← Always Be Testing
Loading summary
Transcript79 lines
- [00:00]
A
Foreign. Hello, hello, hello. Welcome to another episode of the Always Be Testing podcast. I'm your host Ty DeGrange and I'm absolutely thrilled to have Guillaume Caban on today. Guillaume, how you doing?
- [00:18]
B
Hey Ty, thanks for having me. I'm doing great.
- [00:20]
A
Absolutely. It's awesome to have you. Guillaume is notorious in B2B SaaS growth. He is the CEO and founder of Hypergrowth Partners. Hypergrowth Partners helps venture backed SaaS companies scale and grow. It's a collective, it's an investor, it's a growth advisor. And he's had a litany of amazing experience and growth from Zapier to Reddit to G2 Crowd to drift to ramp to many. So you're in for a treat today.
- [00:48]
B
That's awesome. Let's go.
- [00:49]
A
Yes, this is Intercontinental. It's calling in from Europe, so it's big time. One of my favorite memories, I think it was when I was introduced to you was when you were speaking in SF back in the glory days or the early days of quote unquote growth, talking about the year at Dev Guild, talking about instrumenting Madcudu and Drift at the time and some really interesting things that were, I think novel in B2B and B2B I think was sort of exploding in its own way. A lot has changed since then. But do you recall those days?
- [01:24]
B
October 2016 is the date you're looking for. So I'm afraid the 10th anniversary of that event is this year. I just want to say time flies. Oh my God. I'm sorry.
- [01:37]
A
Wow. I feel old. And I am now in Texas and you are now in France.
- [01:44]
B
Yeah, 10 years. And I was at segment, I was the view of growth at segment back then. And I mean none of the things that we know today, not even speaking about AI, but not even like GTM engineering, not to be like even growth was like a very, very small niche kind of knowledge and function. And I was part of the few people trying to like make it exist and battle for its existence and to prove that there was value there. But I absolutely remember those days and maybe me and a bunch of the other people who are involved back then, like we succeeded at it.
- [02:22]
A
Yeah, absolutely. One of the things we talked about was kind of this conundrum of stage and do you have a growth team or not? Right. You think about SaaS, you think about varying stages of funding. What are some of the like scale ups, if you will, that that don't have a growth team versus those that do. How do you kind of differentiate? What are you seeing right now?
- [02:45]
B
In the world of BTB SaaS, which is the one that I know well, most don't now it's very different from what I've heard and some of the things that I know in B2C it's quite different but I have much less of a knowledge there. In B2B SaaS most scale ups do not have growth team. Those that do in this example is basically all the ones where I have worked office bias there. And aside from that Rippling also has a great growth team. So there's, there's a, there's a, there's a couple. But growth teams tend to be very popular around the high growth phase called like late series A, series B, series C and then kind of like, you know, disappears after that.
- [03:33]
A
For people following the home that are not as familiar with growth like you and I and you in particular, can you just give a distinction for folks about growth teams versus a performance marketing team or a marketing team?
- [03:46]
B
Yeah, and I understand this confusion because some people use it interchangeably for good or bad reasons. So here's my definition or how I like to think about it and I think it's now become probably the Most say Accepted 1. A growth team is a group of generally fairly technical people, which means there are engineers, they are marketers and copywriters and this is how they work. They're going to build quasi products, some backend, some front end. They're going to have a product like way of approaching and say building those products on a marketing service area. So a website campaign app on email and they're going to be measured like a sales team in pipeline. So you take the overlap of those three little bubbles, this Venn diagram of sorts and at the center that's what a great growth team is. And the reason why that differentiation is important is well if you just take the engineers that build a product then you have a product team and they're building product. Obviously it has to be applied to a marketing service area. Otherwise like you're just building product. If you don't have the pipeline then basically what you have is a bad marketing team. And so it's important to have those three aspects. The last thing which is not defined by, by those aspects yet is speed. It is my conviction that the growth team succeeds by going extremely fast through a set of experiments that are non obvious. If the things they're trying are obvious and you know they're going to win, it's not an experiment. You should ship those things as high quality as possible because you know they're going to work, you have conviction and that's great. And that's for the marketing team. It's going to be polished, it's going to be nice. The growth team is there to produce elements, campaigns, quasi products where you have a low conviction of success. And because you have low conviction of success, your competitors also have the same thing. They're probably not going to try that. And so the success of a growth team happens when you find unlikely outcomes. Alpha is what we call it wins that happen to be extremely successful when you did not know the outcome beforehand. That's, that's success. That's important to understand and that's how I've operated over the past 15 years of doing it.
- [06:37]
A
When something's sort of on that verge or how are you, what are you seeing? Kind of tip the scales to say this is worth testing versus this is not. Obviously there's a lot of methodologies or a lot of ranking methodologies. You've probably seen them all or a number of them. What are some things that tip the scale for you with teams that are debating that internally.
- [06:58]
B
So generally most things are worth testing unless it's obviously a terrible idea. Here's how I approach it. So I approach, I have a two week sprint and so every two weeks with the team I sit down and we look at all the new ideas. And I looked and asked them to have this, this mindset which I wrote a substack about it a couple of weeks back. It's called like be Crazy creative and rigorously scientific. Okay, so we couple first principles. We look at what do we, what do we believe to be true? The number of potential prospects that are out there on the TAM full addressable market, the price we sell the product at, the kind of customers we sell to, what's the cost of acquisition that we're ready to accept? There's a couple of like ground rules here. And then we look at any number of potential problems. Some leaks in the funnel, drop offs in the funnel, some audience that is not converting as well as others. And that's where we try to have a very, very creative, and I call it crazy creative mindset. And then we run a hypothesis that's the rigorously scientific. If the hypothesis holds to be true, let's say that we send a much better copy or we fix this funnel leak. What is the consequence? What is the lift? And is this lift statistically significant? Can we measure it with confidence and is it worth the effort to do it? Because I am not saying is this a good or bad idea, Is it Going to work or not? That's not what I'm saying. I'm saying like if it works, is it even worth it? And can we even measure it? And that kills more than half of the ideas. More than half of the ideas. They're legitimately, possibly good ideas. You just don't have the end. You just don't have enough people going through that potential funnel path. So that kills half of the ideas, the other half, and that's my expertise, the decade plus that I've spent doing it is they're over complex, right? They're thinking of like a big beautiful project to fix in a big problem. And in doing that, the hypothesis has many unknowns, many, many unknowns. And you just know how math works. You can't solve multiple unknowns in the same equation. It's not going to work here. And so you got to break it down, got to break it down in separate, smaller experiments. Okay, is this the right audience? Is this the right message? Is this the right offer? And again, split those things apart so you can measure them accurately and then if you want, you can bring them together. Okay, I'll give you an example for the audience that is very commonplace and very traditional. I had a client, not going to name the client, obviously had a client that wanted, the marketing team, wanted to do a very typical white paper, was kind of PDF behind a gate form of some sort. And there was some, they wanted to do it with an agency. So there was like a decent spend of like 15 or 20k back then to do it with all the research, good paper. And then they were going to do some promo on it and the whole, the whole shebang, right. And basically put it on the homebridge with a banner or the little banner, the top and the bottom. And so, okay, I'm not going to debate whether that's a maybe, maybe that's what the audience wants. I doubt, but maybe. And I looked at the hypothesis and the hypothesis goes as well, this is what we're spending. Sure. And this is the funnel performance, step by step. And I told them, just skip almost all the experiments, put the banner on the homepage, put the form, don't do the white paper yet. And tell me, do you see the conversion rate on the form? Are people actually attracted to this potential white paper? And they told me, well, they're not going to get the email, they're not going to care. You're going to have 15 people send them to it later. It's okay, people are not going to sue you because they didn't get the white paper they signed for lo and behold, I was right. The outcome was a tenth of what they had forecasted. And so the experiment was not worth it. So that's a good example of breaking down a big expensive, long project into the important unknown. What's the click through rate? What's the conversion? What's the appetite for your audience for this piece of content?
- [11:30]
A
I love that.
- [11:30]
B
And that mindset can be applied across the board. Like I'm the horrible boss that always comes down and like slashes experiments into chunks and I always go, what's the unknown? What's the N? Is this stat sig? Right. Have we seen some of those things before? Right. And when the team starts having this framework, it means we can test many, many things very fast. If you look at Ramp, for example, like a couple of years ago, the team, the growth team was like what, a dozen people and they were shipping 200 ish small experiments per quarter. So like it's a big number but like there's like 60 working days in a, work in, in a quarter ish. And so that's also not that much. It's also not that much. It's a lot, but it's not that much. But you ship 200 ish decent well thought experiments or quarter. Some of them are going to work, some of them are going to be real alpha that none of a competition has.
- [12:34]
A
I like what you said and that's.
- [12:37]
B
That'S why Ramp won.
- [12:38]
A
That's huge. I like what you said about small versus decent. That seems like a big one from what I've gathered from folks and seen and I think you are saying pretty clearly here, like you've got to have the small ones. Okay, you're okay with some small ones, but I would imagine you're saying you don't want all small ones.
- [12:59]
B
No, for sure. And generally I let people build on top of like I break down something. We have some learnings. And so because those learnings are rigorous and so now I know, hey, this audience shows clear preference for this message, this offer, this pricing, right. When all the things controlled for, I can now apply that to a lot of other potential experiments. A lot of different channels. All right. On which I'm like, hey, I already know the outcome of the message now can test different channels, I can test different like so a lot of things I can build on compounding. Okay. And that's the issue. Yeah, it absolutely. And many teams out, they're just like fishing around. They like they put four or five unknowns together and they're shooting in the dark. Right. And maybe they get lucky, but I, I don't like luck. I don't like being lucky.
- [13:52]
A
Well, I love, I love there's a de risking element to what you're talking about and you're semi self critiquing, like, hey, I'm coming in as kind of the quote unquote bad guy to like poke holes in the assumptions of these hypotheses. And that, that's a lifesaver and a de risker, it sounds like, Right?
- [14:08]
B
Absolutely. And if you look at a team after a couple of quarters working with me, like I don't need to be the bad guy, like they know why the framework works and they know they, they it's self sustainable.
- [14:23]
A
Yeah. No, in some way you're, you're coming in to make it more sustainable for everyone. In terms of, like you talked about speed and experimentation. Is there like a framework in your mind around experiments on a particular cadence or a particular level that you feel like is interesting for the, for the audience around. Okay. You should probably be about this number of them or should probably thinking about the volume and you kind of touched on it.
- [14:48]
B
But yeah, I think ramp is an extreme case and people shouldn't be like, you know, stressed about that because like most teams are not going to have a dozen people working on growth. This is not going to happen. And so I think at the early stage, got to be realistic. Most of the people who could be listening to this, founders or the marketing executives. So you're going to have one person running growth. And if you have one person running growth, you got to be, I think, realistic with expectations here. And I would probably go for something around a handful of experiments every week or every two weeks. And those are like not too big. And then as that growth team gets some wins and you get to like a two, three people, I think you can start having a more rigorous cadence of every two weeks and review every month of the outcomes. And sure, the number of experiments go down, it's not linear, but you're doing a bit more meaty experiments.
- [15:53]
A
Yeah, yeah.
- [15:55]
B
And so that's kind of how I think about it. At the latest, at the scale up stage, like reamp, the experiments per person actually drop because there's a lot more maintenance. But that's okay, that's okay. As long as you're shipping more experiments than all of your competitors, that's fine. You're going to win.
- [16:13]
A
Yeah. And for those trying to kind of break out and increased data learnings and the enrichment of that and like the impact is there some, is there some things that you counsel people on, around. Okay, how do you break out of maybe a slow experimentation cadence or maybe tests that are not as insightful?
- [16:33]
B
Break out of the mindset of a slowness situation?
- [16:35]
A
Yeah, like the mindset are teams that are just not getting as much. I think it's mindset, but also number of tests in a period of time or the learnings and impact of said tests typically.
- [16:47]
B
And I'm not always successful. Like I had an engagement recently where I failed to change the culture. And I can tell you what, why it happens. Generally these are marketers that are being repurposed into growth teams and it's not, it's not, it's not a personal issue, it's a team culture issue and expectations issue. Okay. Those people have been trained to expect a risk to reward ratio where if they fail, they risk of being fired. That's what happens to a marketer. If you're a marketing leader and you fail three, four times in a row, you're out. Makes sense. You're a terrible marketing leader. But if you're a growth person, your failure rate should be north of 50%. 50, 60, 70%. Time and time again, I go in, I get a deal done with a company with a CEO founder, and they assign me a small group of two or three people and I'm like, okay, sounds great. And then the people, what do they do? They de risk all the experiments because they're afraid of failing. De risking means they do more research, they pop it up, they make it better, they think about it, they improve the copy, they improve the design and, and so it gets unbearably slow. And at that point I'm like, well, there's so much sunk cost here that now this has to succeed. You got no other choice. And you're talking like one experiment a month, maybe sometimes less. I worked with a team decently, large scale. They couldn't get outbound like outbound. Outbound emails started in a quarter. They just kept going back to the copy and the audience and revisiting the strategy and they just couldn't. And I said, just ship the damn emails, just get them out. We'll learn, we'll see by the reply rates. Anyways, those people are not customers. And if they don't care and it goes through spam, we'll learn something. They're afraid of the outcomes. They wanted to have the best outcomes. And so what happened is that they trimmed down the prospect list from tens of thousands to thousands, from thousands to hundreds. And they ended up with a seven series email to 20 prospects. And I'm like, what are we learning here? Sure, you might be creating leads. Those emails are great. And this is the best of the best of the quality. I'm sure you're going to get a handful of good, decent leads. You're going to create pipeline. There's nothing to be learned. I can't scale anything from that. And that is the best example I have of me failing, of a company failing at growth culture.
- [19:36]
A
That's a big one.
- [19:38]
B
And it's a big company. It's a company you've all heard of. It's like over a thousand employees. It's in tech. We all heard about them and they have a marketing team and yada, yada.
- [19:48]
A
Yada, the whole thing I don't even know so far as it's a hypothesis. But maybe the question for you is there, there it seems like there's not a lot of growth teams, if you really get down to it. And is that how much is culture play into that. It seems like it's a big one.
- [20:05]
B
So going back to what I said here about the risk to reward ratio and what people. People are humans are humans. They react to like incentives and fear. Right. And the stick in their carrots. Right. In other ways, this has to come from the top. This has to come to the fact from the founder. If the founder or CEO depending on scale, isn't creating a safe place for those people to succeed and take risk, it's not going to happen. It's just not. And that is why most scale ups do no longer do not have a growth team is because the risk to reward ratio has shifted and the person in charge no longer supports taking that amount of risk. Okay? That's basic. And every time when founders come to me, founders come to me and say, hey G, you're the CMO of Reamp. Yep, sure thing. Like, great. I want the same outcome. Can you help us? And I walk them through, like, are you ready for this? Like, are you sure that you want a growth team? Are you sure that you want like both the good and the bad? Yeah, yeah, yeah. The growth rate is great. Like, absolutely. I'll support it. All right. But they don't. But they don't. That's the thing. Like, they want the muscles, but they don't want the weight. Lift the weights. That's the thing. They want to skip the leg day here. So the way that I express it to them is to get the alpha to get those wins. You're going to have to have some failures and what a 70% fail rate means. Just to be very clear, it means that on a small growth team shipping what, five experiments per sprint, 10 experiments per sprint, you could go multiple months without a single win and still support their work. Are you ready for that? And is the rest of the team ready for that? Are you going to support them and say they are doing the right thing, they're on track instead of just shifting their priorities? Think about it and get back to me. I can tell you that most don't get back to me because they're not ready for that. But that is the reality. I can't make magic happen. All right. What I can do is tell them this is a framework that does work. If you're ready for it.
- [22:24]
A
Yeah, it's amazing. On the positive side of where this can go, is there an experiment in your career that looked small in the outset but ended up being just a massive change in trajectory for the business?
- [22:38]
B
Yeah, I think a lot of them eventually work because I've built, I cheat a bit in the sense that I obviously bring the wins from one company to another because I just build upon the successes. So if you look at Outbound, a bit more than half of Reapp's pipeline comes from Outbound. Talking about a company that has publicly said they're doing more than a billion in revenue. So that's massive. It just like accounts for typical close rates. You're talking of multiple billions of pipeline being created through Outbound.
- [23:10]
A
Yeah.
- [23:10]
B
So that's the probably the biggest win. And sure, everybody does. Everybody does Outbound. But to do it successfully at the scale of ramp and to crush competition like Brex was crushed, it takes a growth team working hard on it. And to be clear, like Reamp has multiple engineers, like real engineers working full time on supporting the outbound program. It's not just like, hey, we bought something off the shelf and sending emails. There's like multiple engineers in New York at New York rates, plus copywriters, plus program managers. Like supporting your team of like nine DSDRs all in house built. It is very custom and it works extremely well.
- [23:56]
A
I love it. I love it. You talked about culture previously. When you think about helping structure the relationship between growth and sales. What are some ways that you, you talked about SDRs, obviously with the RAMP example, like dial in speed and process in a way that balances the two.
- [24:15]
B
This is always age old question of like where SDRS should live. But I think I resolved it in, in a slightly different way. I like to have SLAs service level agreements with my, with my partners, in this case the sales team, which means that I will commit to something and they will commit something back to me by bi directional agreements. Okay. We agree on a common lead scoring model that we both like. I will commit to a number of leads at that quality which you could translate into a pipeline dollar figure. All right. Because at that point, like if you're comfortable with like the, the lead scoring might, might as well assign like a pipeline fear to each of those deals. So I commit to that and they commit to, to a, a speed of like reaching out and closing. Okay. And so together we can commit to revenue.
- [25:07]
A
Love it.
- [25:08]
B
Most companies don't do that. Most companies, the marketing team commits to leads but without quality and their sales team doesn't commit to like a service level agreement. And there's even more granularity to it generally. Like when I at the size of ramp, there's like a distribution of qualities of leads because I can't just have the AAAs. That's not how the world works. And so I say, hey, like okay, I can w those much of AAA's and those much of like AABS and that on. And I think that we look at the close rates. I think to maximize like revenue and team effort, the triple is you should reach out in like 30 minutes, like one hour and the other ones and you have. And so like, and if we do that all those things with the best people reaching through those things and right distribution, then we agree together and we do that per quarter. We agree to it and it's the best plan because then I have a really good partnership and I know who's accountable for what. Is my team failing at delivering the right distribution of the quality or are they failing to do things and I'm not seeing them? Sometimes it's me. But at least we're not in endless debates about like, why are we not hitting revenue? I can go back to my team, say, hey guys, we committed to this. Let's pull off some magic. Or if we have done our job and they're not calling, I go to the VP sales like, dude, like I'm on vacation now. Like, you got your good leads. You're not calling them. I'm out. Call me back.
- [26:34]
A
It's such a good reminder for it sounds simple. It's not easy to do, but it's doable. And I think it's such a great lesson that can be applied to so many areas of the business and cross functional teams. It's a great reminder to people to set those up and refer back to them. You've had obviously a host of amazing experiences. What's a belief that you held 10 years ago? We're thinking about how old we are and how long we've been doing this stuff that you no longer believe now.
- [27:03]
B
I thought that by now we would no longer have SDRs. A couple years ago when we saw all those AI agents for SDRs coming out, I was like, yeah, it's going to improve. And by 2014, 2015, that job function is out.
- [27:19]
A
You were thinking it back then. Wow.
- [27:21]
B
Yeah, yeah, for sure. Let's start with 2024.
- [27:24]
A
Okay. That feels. I feel better now.
- [27:26]
B
I thought so. Sorry. When we saw that in 2022. Yeah, the decade got me wrong. When you saw AI stuff in 2022, 2023, and we saw all of those AI tools for agents come out around 2023, I thought that by 2024 the job function would be out. And it's not the case. It's clearly not the case. And when I look at the companies, I look at which of the companies that I talk to, the details of has a fully automated AI agent for SDRs done. None of. Wow, there are none. I have not seen a BTP company successfully roll out an SDR AI agent and remove the humans from the function entirely. So that is a surprise. And I think it says something about, yes, there's a huge space of progress, but no, it's not working as well as.
- [28:19]
A
That's really insightful on the converse end. Are there areas that you have found to be very much automated, very much AI driven, very much software eating that segment of the flow?
- [28:32]
B
Yeah, content, content, content was. And there's been ebbs and flows when content. I've been a bit in the SEO side of like some. There were some errors of good quality, some errors of low quality content for the purpose of like SEO AEO and all of that. It's like there's no future there for humans. Humans built workflows, oversee workflows and slightly improve it. But do you know somebody who is tasked with a host for that company and doesn't hit an LLM once? It looks like a blank page in Google Docs is like, okay, I'm on it. See me in like two hours, I'll have the doc ready. It doesn't hit like LLM once. No.
- [29:20]
A
You're definitely going to utilize it.
- [29:22]
B
No. Yeah. And they're probably just going to build a workflow in Air Ops and get it done. All right.
- [29:28]
A
On the sdr. Flipping back to that, I mean, maybe it's obvious, but what is. Is there a particular piece that you think is just why behind it hasn't gotten there yet.
- [29:40]
B
I think the risk profile of the AI agents is lower than the SDRs because you're quick to blame the AI agent LLM workflow when it's like crap. And so the risk tolerance is lower, the risk profile is lower. SDRs, they're young folks out of college, they've got nothing to lose. Today's today, tomorrow's another day. And so they're going to take quite a bit more risk. That's one second. The tone is just not yet perfect. You try to do things at scale. Somebody like you, somebody like me, people in tech can still get a sense of humanness and we recognize the writing of LLM inherently and so it's not great. You got to remember that the only reason why an anonymous person you reaching out to is maybe going to consider reading Your email, maybe 1, 2, 3% is because there's some amount of reciprocity, a bit, some amount of curiosity, but it's mostly reciprocity and respect for the fact that you're human and that you've put some effort in trying to reach out to them. And it only works in again, low single digit percentages. But that's why it works. There's no other reason. All right, they're not doing it as an act of love for humanity. That's not why they're reading your emails. It's just reciprocity. And so you gotta get that perception, that feeling in the first couple lines. And that's really hard to do without at scale, without some really weird outcomes from the element which is like, whoa, no, that's.
- [31:25]
A
No, no, you talk, I love it.
- [31:29]
B
And it's going to happen. But yet for now it's still hard.
- [31:34]
A
Yeah, hopefully we can kind of shepherd the AI tools to our disposal in a good way. Speaking of the growth teams, with true engineers, you think about growth teams and you talked a little bit about those with real engineers, quote, unquote. How has that thought evolved and what are your thinking around the importance of engineers talent on a growth team?
- [31:58]
B
I think most companies no longer need real engineers on the growth teams these days. It was very important at the point in time where the growth team would manage a lot of the Martech stack and the tech stack and they would make sure the data would come from the website and all the traffic data and all the form data into analytics platforms and HubSpot and whatnot and all the cron jobs and so they all have backend work and, and then you wanted to enrich the data with clearbrit and you wanted to push it out to some outbound platform and maybe some ads platform and there was maybe using some segments and stuff like this. There's some pretty complex stuff going on there. And then you wanted to probably build some landing pages in a semi automated way. Great. Okay. These days there's tools for all of that. There's really good tools and building in house doesn't make sense for those things. Right. And. And you can like maybe even two years ago you could just like leverage a GTM engineer for that leveraging to just like connect those tools. Like in 2026. I'm really debating whether you just can't just like vibe code that stuff yourself. I'm seeing some platforms which are like really awesome and like even like people who think nadan is complicated, like there are tools that are coming out like code words, for example, which is like in LLM entry point into something which is like a workflow platform like any den. And so you just like speak to it and write to it and it works, it builds a workflow for you. And so do you still need engineers? Definitely less and maybe not at all. What you need is creative people. You need bigger scientific people that can understand the math, can understand the logic. But what's more important is the creativity, the understanding of human psychology. Why is it going to work? What's going to make people tick and how can I leverage that principle in, in my content, in my prompts? Right. That's important. The engineering less so today.
- [34:02]
A
That's amazing. Yeah, that's amazing. It's a, it's a kind of harkens back to your substack article and the combination of those two of super creativity and very rigorous science.
- [34:13]
B
Yeah, I think both are difficult. Both are very difficult. And what people don't realize is they're almost opposites in the personalities maybe, obviously. And so having a growth leader that can do both things perfectly is very difficult and extremely rare. You're looking for somebody who is very creative to the point of craziness and will think things out of the box that your competition will not think. And then they're going to approach those things with a scientific method and look at the numbers, do a little maybe some SQL requests, look at the data, understand the math, look at the. That's tough in one human being. That's a lot to ask for. They do exist, but it's not common.
- [35:01]
A
Yeah. I feel like that could be its own future growth leader traits and the future of that skill set and that personality trait that I'm sure you run into a lot.
- [35:15]
B
Absolutely. Those are the traits you're looking for for a growth leader. Those are absolutely the traits you're looking for. Somebody who can, like, eventually have those two aspects, but also get a sense of what all the hacks, what are the strategies that they've learned at an early stage, which can still work at a scale up. All right. Most people fail at that.
- [35:39]
A
Love it. Coming down the home stretch here, we've dropped some amazing knowledge. What's your biggest learning for 2025 influencers?
- [35:50]
B
I didn't bet on it a few years ago because it's a thing that comes from B2C. And I saw it, like, slowly come in. People asked me, I was like, no, no, no, no. I was at Ramp. And people were like, hey, do you think you could use influencers to, like, bring CFOs to ramp? And I was like, did you just say CFOs and influencers in the same sentence? Is that what you did? And lo and behold, it's working. I have to admit, my mistake. And RAMP hired a person full time to do that. And it is absolutely profitable and working. And I'm like, okay, that's interesting, let's start. And it's also working. And I'm like, well, there's something there. And if I try to rationalize it, I think it is the counterweave of outbound and personalization at scale. Well, because people trust less what's coming on from anonymous people, because the content quality is great, but it's all automated, so reciprocity is going down. They're looking for social proof, which, sure, like, there's friends, there's your colleagues, but that's a small circle. Eventually people will see a human being that they trust. So micro influencers, and so they trust that. And so I think it's working better because REST is not working as well. Maybe I'm wrong. That's just my hypothesis.
- [37:12]
A
No, I think there's something really to that. And it's. You probably not shocked, but for us, it's such a huge thesis, and I think there is such a trust gap. That's been something that I've talked a lot about and observed as well. And I think that's why you're seeing this clamoring for third parties, recognizable micro influencers of various flavors and quality and authenticity levels that are real enough that those buyers are then Then attracted and it's working enough. So definitely something excited to talk more about and, and uncover. But that's a great learning for, for 2025 and could not agree more. I know you've done a ton of talks, you've done a ton of writing, so people know a lot about Yu Gi. Oh, but what's something that maybe people don't know about you or you want to share with the audience?
- [37:59]
B
Yeah, funny, I'm seeing that in my, in my home and like I'm a technologist, I love technology to like to the extreme for sure in all aspects of my business and personal life. And what's interesting is I see like people like my wife for example, like same age, so like 42, like she's starting to have like a negative reaction towards technology. It's getting too much. It's too much for her. It's like, it's too much. And I was thinking about it and I'm seeing it happen to more and more people outside of our little tech bubble and I wonder whether it's not something we're going to see become more commonplace where some people just, they just can't jump on this wave. Right. And I'm sure, I'm sure that's true, you know, every technological changer and like the parents for those that still have them, like some of them just didn't get like Web 2.0 or something like that. But I think with the prevalence and the importance of this new wave with AI, the people that are going to left behind are going to be left so far behind, it's going to be like a breaking moment in society. There's going to be some people who are going to be, just think about it. Some people are going to be building AI agents and automating parts of our lives. And some people are going to be like, no, LLMs are not. Like, GPT is not for me. I'm out. Is it going to be two different. It's not the same human race anymore. This is two different. You could be living on two different planets at this point. And think about it this way. When is the last time that you took an important decision without writing a prompt or querying an LLM? When's the last time that happened? When's the last time for you, Ty?
- [39:42]
A
I mean, yeah, maybe. I'm going to guess maybe a year. I don't know.
- [39:47]
B
I mean for me it's like, it's like two years.
- [39:49]
A
Yeah, I went back to the adoption end of the story.
- [39:53]
B
Sure, sure.
- [39:54]
A
But like that's Crazy. Two years, that's crazy.
- [39:56]
B
It's important. Think about it that way. I haven't made an important decision in life. Consider the important decision whether it's a purchase or like something that's meaningful for me without like consult, like writing something in LLM to like validate or invalidate my choice.
- [40:12]
A
Yeah, and there's pro. There's definitely some that I'm not consulting at all now. But I'm not. But for most of them, for me, I would say.
- [40:21]
B
And so I'm not saying it's good or bad that should be debated. I'm just saying it is crafting me into a different profile personality than the people who haven't gone there and are rejecting to do that. Again, not judging at all. I'm just saying we're starting to be two different. Like, yeah, we're splitting apart. We're splitting apart like. And society is going to split apart.
- [40:45]
A
Whoa, that's a mind blower. That's a great mic dropper moment. I think the other thing it makes me think of is past generations. And again, we could say there's an exception to this. But I think about from horse to car to PC, there's a percent human involvement that keeps going down. I think that's as humans. I imagine that's gotta be scary and hard for people. It's not just adopting a new technology. It's for some people this is a big shift away from the human involvement. And I think as social creatures, as caveman DNA, that's got to be hard for just people to wrap their heads around. Regardless of how.
- [41:31]
B
Where you are, there's two ways to read it. There's two ways to read it. One way to read it is I'm being lazy. I'm delegating my important choices to an AI. And there's a really good MIT article on that which shows that the brain waves and whatnot, there's laziness happening by delegating all the choices. The other view is I'm doing more with LLM, a lot more than I couldn't do before. And what I'm thinking is what's going on with me both. I am at the same time intellectually lazy because I am delegating things which I could think of and I could work on that I am not because it's good enough. And I'm also doing more because a lot of things I would not have done, I would have given up entirely that I can do now. So both are happening for me.
- [42:20]
A
I would agree with that a lot. And I think that there's, there's a little silver lining of. I think there's a lot of hope for me generally in this that, hey, you get to focus and work and work your brain on the things that you want to. And two, hopefully we can compound together and accomplish more as a society and individuals. But we, we shall see. Amazing.
- [42:42]
B
Exactly. That's, I think, a great place to wrap it up. That's like deep, deep thoughts.
- [42:46]
A
Yeah, we'll just leave that one hanging in the ether for people. Guillaume, you're the man, the Growth B2B expert in the world. As I said earlier, it's true. And it's been an absolute pleasure to have you on. Thanks for joining me.