The Customer Segments Hiding in Your Own Data | ICP Compilation - Episode#106

broad of an ICP, it doesn't just slow you down. It drains you. It frustrates your sales team because they're chasing the wrong logos. It produces very generic messaging that really doesn't speak to anyone. It convinces your product team to build features for customers you don't even really want or shouldn't want. The companies that do this well, they don't guess about it. They go back through every customer they've ever had, every customer they've won, lost, and they look at retention rates, uh, sales cycle conversion rates from opportunity to close one, and they really get ruthless about which segments, which ICP they should go after. If I'm talking to a founder for the first time, one of the very first questions I'm going to ask them is the same. What's your ideal customer profile? And the answers that I get range from things like all CMOs, all SMBs, small companies in tech and large companies, enterprise companies and professional services, all companies that use ServiceNow in the US, IT or security people at large regulated companies. All of those too broad, too many segments, too many buyers that they're trying to please. And I get it, I get the instinct behind this. You want the biggest TAM, the tot— the biggest total addressable market. You want to build the biggest business that you can. You think this is the way to do that. You had conversations with different types of customers or buyers who said that they need what you have, so why would you say no to anyone who's asking for your product. Here's why. If you try to sell to everyone, you really end up selling to no one effectively. And it's not just selling. Your product ends up becoming a Frankenstein product because it's serving too many different types of companies and buyers. You have an issue perfecting the post-sales motion because you're trying to make it plug and play for the smaller customers, but larger ones need a customizable, configurable product that is specific to their environment, and they need your team's help to do that. Your brand content community can't be tailored to a specific persona because you're trying to appeal to multiple, and those different personas truthfully may not even like each other. And I know this because I've lived it at every company that I've run and every company that, that I've advised. I mean, my first company, we were focused on and had amazing traction with micro SMBs, sub-10 employees. And then we landed one large account of 100 employees and thought, that's it, this is the way to go. We— this is much faster way to get to our numbers if we sell these bigger customers. Let's just go all in there. And so we started building products for them, we changed our go-to-market motion, we hired more senior reps, and honestly, we ended up alienating our SMB customers. And we never fully built the capabilities the large companies needed because there were other people in the space doing that. And then at my last company, we were selling to IT and we were great at that. And then we heard from some CISOs, they said, oh, we, we need these security features that you have, this is really cool, this is really cool. And so we decided, okay, let's go after IT and security simultaneously. But when we tried to straddle that line, we really lost what made us special to the market, being the champion for this next generation of IT people, because we always had to add things in for the security buyer in our messaging and our approach, everything. And it took an investor literally forcing us into a defined ICP exercise before I saw what a narrow customer profile actually does to a business. It closes more deals, it improves retention, it lifts pretty much every single metric that you should be caring about. And now the whole company is rowing in the same direction. And once I'd been through that myself, I started to notice it everywhere. Everyone I talked to, every founder, every CEO, every company that I have advised, I'd say without exception, has had some version of this problem. Not because founders are, stupid or unfocused, you know, it's, it's because they're trying to grow as quickly as possible. And investors are asking them, how big is your TAM, your total addressable market? They're— I'm only investing in things with huge TAM. They're having conversations with people, potential buyers, who, who are telling them, I really, really like this, I, I want this. And that's giving them new ideas for, for new markets. But a broad ICP, too broad of an ICP, it doesn't just slow you down, it drains you. It frustrates your sales team because they're chasing the wrong logos. Their, their sales, their close rate is low. It produces very generic messaging that really doesn't speak to anyone. It convinces your product team to build features for customers you don't even really want or shouldn't want. Now, I've written about this a lot, um, on Substack, on LinkedIn. And it's always about this, this idea of finding your pockets of gold. The companies that do this well, they don't guess about it. They go back through every customer they've ever had, every customer they've won, lost, and they look at retention rates, sales cycle, conversion rates from opportunity to closed one. And they really get ruthless about which segments, which ICP they should go after. So today's episode is a compilation of 7 guests that I've had on the podcast over the last 2 years, and all of them talking about ICP, how they found ICP, how they help other people find their ICP, what does it really cost to commit to an ICP, and, and what benefits they got. So let's jump right into it. We're going to start— you should really start with this, which is a diagnostic. So this is Marcus Rue, co-founder and former CEO of Guidewire. He's now an investor. And he runs offsites for his portfolio founders. And when they're offsite, he runs founders through this very simple exercise. Just write down your ICP on a whiteboard, write it out loud, speak to it in front of your peers, these other founder CEOs, and listen to what he says happens when people try it.

I ask them, it's a very simple kind of questionnaire, but of about 15 questions, and I ask them to write answers to them. And the very basic things, which are like, who are, you know, who, who is the user persona? Who is the buyer persona? What is the value proposition to each of those? In what order of priority, et cetera? These kinds of questions. But I say, just remember this one time in this setting, you don't have to promote anything. You're not selling. When you're a CEO, you're selling 99.9% of the time. You are constantly selling. You're selling to your investors, you're selling to your team, and of course you're selling to your market, right? It's so rare to have this little clearing where you say, I'm not selling anymore, I'm just thinking. And I said, I want to create a setting where we're just thinking. But get the spaghetti out of your mind and put it into something clear and coherent on a piece of paper. And now we're going to discuss it. Such a simple exercise. And yet, in my experience, there's a lot of revelation that comes out of that.

Yeah.

Right. And you just ask each other basic questions. And I try to do this in a peer-to-peer way, you know, with other CEOs talking and not me dominating the conversation. But to say, well, on the one hand, you say that, you know, that this kind of market structure is the thing that bothers you. And yet, You seem to be trying to, you seem to be confused about are you competing or cooperating with them? And you know, these kinds of questions inevitably come out, out of that very simple explication of what does this company believe and what is it doing?

I've seen that actually with some of the companies I work with, the founders, when I ask them, tell me just your ideal customer profile. Let's just start there. And like, yeah, we have it, we have it. You know, it's this. I'm like, can we start writing on the whiteboard? I never really thought about just the, the actual act of externalization, they start telling you like, this is 6 ICPs, not one.

Yeah.

And yeah, that's a lot of incoherence emerges that way.

This is 6 ICPs, not one. That might be the most common diagnosis in startup land, and the founder usually doesn't know it or realize it until they're forced to write it down and talk about it out loud in front of someone else. And that's the value of Marcus's exercise. It's not that the founders are lying when they say they have an ICP. It's that an ICP that only lives in your head, never externalized to your team, to your investors, to everyone, isn't actually one thing. It's whatever you talk to, you know, yourself into. And, and on your latest sales call, once you've externalized it and you've started to, to write this down, or at least started to go down that path, the question is, how narrow is narrow enough? And on this particular topic, I quote Tom Biocchi, former CEO of Service Channel, no exaggeration, at least once a week. Segment until it hurts. So let's hear how Tom did this.

I really am trying to emphasize is kind of segment until it hurts kind of thing. And I'll give you an example of what we had at Service Channel, which is an indication. So we had, and we went through a couple marketing departments and a couple sales organizations along the way, right? And we had a mantra that our target market was retail. Right. And that was it. Retail. So in the pipeline process and in the lead generation process, we'd get Walmart, we'd get two bakeries on the Upper West Side, right? A guy that owns a bagel shop on 79th Street, we'd get a small restaurant in Belgium, right? Because that's retail kind of thing. We'd get McDonald's, which is largely franchised, right? We get all kinds of stuff. We'd spend a tremendous amount of time vetting it, having sales calls, etc., etc., only to find out our close rate was terrible and we were letting a lot of things in the funnel that didn't really fit what our target market really— it wasn't really retail. And so we went out and tried to refine it. We ended up with something, and I'll do this from memory, right, that our target market wasn't retail. It was retail. If you retail something related to food or fashion, food or apparel, those two things, right, it had to be 100 locations or more. In North America only, primarily in the US, because that's where we had coverage. It had to be geographically distributed. We didn't want 100 locations in the same town because it wouldn't, it wouldn't adhere to our value proposition. We wanted it multi-state, multi-region, national. The more national, the better. Had to be geographically widely distributed. That's where our value proposition really kicked in. It had to be company-owned. We didn't want franchises because company-owned means centralized decision-making. We can call on somebody who can make a decision. Franchises, I'm not calling on every guy that owns 3 Subways, right? Or 3, you know, Jimmy John's. I can't do that, right? Um, it had to have centralized facilities management. Uh, a great example is Costco. You think Costco would be a great customer? Well, They have different facilities managers at every Costco that operate independently. It's not century, so you could never get someone to make a decision there, right? So that falls out. And we wanted to call on people where consistency of brand mattered. So if you think about an Apple Store retail and the consistency of brand they have, and I'll just make a contrast here, not to be disparaging, Dollar General, Right? It's a different level of brand, um, awareness kind of thing. And so we had all those filters and it turns out we did a, we did an analysis and there were 600, uh, I'm getting close to like 682 companies in the United States that fit that description. 682. And we knew the names of all 682. So long story to say, my, my, uh, my pitch to folks these days that I'm trying to give advice is don't fish in an ocean, fish in a pond and fish in a pond where you know the names of all the fish, right? So segment it down to a level of detail where you can literally name the fish in the pond. Because if I had said we're going after retail, we're going to get a, you know, 3-location church in the Netherlands, right? Versus what we really wanted, which is a Walmart or a CVS or an Apple Store or a Chipotle or a Panda Express, right? Very different.

So you hear that there, right? Tom's team didn't define their ICP with 2 or 3 attributes. They had, I think, 8 or 9 attributes there— geography, ownership structure, location count, brand tier— until the list of qualifying companies in the entire country was It was 682, and they knew every single one of them by name. That's the test that Marcus's whiteboard exercise is actually pointing towards. Can you literally name the fish in your pond? Next up is Craig Rosenberg. Craig is someone who really taught me pretty much everything I know about go-to-market, at least in the early part of my career. He's currently at Scale Venture Partners, and he works with their portfolio companies from, from a go-to-market perspective. And he spent really his whole career inside of go-to-market motions across, I don't know, hundreds, thousands of companies. And he is taking this quote unquote name the fish idea one step further because once you've got your ICP, he says, and I think in today's world this is very true, the real unlock is even the micro-segmentation inside of it.

It really comes down to figuring out your ICP, but then micro-segmenting the ICP. Because one of the issues on relevant messaging is it's hard even with AI. And, um, but if you can start by the entire go-to-market team focusing on a subset of accounts with a homogeneous use case, or the type of similarities that tie back to, uh, use case, then you can have messaging across the board that both is consistent but also is relevant to them. Also, the most important thing is, you know, these kids today that are running these micro-segments, they're data-driven. That is, it's like, yes, folks, like, can we look at accounts we've won? And that, by the way, haven't churned, which is another sort of piece to the segmentation strategy, is like, are they the right customers? And can we find lookalikes to them? And that's you know, a really simple way to think about it. And if I had to take one thing, because, you know, as you know, there are a variety of tactics and a variety of offers and a variety of things that will eventually work, but it has to start at the top. Okay. So that's one thing. Here's the one I know you're going to— well, I don't know if you're going to ask me, but I know you get all the time, which is if I'm a founder of a seed company and I'm getting going, Here's my tip. Talk to 10 people a week that are either in your assumed ICP or close to it, because you're gonna have to close these deals. And the only way you can do that, and the only way you can figure out what works, and the only way you can hear reactions is with that number. That you have the best story of all time. I don't remember how many accounts you talk to in 90 days, but that is like, so, so you guys are like, well, I don't have the data. We'll get the data.

So two things hopefully you take away there from Craig's. First, microsegmenting. Your ICP isn't one homogenous blob. It's made up of a bunch of subsegments. Second, his tactical floor for an early stage founder of 10 conversations a week inside or near your assumed ICP. Real conversations to understand what they need. Are they actually your ICP? All of that. Next, this is where it gets a little bit more complicated. Zach Smith, who ran Packet, an infrastructure company, he drew a hard line between who his users were and who his actual customers were. And this was a distinction even his own board didn't understand and couldn't comprehend for a while.

And we had a nuanced view upon that. Like we, although we targeted the developer user, our customers were the top 1,000 technology-enabled enterprises of today or the future. That's what we were looking for, which means we were not building a product that was actually for everybody. It's pretty exclusive. Why we were doing that? Because I think it was hard for us to identify, um, we felt that the, the kind of biggest, most important companies of today, or that were trying to become the biggest and most important companies of tomorrow, That actually helps you start with a two-person startup, um, uh, where technology is what at the heart of their competitive advantage. Because we had a ton of users, we wanted to create a free tier. We couldn't. Infrastructure is hard. Um, so we gave out where we could to communities, open source ecosystems, et cetera. And then we made it as friction-free as we could for new users to start adopting our value and giving us feedback. But we never were confused. Actually, I had to explain this so many times to my board. Um, they're like, look, you have like $100,000 in revenue. You've figured it out. I'm like, those are just my users who are paying to try and tell me about the thing. They're not my customers. That's not our target. They're paying. I value them. I'm going to support them. But that is not where I'm going. We've got like a multi-layer cake because I got to get to the top. Some of those, and it turns out like one of my first users was Mitchell., um, at HashiCorp and he, you know, became a great, they became a great customer and a great, you know, I'm gonna call it a future company, you know, top, top company of the world based upon technology. Um, but they were like him and Armand and like 2 people at the time. So it was not yet there. And so the confus— like sometimes this kind of like dichotomy of the user, which is adopting and testing and using and telling you versus your customer, um, which was Sprint. Um, you know, they're not the same, right? And so, um, that confused my, my, my board for a little while.

This is the part of the ICP conversation people skip the most. Your ICP is your customer, not necessarily your user. In some cases it could be both, but not necessarily. Zach had real revenue coming in from people who were never going to be his actual business. And he had the discipline to tell his own board that money didn't count the way they thought it did. That takes conviction most founders don't have when the metric in front of them looks good. If Zach's clip is about discipline with your board, this next one is about a different kind of discipline. Sean Henry co-founded Stord as a teenager, and the company's first real pivot meant doing something almost no founder wants to do or ever really does do. And that's FIRE paying customers.

I mentioned that we pivoted early on, and I'll actually talk about both because when we started, we first basically set up an Airbnb-like, uh, website for anyone with any capacity in a warehouse, storage space, whatever. And we saw this rapid deviation from kind of the mission of make a warehousing network for brands to start to compete with Amazon to people wanted to store anything and everything from movie sets to cars to all sorts of things. And we were like, okay, this is interesting, but it's not exactly the use case we thought. That was very guerrilla, kind of like we were on LinkedIn, we were in Facebook groups, we were on Craigslist, we were posting this website everywhere, just attacking, messaging people. And then we said, okay, let's fire all these customers. We kind of got this kind of mixed bag of use cases, and let's just go all in on our ICP of these But we quite literally fired, uh, literally fired, literally fired a few hundred thousand dollars a year in revenue at 18 years old. And we spent the next 6 months so mad at ourselves. Like, what did we just do? This new model of trying to get the perfect customers is so much harder. Uh, but it basically became from like, let's be a marketplace and just see what transactions happen to let's go sell and let's go direct to direct customers. And that became a, an approach of quite literally selling. We got into an accelerator around this time called Dynamo. It was supply chain focused. And I say that just because they had started a freight brokerage and sold it to Coyote, became part of UPS eventually. And a freight brokerage is very much a manual brokerage, like a smile dial, call someone, say, do you need a truck? Call a truck and say, I have a load for you. And so these guys were basically like, cold calling and cold emailing is going to be your life. And, uh, they gave us the advice of basically lock Jacob and I in a room all day, every day, just making hundreds of calls and hundreds of emails. And so Jacob and I literally set quotas for ourselves, uh, 50+ calls a day each, 500+ emails a day each. So probably even higher than that. And then it became, well, where do we get all these leads and how do we find this? And we got super scrappy because we had no money. There was a government website that was the kind of a record of every single ship coming into the US, every container on it. They all have a bill of lading. And we basically built a scraper tool, went through hundreds of millions of documents, got hundreds of thousands of emails and phone numbers from every one of those documents, and just started aggressively outreaching. It went from 500+ manual emails a day to tens of thousands of emails and calls, uh, emails a day and, and hundreds of calls. And so that's really where we ended up meeting our first customer. We were not the company that, uh, could go sell to everyone in our Y Combinator batch because we're just a software company that all of our friends could use. Our first true customer for this business model was a publicly traded solar panel manufacturing company, publicly traded in China.

A few hundred thousand dollars a year gone on purpose by an 18-year-old because it didn't fit the profile they actually wanted to build a company around. That's the single most concrete example I have of what's narrowing your ICP, the real costs of that in real dollars, real time, before you even have a cushion to do that. Most founders talk about focus like this in the abstract. They're, they're not actually doing it. So Sean's story is about having the courage to cut the wrong customer. Melanie Fillet, co-founder and CEO at Spekit, her story is a little— is about a more subtler trap: investing heavily in people who love your product but were never actually going to buy it.

Don't treat all potential buyers, your customers, of your products equal. You know, be intentional around like, what are the communities that you really want to get more deeply embedded where there will be more loyalty? The irony is like a lot of those people never bought the products, right? Many of them were actually too low level in the organization to have the kind of budget. And I learned that the painful way. Be, be aware of that too. Like, is the community the champions that can use the tool, or is it the buyers, Right? Early on I'd say we invested way too much in the champion community. What we've learned the hard way is like you need to be a big enough company that you really have the pain of enablement. Early on we sold to customers that were too small, right? They had like, you know, 30, 40 employees. They were starting to feel, and like they're just so messy and unstructured that they have the pain, but they don't really have someone that's gonna be able to like manage it and be effective and sell. We really start minimum at that like $100, $150 range and then go upwards to that like $500 or so.

Champions are not customers. That's the line to remember. A champion can love your product, amplify your content, and never sign a check because they don't have the budget or authority. And Melanie's other filter matters just as much. Company size isn't about who can use you. It's about who actually feels the pain enough to prioritize fixing it. Small companies have the symptom— symptom. They don't always have the urgency or the resources to treat it. Last one, and it's the clip that really pulls this whole episode together, because Rohit Balani, co-founder and CEO of Fishme, makes the point that your ICP isn't a one-time decision. It moves as your product moves, as the market changes, and it has to stay tied to how the entire company operates, not just sales.

I think it's very important to constantly be thinking about what your ICP is. It, it, over time, as we evolved products, of course the ICP evolved as well, uh, as we entered different markets, you know, we went downstream as well. So we were purely enterprise, we started serving mid-market, but yes, at any given point in time, we always had our eye on the ball in regards to the ICP. I think it's extremely important to get alignment across product marketing, the direction of product, engineers understanding why they're doing something. I think to the success of the business, at the end of the day, look, there's probably 100 good ideas for any given product, and you have to be able to whittle it down to say, what are the 3 things we're doing over the next 3 to 6 months? And how do you get there? Because you have to put that customer focus on and say what really matters to our customer. And if you don't know who your ideal customer is, how are you gonna do that?

The last line is the whole episode in one sentence. If you don't know who your ideal customer profile is, how are you gonna decide anything? Not your roadmap, not your messaging, not your hiring. Rohit's other point is just as important. His ICP moved downstream over time from purely enterprise into mid-market, and he treated that evolution as expected, not as a failure of the original definition. Define it, live by it, revisit it on purpose. 7 different vantage points, and these are just the clips I used here. I have many more, and the same thread runs through all of them. If there's one thing I want you to take away from this episode, it's that a narrow ICP isn't a slide in a deck. It's a decision you have to keep making, especially in the moments when saying no is expensive. Go build your customer data set, go talk to your best reps, find your pockets of gold, and then have the discipline to actually fish in that pond, not the ocean. That's it for this one. As always, thanks for listening to Not Another CEO podcast. If you like this episode, please share it out with your networks.