big.
21
00:04:13.100 --> 00:04:27.700
Tens of thousands of AI people at just one firm. If I'm a 12-person shop, that number really scares me, shouldn't it? Eh, no. H-Headcount is not a capability. Tens of thousands of new hires sounds like capability.
22
00:04:28.380 --> 00:04:38.320
In practice, it is an onboarding problem. All right. Tell me more about that. Many of those people c- uh, come from adjacent roles inside of large organizations.
23
00:04:38.780 --> 00:04:53.480
Good people, but becoming deployable on real client work at the level the market expects takes time. Meanwhile, the agency that has spent the last few years using AI on live client work has a learning curve.
24
00:04:54.440 --> 00:05:04.400
Um, you cannot hire that learning curve away. All right. So what is the thing that clients are actually short on? Because it's not tools. Everybody has the tools now. Correct.
25
00:05:04.700 --> 00:05:17.780
Every client has access to the same models or tools or pick, pickaxes or [laughs] Shovels... or, or shovels or genes or... [laughs] What, what they are short on is the judgment.
26
00:05:18.020 --> 00:05:30.540
They need someone who can walk in the room and say, "This use case in your organization with your data. Um, and here's what works. Here's what we fail.
27
00:05:31.020 --> 00:05:40.760
Um, and here's why the thing the vendor is promising will look great in a demo and break three months after they go live." Uh, so where does that come from?
28
00:05:42.240 --> 00:05:53.940
From shipping the same category of things 40 times and learning what breaks. It does not come from hiring announcements. Hmm. All right. You also make the case that judgment and relationships are connected.
29
00:05:54.460 --> 00:06:05.456
That one surprised me a little bit. Judgment sounds like expertise. Relationship sounds like just being more likable. Yeah, they're not the same, but they grow together. Judgment needs context.
30
00:06:06.056 --> 00:06:11.056
Knowing the CEO's stated priority and actual priority is, uh, just two different things.
31
00:06:11.636 --> 00:06:21.736
Knowing which VP will quietly kill a product, knowing the client's internal politics before you walk into the meeting, you only learn that through close contact over months or years. All right.
32
00:06:21.796 --> 00:06:32.736
So is there data on that, or is this just expertise talking? Uh, both. There's a survey from, uh, Setup that asked marketing leaders why they ended agency relationship.
33
00:06:34.056 --> 00:06:47.676
44% said the agency never understood their business. Agencies barely list that as a reason. So the gap is real, and most agencies don't see that in themselves. Right, which cuts both ways.
34
00:06:47.716 --> 00:07:03.136
That's a warning for small agencies too, isn't it? Yes, it is. Being small does not give you judgment automatically. Being close does. The agencies with durable AI practice are not the ones with best capability decks.
35
00:07:03.596 --> 00:07:15.996
They're the ones where the founder picked up the phone when something broke. That one's hard to scale. That's the part that's hard to scale. That is the... Boy, my friend, as the cost of intelligence keeps falling,
36
00:07:17.076 --> 00:07:28.456
this is what gains value. The trust between a practitioner and a client does not transfer to a new account manager. The big firms cannot manufacture it at scale. Yeah, yeah.
37
00:07:29.156 --> 00:07:41.376
Okay, you keep saying, "Show up on Monday with a working version." Give me a time when that actually happened at Sumo Heavy. Um, back in the early days. Early days too. Uh, yeah, we're in early days or, or things.
38
00:07:41.376 --> 00:07:51.176
Yeah, I think, so I think film was still black and white. Early days. Yeah. [laughs] We did it in chisels. All the computer- We had the big, we had the big buggy whip account. You remember? Um, yeah.
39
00:07:51.216 --> 00:08:04.196
We actually, uh, worked with a client that hired a big consulting firm, and it's your typical hired a big consulting firm, but the people who they send are youngsters straight out of school that have no knowledge or things.
40
00:08:04.276 --> 00:08:11.456
Mm. And we came in with a very nimble team. And back in the day, we were trying to, you know, we were still young and still doing the, the work, but,
41
00:08:12.776 --> 00:08:17.176
you know, we were going against a big, big, big, huge A, uh, uh, consulting firm.
42
00:08:18.356 --> 00:08:27.856
And what happened was one of the departments, [laughs] this was the fun part, um, we built some tools out, and then the, the consulting firm kind of took over.
43
00:08:28.696 --> 00:08:39.976
Um, but they didn't have any judgment or background or any of it, so they started doing, like, prescribed thing, you know, building stuff, and it was just off. Okay. And where were you in all this?
44
00:08:41.016 --> 00:08:48.896
Well, we kind of got sidelined a little bit because the big, you know, corporation took over, and we were there building other things with another department.
45
00:08:49.016 --> 00:09:02.316
But the department we worked with beforehand was like, "Can you come in and fix it? Can you look at this?" And it was exactly that scenario. We looked at it on Tuesday, and by Wednesday [laughs] it was flowing already.
46
00:09:02.796 --> 00:09:10.476
You know what I mean? Like, we had the background and the knowledge, understood the company, and we're a little tiny place that did it, and we were nimble about it because we jumped in and did it. All right.
47
00:09:10.596 --> 00:09:23.436
So what happened after that? Uh, we stayed there for another three years, [laughs] doing a lot of work for a while. They got purchased. Wow. Okay. And the conversations change.
48
00:09:23.576 --> 00:09:29.896
Before that, we're the vendor executing someone's strategy because that strategy was given us by big company.
49
00:09:30.416 --> 00:09:40.496
After that, w- you know, the, the one department or multiple departments in that client started calling us because they understood that we are a lot nimbler- Okay... and could do it and go faster. All right.
50
00:09:40.596 --> 00:09:49.356
So the consulting firm, in essence, wasn't wrong. They were just far away. Exactly. The strategy was fine for the most part.
51
00:09:49.356 --> 00:10:02.136
[laughs] The judgment about the what would break the company was not in a deck because it could not be. It was in the people who had been there every week, and that happens a lot.
52
00:10:02.576 --> 00:10:11.396
A deck, if you don't have proper documentation, it's not gonna tell you anything. You have to be in there nitty-gritting. Right. So is there a cost to working that way?
53
00:10:11.476 --> 00:10:21.615
Because it sounds great, but also sounds like the founder in every single room. Yes, there was a cost. For a lot of these years, I was the last set of eyes on everything that left my agency.
54
00:10:22.196 --> 00:10:38.556
I told myself it was about quality. It was also about control. So the lesson is not founder must be in every room. The lesson is that the closeness is the asset. Your job is to build a team that can carry it. Very good.
55
00:10:38.736 --> 00:10:48.736
All right. So someone listening that runs a $3 million shop, they buy all this. What do they actually do differently on Monday? We always use Monday as an example, but what do they do Monday?
56
00:10:48.736 --> 00:10:58.236
[laughs] Well, I think Friday. They can do, they can do it Thursday if they want. I don't know. Or Friday. Don't do it Friday night. Friday night? [laughs] No. Think about it Sunday, Sunday evening around 4:00. Oh, man.
57
00:10:58.236 --> 00:11:08.076
[laughs] Uh, that's, that's me every day. First, stop competing on the terms. You cannot outstaff Accenture. You cannot go against the nobody gets fired for hiring IBM.
58
00:11:08.536 --> 00:11:23.716
Steal the fricking thing, which annoys the hell out of me. Yeah. I want that burned to the ground. Um, you will not outsource them. You will not write a better capability presentation than BCG or, um, McKinsey.
59
00:11:24.176 --> 00:11:32.956
Let's do McKinsey- Right... 'cause that's the top standard, right? So don't even try it. Well, what, what- Um, and, and, and before we go further- And there's always an and...
60
00:11:33.376 --> 00:11:44.316
and there's an and because it is my favorite thing lately is you can just do things- You can just-... and be faster at it... do things. Exactly right. All right. So then what do you actually compete on?
61
00:11:45.276 --> 00:11:52.776
You could just do things. [laughs] You could just do things. A position so specific that you're not competing with them at all.
62
00:11:53.576 --> 00:12:00.196
Um, this client, this client type, this problem, this method, this outcome, four things, one sentence.
63
00:12:00.736 --> 00:12:13.972
But also- When I say you can do things, it's the idea of you can change your billing terms and how you deliver the work or what you deliver, and be creative about it every day. Hmm.
64
00:12:14.952 --> 00:12:28.172
All right, so every founder hears niche down and gets nervous they're leaving money on the table. What do you say to that? [laughs] I twitch. You twitch.
65
00:12:28.512 --> 00:12:43.672
[laughs] The data says the opposite. Promethean Research surveyed agency owners to share. The average agency grew about 7 to 7.5%, with around 13% net margin.
66
00:12:44.252 --> 00:12:58.572
The agencies that narrowed their service line grew 13% and hit 30% margins. Wow. Focus is not sacrifice. It is a strategy. Okay, so who's that specific client? Where do they live? Where can I get in touch?
67
00:12:59.732 --> 00:13:10.492
Uh, everywhere. I don't know. [laughs] Mostly the mid-market. Big firms are going upmarket. Their model is built for enterprise budget and long timelines.
68
00:13:10.712 --> 00:13:23.552
Right That leaves companies that cannot absorb consulting overhead, cannot wait a year for a capability to come online, and I... and need someone who shows up with real answers. That is your ICP. I like that.
69
00:13:23.652 --> 00:13:35.972
All right, so you make a big prediction in the article that I wanna push on. You say that big firms will start buying boutique AI practices. Sounds like a little wishful thinking from someone who has run a boutique.
70
00:13:35.972 --> 00:13:45.612
[laughs] Yeah. I understand why it sounds that way, but look at the history. Digital transformation, agile delivery, UX as a practice.
71
00:13:46.012 --> 00:13:56.832
Every time that big firms announcement their, their plans, hire talent, and then acquire the shops that have built their real expertise, they can build it themselves at their scale.
72
00:13:56.932 --> 00:14:06.692
The math on how long it takes to earn credibility does not work. Yeah. All right, so who, who actually gets bought? Not the ones with the best marketing.
73
00:14:06.952 --> 00:14:20.332
The ones with specific expertise, real client relationship, and a track record measured in production deployments, not pilots. Right? So the small sel- stealthy firms that are specializing in, you know,
74
00:14:21.832 --> 00:14:26.212
this level, this level, this level. Yeah. Yeah. And, and the buyer's already pricing it.
75
00:14:26.872 --> 00:14:41.432
Um, FE International valuation data says an agency that has truly put in, put AI into its delivery, not just played with ChatGPT, um, can add one to two turns on its multiple.
76
00:14:42.472 --> 00:14:54.952
And if you are building something specific, making it work, uh, for real clients, you are building something someone will want to own. Okay, so, uh, just slapping on "We, we use AI" doesn't work anymore.
77
00:14:55.032 --> 00:15:05.652
You said production, not pilots. Say more about that, 'cause we talked about pilots before and, and I think that's a real headline. It is. Uh, most organizations have pilots.
78
00:15:05.932 --> 00:15:21.472
Uh, far fewer have anything running in production. Digital Applied surveyed 250 agencies. 41% now have at least one AI agent live. The median return is about three times the spend.
79
00:15:22.212 --> 00:15:36.392
But the bottom quarter is getting about 70 cents back on the dollar. Well, okay, so a quarter of the agencies that adopted are losing money on it? Yeah. And the time blocker they named was not AI.
80
00:15:37.052 --> 00:15:49.172
Uh, it was testing, knowing whether the thing actually works. [laughs] That is judgment again. The gap between pilot and production is where, you know, the work is.
81
00:15:49.752 --> 00:16:03.172
It is where agencies get called back and where their relationship shifts from vendor to an advisor. Hmm. So the move is not putting AI native on the homepage. No, not the homepage. Not the pitch deck.
82
00:16:03.672 --> 00:16:17.712
Not trying to get on a list of next, you know, to Accenture. The move is to get every engagement to a working production deployment. The pilot era is just over. Okay. And we also did a podcast on that two weeks ago.
83
00:16:17.812 --> 00:16:24.552
Mm-hmm. So if you wanna look that up, we'll also link that back in the show notes. Bart, if a founder walks away from this episode with one thing, what do you think it should be?
84
00:16:25.572 --> 00:16:41.512
That consulting land grab is not a threat. It is a clarification. It tells you exactly where you win. Be specific. Stay close to the client. Ship to production. Let the big firms have press releases.
85
00:16:41.512 --> 00:16:52.692
[laughs] And their theater. And their theater, theater kids. All right, theater kids. All right, for the founder who still feels left behind, what do you say? Look at your last five engagement.
86
00:16:53.052 --> 00:17:05.872
How many are you running in production right now with number attached to them? If the answer is most of them, you're not being the big consulting wave. You were doing this work before the press release.
87
00:17:06.352 --> 00:17:16.792
Okay, so what if the answer is not many? Then that is conversation to have. That is what we do. Okay. Uh, which is a good place to mention it.
88
00:17:16.872 --> 00:17:26.712
If you run an agency and growth feels like it's held together with duct tape, string, and Band-Aids, or you have a specific position in there somewhere and you can't quite name it, book an insight call with us.
89
00:17:26.992 --> 00:17:40.032
30 minutes, zero pressure. The link is in the show notes, or go to, get your pencils ready, tidycal.com/agencyfocus. Bring the five engagement with you. [laughs] Bring them with you. That, that's the cost of admission.
90
00:17:40.092 --> 00:17:46.672
Bring those five engagements. Uh, it's coffee. Coffee is usually the cost of admission. I think that's what we do. Cost of admission is coffee. We even say that in the, in the disclaimer.
91
00:17:46.792 --> 00:17:57.792
All right, bring the five engages. Bart, as always, thank you. Thank you, sir. All right, we'll see you next week. The Playbook Podcast is produced by Agency Focus at 14:59 PM. New episodes every week.
92
00:17:57.952 --> 00:18:14.252
Watch and subscribe on YouTube or on your favorite podcast app. And remember, sign up for our newsletter at playbook.agencyfocus.xyz. New issues every Tuesday morning. See ya. See ya. Wouldn't wanna be ya.
93
00:18:14.252 --> 00:18:20.312
[outro jingle]
1
00:00:00.120 --> 00:00:15.900
[upbeat music] Welcome back to the Playbook Podcast from Agency Focus. I'm John Suter. And I'm Bart Miraz. Bart, any, uh, opening thoughts today?
2
00:00:16.940 --> 00:00:30.680
A priest walks into the bar. [laughs] Ow. [laughs] All right, let's kick it off. Bart, I wanna read you something. Accenture has added tens of thousands of AI and data people, as we've been saying.
3
00:00:31.040 --> 00:00:44.100
EY put more than a billion dollars behind its AI platform. McKinsey has put AI in the center of how it delivers work, and the phrase all of them are using is AI-native delivery. D- [snorts] Yes. Okay.
4
00:00:44.220 --> 00:00:54.820
So that, [laughs] that laugh tells me you have, uh, some sort of opinion. [clears throat] I have very simple opinion. Take away the language and read what they are actually describing.
5
00:00:55.180 --> 00:01:07.220
Small teams embedded inside the client, building a working version fast and iterate it. Measure the outcome instead of [clears throat] handing over the document and leaving. And what do you call that?
6
00:01:08.360 --> 00:01:23.340
That is called working at an agency. [laughs] All right. Say more. [laughs] I ran one for 13 years. We did exactly that. Uh, we did not [clears throat] need a billion-dollar announcement to do it.
7
00:01:23.500 --> 00:01:32.220
So when the agency world reads these press releases, the right response is not to panic. It is, "Welcome." Welcome.
8
00:01:32.700 --> 00:01:41.040
That's what we're gonna get into today, why Big Consultancy just discovered the agency model, what it means for a small shop, and why Bart thinks this is good news.
9
00:01:41.400 --> 00:01:50.280
Let's start with the language, because I think a lot of agency owners read AI-native delivery and feel like they are already behind. What's actually inside that phrase? Uh, three things.
10
00:01:50.540 --> 00:02:03.780
Uh, rapid prototyping, practitioners embedded inside the client's team, and working towards measurable outcome. That is it. The AI part is real, but the model underneath it, it's not, like, new.
11
00:02:04.200 --> 00:02:16.400
All right, so the delivery model is the old part, and AI is the new part. Is that what you're saying? Correct. And here's a piece people miss. Small agencies did not build that model because they were ahead of the trend.
12
00:02:17.160 --> 00:02:31.520
Uh, they built it because they had no choice. What do you mean they had no choice? When you are a 10-person shop going after the same work as 10,000-person firm, you cannot win on size. It-- You cannot win on brand.
13
00:02:32.160 --> 00:02:49.720
You win on specificity and speed. So you embed, you iterate, you ship, you show up on Monday with working version instead of a slide deck. You do not have to have the budget for theater.
14
00:02:49.720 --> 00:02:59.760
[laughs] All right, theater kid, let me push on that. Isn't that a little defensive? These firms obviously have a lot more money, more people, more data than any agency listening to this.
15
00:02:59.920 --> 00:03:09.720
"We did it first" doesn't pay the bills. It is not defensive. It's an advantage, and it's a specific one. AI did not create this model.
16
00:03:09.880 --> 00:03:24.940
AI plugged into it, and it plugged in first at shops that were already in a room when the client's problem showed up. If you are embedded, you see the problem on Tuesday, and you test a fix on Wednesday.
17
00:03:25.580 --> 00:03:34.460
If you're writing a capability deck, you see the problem in quarter reviews. Right. So proximity's actually the head start. Proximity, proximity is the head start.
18
00:03:34.980 --> 00:03:44.860
The big firms are now spending billions to buy proximity. Small agencies already have it. Well, then why does it feel like the agencies are the ones playing catch-up? Why, because they're smaller?
19
00:03:45.720 --> 00:04:00.420
Eh, because the big firms are better at announcing things. [laughs] That is a real skill. They will win with enormous contracts. They will write the white papers. They will host events where CMO is not along.
20
00:04:00.940 --> 00:04:12.800
I'm not saying they will fail. I'm saying we should call rebranding what it is. All right, it's me nodding along. All right, so let's talk about the hiring numbers because they're big.
21
00:04:13.100 --> 00:04:27.700
Tens of thousands of AI people at just one firm. If I'm a 12-person shop, that number really scares me, shouldn't it? Eh, no. H-Headcount is not a capability. Tens of thousands of new hires sounds like capability.
22
00:04:28.380 --> 00:04:38.320
In practice, it is an onboarding problem. All right. Tell me more about that. Many of those people c- uh, come from adjacent roles inside of large organizations.
23
00:04:38.780 --> 00:04:53.480
Good people, but becoming deployable on real client work at the level the market expects takes time. Meanwhile, the agency that has spent the last few years using AI on live client work has a learning curve.
24
00:04:54.440 --> 00:05:04.400
Um, you cannot hire that learning curve away. All right. So what is the thing that clients are actually short on? Because it's not tools. Everybody has the tools now. Correct.
25
00:05:04.700 --> 00:05:17.780
Every client has access to the same models or tools or pick, pickaxes or [laughs] Shovels... or, or shovels or genes or... [laughs] What, what they are short on is the judgment.
26
00:05:18.020 --> 00:05:30.540
They need someone who can walk in the room and say, "This use case in your organization with your data. Um, and here's what works. Here's what we fail.
27
00:05:31.020 --> 00:05:40.760
Um, and here's why the thing the vendor is promising will look great in a demo and break three months after they go live." Uh, so where does that come from?
28
00:05:42.240 --> 00:05:53.940
From shipping the same category of things 40 times and learning what breaks. It does not come from hiring announcements. Hmm. All right. You also make the case that judgment and relationships are connected.
29
00:05:54.460 --> 00:06:05.456
That one surprised me a little bit. Judgment sounds like expertise. Relationship sounds like just being more likable. Yeah, they're not the same, but they grow together. Judgment needs context.
30
00:06:06.056 --> 00:06:11.056
Knowing the CEO's stated priority and actual priority is, uh, just two different things.
31
00:06:11.636 --> 00:06:21.736
Knowing which VP will quietly kill a product, knowing the client's internal politics before you walk into the meeting, you only learn that through close contact over months or years. All right.
32
00:06:21.796 --> 00:06:32.736
So is there data on that, or is this just expertise talking? Uh, both. There's a survey from, uh, Setup that asked marketing leaders why they ended agency relationship.
33
00:06:34.056 --> 00:06:47.676
44% said the agency never understood their business. Agencies barely list that as a reason. So the gap is real, and most agencies don't see that in themselves. Right, which cuts both ways.
34
00:06:47.716 --> 00:07:03.136
That's a warning for small agencies too, isn't it? Yes, it is. Being small does not give you judgment automatically. Being close does. The agencies with durable AI practice are not the ones with best capability decks.
35
00:07:03.596 --> 00:07:15.996
They're the ones where the founder picked up the phone when something broke. That one's hard to scale. That's the part that's hard to scale. That is the... Boy, my friend, as the cost of intelligence keeps falling,
36
00:07:17.076 --> 00:07:28.456
this is what gains value. The trust between a practitioner and a client does not transfer to a new account manager. The big firms cannot manufacture it at scale. Yeah, yeah.
37
00:07:29.156 --> 00:07:41.376
Okay, you keep saying, "Show up on Monday with a working version." Give me a time when that actually happened at Sumo Heavy. Um, back in the early days. Early days too. Uh, yeah, we're in early days or, or things.
38
00:07:41.376 --> 00:07:51.176
Yeah, I think, so I think film was still black and white. Early days. Yeah. [laughs] We did it in chisels. All the computer- We had the big, we had the big buggy whip account. You remember? Um, yeah.
39
00:07:51.216 --> 00:08:04.196
We actually, uh, worked with a client that hired a big consulting firm, and it's your typical hired a big consulting firm, but the people who they send are youngsters straight out of school that have no knowledge or things.
40
00:08:04.276 --> 00:08:11.456
Mm. And we came in with a very nimble team. And back in the day, we were trying to, you know, we were still young and still doing the, the work, but,
41
00:08:12.776 --> 00:08:17.176
you know, we were going against a big, big, big, huge A, uh, uh, consulting firm.
42
00:08:18.356 --> 00:08:27.856
And what happened was one of the departments, [laughs] this was the fun part, um, we built some tools out, and then the, the consulting firm kind of took over.
43
00:08:28.696 --> 00:08:39.976
Um, but they didn't have any judgment or background or any of it, so they started doing, like, prescribed thing, you know, building stuff, and it was just off. Okay. And where were you in all this?
44
00:08:41.016 --> 00:08:48.896
Well, we kind of got sidelined a little bit because the big, you know, corporation took over, and we were there building other things with another department.
45
00:08:49.016 --> 00:09:02.316
But the department we worked with beforehand was like, "Can you come in and fix it? Can you look at this?" And it was exactly that scenario. We looked at it on Tuesday, and by Wednesday [laughs] it was flowing already.
46
00:09:02.796 --> 00:09:10.476
You know what I mean? Like, we had the background and the knowledge, understood the company, and we're a little tiny place that did it, and we were nimble about it because we jumped in and did it. All right.
47
00:09:10.596 --> 00:09:23.436
So what happened after that? Uh, we stayed there for another three years, [laughs] doing a lot of work for a while. They got purchased. Wow. Okay. And the conversations change.
48
00:09:23.576 --> 00:09:29.896
Before that, we're the vendor executing someone's strategy because that strategy was given us by big company.
49
00:09:30.416 --> 00:09:40.496
After that, w- you know, the, the one department or multiple departments in that client started calling us because they understood that we are a lot nimbler- Okay... and could do it and go faster. All right.
50
00:09:40.596 --> 00:09:49.356
So the consulting firm, in essence, wasn't wrong. They were just far away. Exactly. The strategy was fine for the most part.
51
00:09:49.356 --> 00:10:02.136
[laughs] The judgment about the what would break the company was not in a deck because it could not be. It was in the people who had been there every week, and that happens a lot.
52
00:10:02.576 --> 00:10:11.396
A deck, if you don't have proper documentation, it's not gonna tell you anything. You have to be in there nitty-gritting. Right. So is there a cost to working that way?
53
00:10:11.476 --> 00:10:21.615
Because it sounds great, but also sounds like the founder in every single room. Yes, there was a cost. For a lot of these years, I was the last set of eyes on everything that left my agency.
54
00:10:22.196 --> 00:10:38.556
I told myself it was about quality. It was also about control. So the lesson is not founder must be in every room. The lesson is that the closeness is the asset. Your job is to build a team that can carry it. Very good.
55
00:10:38.736 --> 00:10:48.736
All right. So someone listening that runs a $3 million shop, they buy all this. What do they actually do differently on Monday? We always use Monday as an example, but what do they do Monday?
56
00:10:48.736 --> 00:10:58.236
[laughs] Well, I think Friday. They can do, they can do it Thursday if they want. I don't know. Or Friday. Don't do it Friday night. Friday night? [laughs] No. Think about it Sunday, Sunday evening around 4:00. Oh, man.
57
00:10:58.236 --> 00:11:08.076
[laughs] Uh, that's, that's me every day. First, stop competing on the terms. You cannot outstaff Accenture. You cannot go against the nobody gets fired for hiring IBM.
58
00:11:08.536 --> 00:11:23.716
Steal the fricking thing, which annoys the hell out of me. Yeah. I want that burned to the ground. Um, you will not outsource them. You will not write a better capability presentation than BCG or, um, McKinsey.
59
00:11:24.176 --> 00:11:32.956
Let's do McKinsey- Right... 'cause that's the top standard, right? So don't even try it. Well, what, what- Um, and, and, and before we go further- And there's always an and...
60
00:11:33.376 --> 00:11:44.316
and there's an and because it is my favorite thing lately is you can just do things- You can just-... and be faster at it... do things. Exactly right. All right. So then what do you actually compete on?
61
00:11:45.276 --> 00:11:52.776
You could just do things. [laughs] You could just do things. A position so specific that you're not competing with them at all.
62
00:11:53.576 --> 00:12:00.196
Um, this client, this client type, this problem, this method, this outcome, four things, one sentence.
63
00:12:00.736 --> 00:12:13.972
But also- When I say you can do things, it's the idea of you can change your billing terms and how you deliver the work or what you deliver, and be creative about it every day. Hmm.
64
00:12:14.952 --> 00:12:28.172
All right, so every founder hears niche down and gets nervous they're leaving money on the table. What do you say to that? [laughs] I twitch. You twitch.
65
00:12:28.512 --> 00:12:43.672
[laughs] The data says the opposite. Promethean Research surveyed agency owners to share. The average agency grew about 7 to 7.5%, with around 13% net margin.
66
00:12:44.252 --> 00:12:58.572
The agencies that narrowed their service line grew 13% and hit 30% margins. Wow. Focus is not sacrifice. It is a strategy. Okay, so who's that specific client? Where do they live? Where can I get in touch?
67
00:12:59.732 --> 00:13:10.492
Uh, everywhere. I don't know. [laughs] Mostly the mid-market. Big firms are going upmarket. Their model is built for enterprise budget and long timelines.
68
00:13:10.712 --> 00:13:23.552
Right That leaves companies that cannot absorb consulting overhead, cannot wait a year for a capability to come online, and I... and need someone who shows up with real answers. That is your ICP. I like that.
69
00:13:23.652 --> 00:13:35.972
All right, so you make a big prediction in the article that I wanna push on. You say that big firms will start buying boutique AI practices. Sounds like a little wishful thinking from someone who has run a boutique.
70
00:13:35.972 --> 00:13:45.612
[laughs] Yeah. I understand why it sounds that way, but look at the history. Digital transformation, agile delivery, UX as a practice.
71
00:13:46.012 --> 00:13:56.832
Every time that big firms announcement their, their plans, hire talent, and then acquire the shops that have built their real expertise, they can build it themselves at their scale.
72
00:13:56.932 --> 00:14:06.692
The math on how long it takes to earn credibility does not work. Yeah. All right, so who, who actually gets bought? Not the ones with the best marketing.
73
00:14:06.952 --> 00:14:20.332
The ones with specific expertise, real client relationship, and a track record measured in production deployments, not pilots. Right? So the small sel- stealthy firms that are specializing in, you know,
74
00:14:21.832 --> 00:14:26.212
this level, this level, this level. Yeah. Yeah. And, and the buyer's already pricing it.
75
00:14:26.872 --> 00:14:41.432
Um, FE International valuation data says an agency that has truly put in, put AI into its delivery, not just played with ChatGPT, um, can add one to two turns on its multiple.
76
00:14:42.472 --> 00:14:54.952
And if you are building something specific, making it work, uh, for real clients, you are building something someone will want to own. Okay, so, uh, just slapping on "We, we use AI" doesn't work anymore.
77
00:14:55.032 --> 00:15:05.652
You said production, not pilots. Say more about that, 'cause we talked about pilots before and, and I think that's a real headline. It is. Uh, most organizations have pilots.
78
00:15:05.932 --> 00:15:21.472
Uh, far fewer have anything running in production. Digital Applied surveyed 250 agencies. 41% now have at least one AI agent live. The median return is about three times the spend.
79
00:15:22.212 --> 00:15:36.392
But the bottom quarter is getting about 70 cents back on the dollar. Well, okay, so a quarter of the agencies that adopted are losing money on it? Yeah. And the time blocker they named was not AI.
80
00:15:37.052 --> 00:15:49.172
Uh, it was testing, knowing whether the thing actually works. [laughs] That is judgment again. The gap between pilot and production is where, you know, the work is.
81
00:15:49.752 --> 00:16:03.172
It is where agencies get called back and where their relationship shifts from vendor to an advisor. Hmm. So the move is not putting AI native on the homepage. No, not the homepage. Not the pitch deck.
82
00:16:03.672 --> 00:16:17.712
Not trying to get on a list of next, you know, to Accenture. The move is to get every engagement to a working production deployment. The pilot era is just over. Okay. And we also did a podcast on that two weeks ago.
83
00:16:17.812 --> 00:16:24.552
Mm-hmm. So if you wanna look that up, we'll also link that back in the show notes. Bart, if a founder walks away from this episode with one thing, what do you think it should be?
84
00:16:25.572 --> 00:16:41.512
That consulting land grab is not a threat. It is a clarification. It tells you exactly where you win. Be specific. Stay close to the client. Ship to production. Let the big firms have press releases.
85
00:16:41.512 --> 00:16:52.692
[laughs] And their theater. And their theater, theater kids. All right, theater kids. All right, for the founder who still feels left behind, what do you say? Look at your last five engagement.
86
00:16:53.052 --> 00:17:05.872
How many are you running in production right now with number attached to them? If the answer is most of them, you're not being the big consulting wave. You were doing this work before the press release.
87
00:17:06.352 --> 00:17:16.792
Okay, so what if the answer is not many? Then that is conversation to have. That is what we do. Okay. Uh, which is a good place to mention it.
88
00:17:16.872 --> 00:17:26.712
If you run an agency and growth feels like it's held together with duct tape, string, and Band-Aids, or you have a specific position in there somewhere and you can't quite name it, book an insight call with us.
89
00:17:26.992 --> 00:17:40.032
30 minutes, zero pressure. The link is in the show notes, or go to, get your pencils ready, tidycal.com/agencyfocus. Bring the five engagement with you. [laughs] Bring them with you. That, that's the cost of admission.
90
00:17:40.092 --> 00:17:46.672
Bring those five engagements. Uh, it's coffee. Coffee is usually the cost of admission. I think that's what we do. Cost of admission is coffee. We even say that in the, in the disclaimer.
91
00:17:46.792 --> 00:17:57.792
All right, bring the five engages. Bart, as always, thank you. Thank you, sir. All right, we'll see you next week. The Playbook Podcast is produced by Agency Focus at 14:59 PM. New episodes every week.
92
00:17:57.952 --> 00:18:14.252
Watch and subscribe on YouTube or on your favorite podcast app. And remember, sign up for our newsletter at playbook.agencyfocus.xyz. New issues every Tuesday morning. See ya. See ya. Wouldn't wanna be ya.
93
00:18:14.252 --> 00:18:20.312
[outro jingle]
