The ICP Blueprint

Drew: Hello, Renegade Marketers! If this is your first time listening, welcome. If you're a regular listener, welcome back.

You're about to listen to an expert huddle, where experts share their insights into topics of critical importance to our flocking awesome community, CMO Huddles. In this episode, Drake Lenhan of Sitecore gets into what it takes to operationalize an ideal customer profile. He explores how to define where you truly win, how to get sales, product, and marketing aligned around it, and how to turn ICP into a practical guide for smarter go-to-market decisions. If you like what you hear, please subscribe to the podcast and leave a review. You'll be supporting our quest to be the number one B2B marketing podcast. All right, let's dive in.

Narrator: Welcome to Renegade Marketers Unite, possibly the best weekly podcast for CMOs and everyone else looking for innovative ways to transform their brand, drive demand, and just plain cut through — proving that B2B does not mean boring to business. Here's your host and Chief Marketing Renegade, Drew Neisser.

Drew: Hello, Huddlers. Today's conversation builds on a topic many of you have been wrestling with: how to move from an aspirational ICP to one that actually drives win rates. Drake Lenhan from Sitecore has been doing exactly that, combining data, operational rigor, and now AI agents to turn ICP into a practical growth layer. In our prep, it became clear that this isn't just about defining the ICP — it's about making it usable across marketing, sales, and the entire go-to-market engine. So first, I just want to welcome Drake and thank him for being here. How are you and where are you this fine day, Drake?

Drake: Thanks, Drew. I'm in Dallas. Beautiful weather in Dallas, Texas. Really happy to be here. So thank you.

Drew: At this point, we're going to just — I'm going to turn the show over to you. You can start sharing your slides, and we'll wait for you to pause when you're ready for questions.

Drake: So, thank you for the time today, everybody. I'm Drake Lenhan. I lead market intelligence and competitive intelligence for Sitecore, which is the second largest DXP, or digital experience platform company, in the world. So, I'm really excited to be here with you guys today, glad to see some familiar faces on the call, and new ones as well. So, like Drew said, today I'm going to walk us through an actionable and repeatable approach to identifying, validating, and then aligning your ICP across the organization to really kind of drive that meaningful impact across your full GTM initiative. So, I'm going to start with how ICP refinement became a really critical need for us here at Sitecore — how we identified a shift in kind of demand that required us to rethink, strengthen, and redefine our ICP. And then from there, I'll walk through how we use data, both qualitative and quantitative, to validate our assumptions and then build confidence in where we should focus, before taking that story to the business to really drive that cross-functional alignment and executive buy-in. And then I'll show us how we operationalize that ICP across our systems of record and positioning and messaging, and ultimately into agentic frameworks that we use to create something that's not just strategic but scalable and repeatable across the organization. And then I'll close with a three-part kind of blueprint that hopefully you can take and apply to your business as you see fit, to validate your own ICP, align your teams, and then ensure your programs and sales motions are focused on where you're most likely to win. So, like Drew said, I'm going to aim to pause at the end of each of my sections to answer any questions you want to dive into a little bit deeper, but we'll also have some time at the end to explore any of those sections that you're a little bit more interested in. So with that, I'll just reframe: we have refined our ICP a couple of times at Sitecore in the last three or so years, so I'll start with just why it was such an important initiative to us.

So, rewinding the clock — Sitecore uniquely benefited from the pandemic. I joined initially in 2021. Every organization was kind of forced to accelerate their DXP stack, their digital experience stack, in some way, shape, or form to really meet that digital-first buyer's preference, and because we were all digital-first at that time, it drove a massive spike in digital traffic, engagement, and ultimately conversion for us. So between 2020 and 2023, almost every enterprise was in the market for some sort of DXP technology. Sitecore was uniquely positioned as one of those very few vendors who could truly deliver the number one thing they were all after, which was personalization at scale, while also meeting enterprise-grade performance requirements. But as with any rapid market growth or shift, competition caught up, new entrants entered the market, and they expanded their capabilities quickly, closing the gap on those core features that were needed to kind of compete in this market. So, on paper, what happened — everything started to look the same: the same features, similar pricing, seemingly similar outcomes that everybody was driving. And that's when we started to see an impact over the last few years: win rates were declining, new deal sizes were shrinking, pipeline conversion was dropping. So, when we looked at it, competitive intelligence was one of those gaps that we had to solve, and we did that, but the other issue was we weren't starting in the right places to begin with. So we were effectively, as the slide would present, trying to sell a Ferrari to the daily commuter who might have been an inexperienced middle-class driver — so it looked great, but it wasn't the right fit for them, and it was consuming a massive amount of time, energy, and resources of ours without delivering that outcome that we needed. So we took a step back. We knew we needed to make a change. We knew we had to focus, and most importantly, we needed the data to kind of tell us where to go, and that's kind of the kickoff right here: how we validated our old ICP and then turned it into a predictive algorithm for success, which is the first section here.

So, before we did anything, we validated whether the ICP was actually correct. We went back, we looked at our data — where we had the highest win rates, where were the largest deal sizes, where did we see the fastest deal cycles — and then we overlaid that on our install base, where we saw the lowest churn segments, the highest CSAT, the strongest expansion and revenue groundswell across our install base. And then very quickly we saw a gap between where we thought we won and where we actually won. So where it was previously thought to be any large enterprise who needed website personalization — bonus if they're looking for headless, bonus if they're a Microsoft shop — we quickly saw that there was a much more focused refinement that needed to happen there.

So, from there, we analyzed common traits across those customers using the standard ICP dimensions — I think everybody's aware of: industry, firmographics, demographics, operating model, budgets, resources — and then some of the qualitative stuff around what were the business situations and compelling events. And then we took it a step further: we ran a regression analysis to understand which of these variables had the highest correlation to win rate, to ASP, to deal velocity, and where was that TAM. And that SAM was a very fine balance into precision, but giving you a big enough universe to go after. That's where we started seeing some of the interesting stuff. So, we identified diminishing returns, right? You go until the diminishing returns. So, what we saw was ARR — no brainer, one of the biggest weighting variables. Sub-industry, years in business was an interesting one, country and tech stack — that allowed us to then build a scoring model, which surfaced segments where win rates were nearly 3x higher than our baseline and industry benchmarks as well. And once we had that clarity on who the ICP was, we needed to make it actionable. So we took that scoring algorithm — all of the weighting variables that most correlated again with win rate, deal size, and velocity — and we translated that into a simple A through D scoring system: A = highest propensity to buy, D = lowest propensity to buy. And then we tested it across all three of the geos that we operate in — Americas and APJ — created more dynamic scoring for them to make it very specific to their region and historics. And the results were consistent in each geo and each segment that we were targeting: higher win rates, more predictable, higher ASP, deal cycles slightly faster — not a huge notable improvement there. And then we continued to validate that quarter over quarter in all of our win-loss reporting.

But we didn't stop there. So that same win-loss program — we utilized the sentiments from that to really understand the why behind the data. What were the compelling events? What outcomes were buyers actually trying to achieve? What decision criteria mattered most? And most importantly, which use cases consistently drove that differentiated value that Sitecore could uniquely fill? And that's where I'd say the most critical thing came to light: that the use case and scenario fit were just as important, if not more important, than all of the ICP segmentation fit. So, when we combined those together, we were able to identify — as you can see here — the perfect scenario, the conditions where Sitecore became that obvious choice. And when those conditions were present, we saw up to a 5x increase in our win rates. So we also validated that with our sales leadership, working very closely with them as thought partners here to ensure that alignment and adoption was going to be successful from day one. And through that process — last slide in this section — we identified industry specifically. Industry was one of the strongest predictors of our success, but our industry categories were maybe defined very, very broadly, and we couldn't go after everything. So we created a composite score that then normalized different scales of TAM, win rates, and ASP across all of our sub-verticals, which then allowed us to mathematically identify the top industries and where we should concentrate our efforts globally or in specific geos. And at that point, we then had a validated ICP, we had the predictive scoring model, we had the clear priorities for our industry focus and defined high-value use cases. So then the question became: great, we have this — how do we operationalize it across the business? Which is the next section. So I'll pause here if there are any questions the team might have.

Drew: Well, I want to start with one, which is just restate what you just did, so we get that on record one more time. Just to make sure — you talked about predictive scoring — go ahead and just rephrase what you just did.

Drake: Yeah, so we reviewed our existing ICP as we understood it, to validate it through kind of statistical analysis and correlation to win rate, ASP, et cetera. We validated that model across three different geos, worked with sales to identify the industries in those geos we should focus on, and layered in sentiment analysis from our win-loss to validate a strong, statistics-backed ICP and predictive algorithm — which we created a scoring model for — identifying the top 10 sub-verticals where our marketing efforts and teams should be focused and verticalized.

Drew: And I'm imagining that by having those highly focused sub-verticals and increasing the win rates within those, that you have also better, more relevant case histories for the next group of folks in that same target. I mean, so you could — by really getting down to that particular definition — I'm guessing now that we may be early in this process to know that.

Drake: No, we — we have, just like you said, case studies, specific use cases, putting the vertical kind of jargon on all of it — we started doing all of that. Some of the newer stuff is how do we verticalize our sales teams around that? So, creating vertical sales plays are new motions for us, but you know, you're spot on.

Drew: Okay, Alan, you have a question — come on camera, join us.

Alan: Yes. Hello, thank you, Drew and Drake. Great to meet you. So, couple of questions. You know, you spoke a lot about win-loss rates, which is great. How do you balance gross revenue retention and net revenue retention? So you take a look at customers: easiest to get, easiest to keep, and easiest to grow — you can kind of triangulate those three. And then the second question is, once you've done all this work and you're achieving 5x the win rates, how are you putting governance in place so your sales teams really focus on those ICP great fits and not anything just to make the quarter?

Drake: Yeah, the second bit I will touch on in one of my sections. I'll give you a quick preview, which is: stakeholder buy-in from the get-go is very, very important. It's also a prominent lens in all of our win-loss recordings, so they see it. The first question was: how do we balance NRR, GRR, and some of that net new revenue? That was part of the analysis. We do actually have two different scores — one is specific for net new, one is specific for migration. So we are moving from an on-prem to a cloud-based technology right now, which is a major focus for us, and we just relaunched a new platform as well for them to do that, to add more complexity. So part of the install base was looking at who has the lowest churn, who is using the most of the product that would benefit from this new motion — where we have now recreated a platform in the cloud. And so NRR, GRR, all of that was part of it, and I would say that is the biggest weighting we use. So, when we look at the algorithm, the largest weightings come from revenue impact. So, for our install base, that is like the biggest focus for us.

Alan: Great, thank you.

Drew: Just to clarify — then the process of this meant you were winning bigger deals that you are more likely to keep. Yeah, which also has to speak to sort of the efficiency of a sales motion, right? Because — you would all rather win bigger customers than smaller ones, right, particularly if they'll stay. So just confirming that trend: by narrowing the ICP and getting at these folks with a high propensity to buy at a good match with your solution and their need, you have been able to increase essentially your value per customer.

Drake: Exactly, and we've also decreased our loss reason due to pricing constraints.

Drew: Right, right. Bindu Chellappan.

Bindu: My question to you is: how did you get to identifying the use cases and the events? I mean, you can have intent platforms. What about the use cases? Was it like some kind of research and you applied it to the entire population, or was it feedback from the sales calls?

Drake: So it was a combination. So I lead market intelligence and competitive intelligence, which has also been rolled into it. A lot of it was directly from our prospects and our customers. Some of them were user reviews. We do a lot of market research, so some of it came from there, and a ton of it was direct feedback from sellers — especially those who have been with us for quite a while and have been really, really successful. So understanding direct interviews with them — regular, we call them sales champion meetings with them — what they're seeing. But also, we're a little bit unique now: our sellers are putting stuff in Salesforce. When we look at win-loss in Salesforce, they're actually capturing things, and we get a lot of valuable insight from there too.

Bindu: So that was my next question. What's the tech stack that you're using to get that feedback loop going?

Drake: So, last — but not to steal my own thunder — the last thing I'll talk about is Gong. We're implementing it right now. I'm really excited to get those trackers in, but Salesforce is the big one for just breadth of notes from sellers. I will say what made the pendulum kind of swing into that — where they actually started putting notes in there — is they started seeing how we were using it and feeding it back to them. We made that really notable in our reporting and our recommended actionable recommendations. So we've been seeing — it's hit or miss — but a lot more great notes, great depth, great insights, just from our Salesforce notes.

Bindu: Thank you.

Drew: Okay, if there aren't any more questions, let's keep going.

Drake: Okay, so we have talked about validating the ICP, creating a predictable algorithm, and validating the industries, and now we've got all of that good stuff. Now, how do we actually use this? Turn it from a static artifact to an operationalized business tool. And so this is where I would say one of the most important things is cross-functional alignment became really, really critical. We brought together stakeholders from across sales, from marketing, product operations, and we didn't just present the model that we had created, we tied it directly to the business impact specific to their organization, so you know what happens if we focus here versus where we are today. How will that impact pipeline, win rates, revenue? The one that got everybody's attention: what would your fiscal finish look like last year had you done this? And that got us really strong buy-in from our then COO, who is now our CEO, our chief marketing officer, and our chief product officer. And so from there, then we operationalized it. We embedded the scoring model into Salesforce, so every account had an ICP score. When we created the algorithm, we enabled marketing and sales teams to prioritize those targets based on that ABCD score. We then made the ICP a very at-the-core lens in our win-loss reporting, and then we refreshed our positioning, our messaging really to align with the ICP and the evolving use cases and scenarios that made up that perfect fit scenario.

Drew: Can you — I'm sorry, I got to go back, go back to that slide for a second. So there are a couple of things that you said just that I wanted to make sure I — so I really love the idea of the, as you called it, the fiscal fitness report, where you retroactively applied this to say the last quarter, the last year, to show that had we done this — that had to have been very powerful.

Drake: It's also what got our competitive intelligence program a lot of attention too, because we could literally point back to a specific competitor, call it Adobe or Optimizer. If you were "Audible ready," if you had this differentiation strategy, and you did this motion, that would have been worth, call it, you know, random number, a million euro in bookings last quarter from them alone. So that's — it's very powerful when you speak numbers and you speak it related to what they care about, which is in this quota.

Drew: Right. There was one other point that I wanted to make that I think is fascinating, which is by tightening up the definition of the ICP, you were then able to sort of tighten your positioning and the subsequent messaging, and I imagine that that goes all the way down to sort of the content that you created, because suddenly you can make it more relevant to a tighter group of people.

Drake: Exactly. It fit into the messaging house, which then, as you know, cascades down into everything that we're building for specific personas. Personas were part of the work that we were doing also, so specific personas for specific product lines, and then the new lens, which is the sub-verticals, so very, very specific categorization.

Drew: And it's just interesting because we've been putting a target and defining a target — that is not a new exercise. I think the newness of this exercise is the precision with which you were able to define the target and link it to actual sales as opposed to guesswork, which is a lot of, or wishful thinking. So, anyway, continue.

Drake: All right, so now we had identified and validated that ICP. We secured the input from across our business stakeholders and executive stakeholders, which was critical across the business, and now we operationalized our ICP, and everything we did to take it from a static marketing artifact to a dynamic embedded operating model that pointed us in the right direction for each of our initiatives, from marketing, sales flows, etc. right out of the gate, and then hot topic, we extended it further into AI.

Drew: Wait, wait, I'm sorry. Before you go to that — I just want to make sure that you validated it, you created the composite scores, you embedded it in the CRM, then we got to real-time AI scoring, and the result was more predictive and a higher win rate.

Drake: And operationalized, so sales can actually see it on their accounts.

Drew: So feel free, anybody listening right now to clap or say, "Hey, if I had this operating model with the results of this thing, would I be doing better than we are right now? If marketing could help drive such a model, does this feel like a very positive outcome?" It feels like this is a really interesting approach.

Ad Break: This show is brought to you by CMO Huddles, the only marketing community dedicated to B2B greatness and that donates 1% of revenue to the Global Penguin Society. Why? Well, it turns out that B2B CMOs and penguins have a lot in common. Both are highly curious and remarkable problem solvers. Both prevail in harsh environments by working together with peers, and both are remarkably mediagenic. And just as a group of penguins is called a huddle, our community of over 300 B2B marketing leaders huddle together to gain confidence, colleagues, and coverage. If you're a B2B CMO, why not dive into CMO Huddles by registering for our free starter program on cmohuddles.com? Hope to see you in a Huddle soon.

Drew: Okay, now let's get to the sexy AI agent component.

Drake: Of course. One of our core pillars is making AI our operating rhythm. We have a lot of resources when it comes to Copilot and AI throughout the business. One of the things we saw quickly — you know, I'll equate this to Salesforce. A lot of people will use Salesforce, but people don't create in Salesforce. That's the Salesforce admin's job. So, though we all have access to Copilot Studio, not a lot of people are going to create in Copilot Studio. So we formed a small COE for go-to-market agents in my business, where we're looking at what's going to drive the impact for our programs through agentic scale. The first one we built was the ICP agent. It evaluates things like account fit, use case and scenario alignment, messaging effectiveness, recommendations to improve win probability for deals in flight. That became then a foundation for a lot of what we built on top of that, which was the brand assistant agent that then takes all of that ICP and persona content grounded in our positioning and messaging, and allows our brand teams to create content that is aligned to the ICP. It is the grounding for our differentiation agent, which is then coupled with a competitive intelligence agent that takes all of that ICP content, pairs it with the competitive content, and then spits out competitive deal strategies. We're in a very mature market, so differentiation is like the number one thing for us in competes, and now we're creating a product overlay agent. We just launched a new product that is the combination of what used to be five composable products. This product overlay agent will then chain a lot of these different ones I've named, and some others that we're building, to act as that product expert to support discovery through proposal and demos and all of that good stuff, but everything used the ICP and those personas as the foundation to build on. So I'll pause there, because I'll get into how we maintain this ongoing and refine it.

Drew: Yeah, I'm really curious on the core engine creating this ICP. I'm just curious what you populated it with, and how you validated that this agent was able to then do everything else.

Drake: Yeah, so that scoring model — we took it and we made it kind of more conceptual. Like your firmographic data, you're looking for this, those types of things were ingested into it. Part of the ICP exercise was redoing our personas as well. We got really, really granular with user and buyer personas that fed into there. Where we had redone the positioning and messaging, that fed into there, so it's grounded only on these things to make it an expert in this small, small space, so it's not hallucinating into other things. Before we created it, we looked at what are we trying to accomplish, right? What's the objective? We're trying to impact win rates by targeting the right people. Who are our audiences, and what do they care about? So we go down to sales — sales wants to know that they're in the right account or in the right opportunity. If they're not, how can they strengthen it, qualify it, or exit? Those types of things. Our SEs, our brand copywriters, etc. We kind of grounded it in only what it needed to know, so it wouldn't hallucinate into other things. And then we started chaining it to the other kind of expert agents that create a function.

Drew: So we have this overarching thing, which is essentially personas and customers or prospects, really, really well defined on all sorts of levels. Okay, talk a little bit more about the brand assistant, and what it means when you say generating content perfectly aligned. We all know that LLMs generate so-so content that sounds and looks the same as all the other LLM so-so content. Talk about that, and what that means — is it helping identify the topics that you should be writing about, the white space out there? What's the process of using that so that you are not creating slop?

Drake: So part of those signals that you're referring to is my team's responsibility to surface those signals. We do also have those types of tools in our own product suite, so the suite we sell. But this was also a product of working with our content and digital strategy teams for how they ensure things are on brand every time, so the brand guidelines and all of that is ingested into this. Lots of testing with them to make sure that the output was consistent with what they expected for blog posts, white pages, or white papers. But also our SDRs, as an example, can use this for outreach — LinkedIn, email copy, et cetera — stuff that's going to resonate with a specific target in mind within that ICP.

Drew: Yeah, I'm going to just take a deep breath and just sort of imagine that the content — these things are only as good as the human eyes that are editing them. If you have weak editors who say, "Oh, that looks good," there's some of that that has to exist. The human in the loop on this to judge it has got to be important. But I don't want to get dwelling on that. Let's keep going. The differentiator agent — I'm thinking competitive battle cards that the salespeople can sort of dial into. How is that agent?

Drake: This one's my favorite. So we have a competitive hotline. That hotline is the white-glove service for differentiation strategy for all of our competitive deals and sellers to reach out. So we've got real quantifiable, backed information for the efficacy of that program. One of the things we saw is because it's a mature market, a lot of the requests we were getting were a little bit too far downstream in the deal cycle to really make a difference — like, if you've already shown everything you've got, how are you going to convince them it's different from everything they also saw? So upstream, what we're seeing is we've got all of these battle cards, we have systems, Crayon, all that good stuff, and they can use them, but they're not putting it in the realm of context. So in sales, if you show something that your customer doesn't care about, they're just going to say that that's too much, it's too expensive. And so what was happening is they were taking all of this stuff — over-differentiated and valuable to us — and they were presenting it as if their customer cared, without validating it. So the differentiator agent helps them plug in the context. It's actually associated with Salesforce. So, Bindu, this is another reason they're actually putting stuff in Salesforce — it's because it's linked to their opportunity. It pulls from the notes in their opportunity, and that's the context it's grounded on. It lets them know this, it asks for more context that might not be in Salesforce, and then it refers to our competitive repositories, battle cards, et cetera. And the output is consistent. It is a consistently formatted deal strategy. What is your objective statement? What are the discovery questions to lead you into value where you know your competitors are weak, and to validate your customer cares? Once that's done, what should you showcase in the product, and all the way down to how do you get to the next steps? So that's my favorite one. That's the one I spent the most time personally on. It's had hundreds of deals go through it since we've launched it.

Drew: Which is so — it gets more and more accurate because you did talk about there being a learning loop in this thing. It is interesting, because that does prevent the sort of kitchen sink approach. "Oh, well, we got this feature too." Because, yeah, but I don't care, so move on. Okay, so product overlay agent — what's that do?

Drake: So, if you imagine what a product relay does — they are the expert on all things product: technical documentation, how do you position it in a certain light, which you could probably guess some of the differentiator agent comes into play there. Roadmap, the technical documentation and comparisons, all of the things you would expect a product expert to do. It takes the grounding of the ICP agent, the differentiator agent. Now we're building agents about specific product sets, so we have — remember those five specific lines of products that are now incorporated into one — so it can source from the right product sets and the right use cases and produce the right deal strategy. Like actual deal strategies, not just the competitive strategy — deal strategy. What discovery questions, even down to mapping to the solution itself in the demo. How to demo that. What's coming up on the roadmap to be the next thing you want to kind of sell the vision of a long-term partnership. That one's in testing right now. It's chaining a lot of different agents for that type of output.

Drew: And is the ultimate win — you know, you've got this win — a salesperson can actually respond to a detailed technical question in almost real time? Because one of the big sort of fall-apart parts of the buyer experience is you get to a salesperson and they ask a question that's technical, like, "How does this integrate with X?" and they don't really have a technical answer, so then you have to meet with a technical person, and it just goes on and on. So does this enable salespeople to offer better product-related answers? Is that the idea?

Drake: Them and their solutions consulting team members — absolutely. The first one was when we saw the first iteration of this. We launched our platform, Sitecore AI. I got to lead that launch at our symposium event, so that was great. The first thing we did there is basically creating a Sitecore AI agent — essentially it was just grounded on a 50-page FAQ. We spent a lot of time getting the documentation correct, obviously. That's static, that becomes outdated. So what we've been doing in the background now is attaching it to the Confluence resources, like the dynamic resources that are always up to date by our product teams, our SE teams, etc. So it's getting the latest and greatest information, and it's retiring the outdated information all at the same time. I'll pause here for the next section. Bindu’s got a question.

Drew: Go for it, Bindu.

Bindu: Okay. Thank you. To build the routing for the ICP, right? I mean, the brand assistant, all those agents are good. How are you actually structuring or architecting, you know, sending it to the right company prospect? How did you architect that on Salesforce?

Drake: So we work with our ET team, or enterprise technology team. Part of it is licensing and security reasons, but also part of it is because they're experts in systems. So I'll give you one of the differentiator flow. This agent doesn't connect directly into our Crayon repository. It can really only access SharePoint. So we're doing a daily sync into SharePoint folders to get the latest and greatest competitive intelligence every single day in there that it's sourcing from. So we work with the product teams and all of the teams that we need content from to understand where do they work out of. I'll say some of it gave us evidence that they don't consistently work out of some place, so it's a forcing function to actually get a little bit stronger in organization there. Then we work with the ET teams to make those connections, and then in Studio you have a lot of really, really solid prompt engineering that's going on in the background for my team to drive to those outcomes. We're just leaning on our planning docs that we're trying to drive toward, so everything is leading up to those desired outcomes.

Drew: All right, moving ahead.

Drake: So we have all of our stuff — it's validated, stakeholders bought in, we've put it into systems of record, Salesforce, CRM, marketing motions, the sexy agents, as Drew called them. We knew we couldn't make it static. We had to refine it ongoing, because we know markets shift, technology and use cases evolve, buyer expectations change. So we built a continuous refinement loop. We run regular win-loss reviews with the field teams. We collect direct feedback from our sellers on ICP fit. We evaluate that feedback against our model and see if it improves the predictive accuracy, and when it does, we update that model, we update it in CRM, which then flows through all of our reporting, all of our targeting engine, all of our deal execution, and that ensures that our ICP stays dynamic and continues to be refined and improve over time. And I'll say I alluded to this earlier, but the next step for us is then to bring in real-time data. So we're implementing Gong trackers right now. I'm super excited about it. I've implemented Gong at other companies. I swear, I don't get paid by Gong to say these things. Brilliant tool, love it. And we're going to be implementing those trackers across prospecting calls, discovery calls, and our demos, so we're capturing use case signals, buyer intent, key questions, and patterns all in real time, that then feed directly into ICP refinement, win-loss analyses, and our AI agents. So we're moving from those periodic insights to that continuous real-time optimization, and that's the last piece I have for maintaining and refinement before going into kind of the blueprints of how you could take this and replicate it in your own business.

Drew: Let me ask you a question about — and this maybe goes back to what Alan was asking about the rogue — I'm imagining with Gong that because it's recording all these calls, it could probably very quickly identify, "Wait, this person is in the ICP, this is all wrong, this salesperson went off the reservation," so to speak. Is that part of this too? Is that, you know, just knowing that Gong is listening, that maybe that will keep the salespeople aligned on target, or just a wonderful sort of...

Drake: Early days, early days. But that is the intention, that's the intention. So my motto — I've been a sales manager — but my motto has always been, what you do early on matters way more than what you do later on. I'll equate that to also bringing in the stakeholders early for the ride matters a lot more than doing all the work and saying "this is what we're doing." So I say that because everything that we're doing should be done before they enter a deal cycle or choose to qualify out early. Now, we actually just today — with my sales ops companion — have integrated Gong with Closed, who is our win-loss provider. So we're getting two fronts: one, the Gong transcripts will then flow into my Closed win-loss interviews, and they will combine those, so my sentiment analysis is going to get so much more enriched with all of this real-time data. And then two, it flows back into Gong, which then the sellers can see, and their managers can report on, and sales ops can push. So yes, this is the intent and where we're going, but we're just now getting everything implemented in place.

Drew: Just imagining this scenario where the ICP match and all the other things, the score is way up here, and you lost. There must be sort of this — well, wait, we had everything perfect here. How are you able to sort of understand why? When that has to happen sometimes, right? Find those aberrations — whether maybe a competitor wanted it, it was the end of the quarter, and they said they're just going to cut their cost in half and take the win that way.

Drake: That happens. Yeah. I think maybe your question is like, how do you ensure they don't lose faith in the scoring? Well, deals get lost, and we lost a real deal — a real, solid, qualified deal with all of the right use cases — to a competitor we know can't support any of those resources. Twice that I can recall very, very recently — very high-profile deals. One was to Wix, which I mean, come on. But one was to another provider I won't mention by name, but they are notoriously known to not be able to be used by anybody but developer personas. The marketing team was seemingly leading the charge, so there's relationship aspects, there's politics, there's a lot of stuff you've got to uncover. At the end of the day, it's the math behind it. The win rates are still much higher; you still have a much higher likelihood to win the deal. We show them that you have a much higher likelihood to win a bigger deal, which gets you to your number quicker without needing more deals. And we're always reporting this back to those teams, but also when we have deals like that and we do a win-loss interview on them, we're looking through the lens of what did we do right, what did the competition probably do better, and what is that actionable recommendation we can take away to do better the next time. So this team is all over Closed, which is our win-loss tool — getting in there — but it's also in our win-loss reports, where we do readouts with the sales teams so they can see this. You know, nothing sucks worse than losing a deal that you know you should have won, but if you're doing it the right way consistently with the right target, it's kind of a no-brainer for us, because the converse is much, much lower success criteria or probability.

Drew: Paige O’Neill, go for it.

Paige: Hey, amazing presentation, really, really impressive what you've done. I'm just — I'm sitting here the whole time thinking about data, because you know, I've been at many companies, including Sitecore, where data was always swirling in the background, preventing us from moving forward, because everyone gets wrapped around the axle. Do we have the data quality? Do we want to build on top of this? Do you have any tips for how you moved past that and just got the work done, because it's never going to be perfect, right?

Drake: Yeah, you're right. And yes, Paige, you and I briefly overlapped, so great to see you. But the data here — we have some work to do on that still. That's a partnership between sales operations, marketing operations, and field to get that data hygiene better. But that's why we look at population first and then take sample sizes to kind of smooth for that data hygiene, and we looked for a 95% confidence level when we did that, and then validated across sales. So bringing in sales for that manual input upfront while we were scoring it was one way we got around it. Today, that data hygiene issue is a real concern for scaling this broader in a more automated fashion. As I think what you're alluding to is basically, if you don't fix the data upfront, you're just going to scale out garbage. Right now, we're looking at tools and systems and people in place to maintain the database, but that's how we've done it at the onset — sample size at a very high confidence level, and then some manual inspection from the sales teams on that sample set.

Paige: Thanks. I'm glad to hear it's not perfect there either.

Drake: You guys find out where it is, let me know. That'll be the reverse, right? Bring that to me, please.

Paige: Yeah.

Drew: So to get to this stage, how — and if you were sort of advising a bunch of CMOs who are going to go on this journey — how long would one expect to get to the place where you are now, which is, you know, you've just about implemented your third or fourth AI agent and you're expanding this? What's the timeline?

Drake: So the first time — so I've built this model at three different companies. I worked for Vista Equity Partners for about six years. This is one of their playbooks that they run. If we have any Vista alums or somebody there, you're aware of the CGS approach. So building the model for this business took the longest. I'd say you really want to get it right, so probably a good 60 hours of just building the model. The process from building the model to then getting it to where we are today — there's been a lot of movement in our business, so if I smooth for that and I extract that, I would probably say you're looking at six months. And then again, that's with working with the stakeholders to bring them along for the journey. It takes a little bit longer, but then they buy into it and it actually works down the road. So about six months.

Drew: Well, and it's funny — going back to Paige's point about the data — one of the things that struck me early on in this was that to make the business case you needed that sort of fiscal check that you did, where you would look at the previous deals. If you don't have good data to start with, even on that, you can't do that. So it feels like a fundamental part of all of this is data hygiene is essential. Drake, anything else that you want to cover?

Drake: The last slide I have is that kind of blueprint for what I would offer, just as a recap, and what you can do in your own business to kind of replicate this.

Drew: Go for it.

Drake: Do we want to go to Bindu, or do you want me to wrap up?

Drew: Go for it, and then we'll get to her, okay?

Drake: I've got you. So I'll just start by saying I think the biggest mistake we made from the get-go was jumping straight into activation of all of our programs without revisiting our ICP to validate that it was still actually correct, as things have shifted in our market, our business, etc. You know, previous analysis had been done, but it had been a long time since anybody had revisited it. So if I were to prescribe a Monday blueprint — take this on Monday — start applying things where you can identify, validate, operationalize, and maintain a fresh and accurate ICP. It would be this: validate first where you actually win. So look at your data — win rates, deal size, churn, expansion — and identify that true ICP fit. Refine that with the qualitative insight. So layer in your win/loss, your sales feedback, identify compelling events, high-value use cases, all of the good stuff. And align leadership upfront, like I said, to get sales, product, and operations — get all of them bought in on who you're targeting and very clearly who you're not targeting. And then prioritize focus to find your top segments. Stop spreading resources across everything so you can align your programs to that ICP. So if a campaign or a motion doesn't map to ICP or use case fit, either fix it or you stop doing it. And then operationalize all of this in your system — embed it into your CRM, your marketing automation, your sales deals, and workflows. And then measure and enforce that weekly, monthly, quarterly, through pipeline analysis, win/loss reports — track your win rates, ASP, all of that through the ICP lens. And then continuously refine. Use the data and the feedback loops we discussed to keep improving your model and making sure that it's always dynamic, it's always up to date, and you're not operating on something that's old and static. And so that's it. That's all I've got.

Drew: That's it — easy peasy, piece of cake, eight simple steps, come on people. No, but seriously, how many people does it take? Like, you need a team to do this right. This is not a job where you can say, "Hey, go do this." This is more than a six-month project. There are a lot of people involved in the process, right?

Drake: There's a lot of people involved in the process. Absolutely. I will say this was an initiative I took on personally because I had done it before and so I knew all of the motions. I do have a whole team that helps me with a lot of the other stuff behind the scenes. Our COE for agents — this specific one was right in my wheelhouse, so I did take it on. I have a lot of relationships — four years at Sidecore — so I've got a lot of relationships in the business. So it's sourcing them out to get some of the data points if I needed, but mainly to show what we found and how it aligns to what they care about. I'd say that, yeah.

Drew: So we have to wrap up, and I think I want to get from you a couple of key takeaways, if you will. What do you hope they take away from today's conversation?

Drake: I would hope it would be this — this one slide here: three phases, eight steps. But the three mistakes I have seen where this can fall down: one is — again, the mistake we made — we didn't review it at first, and it was static and stale. So review it, make sure it's up to date, it's up to snuff. I would say subjective versus objective. I've seen a lot of ICPs created from some analysis and then just using gut feel to create a story about who your ICP is, but that doesn't really have math or science or process behind it. So validate it objectively with numbers. I'm a big numbers guy, for those who know me. I love Excel — like, I live in Excel. So that's the second part. And then the third part: bring your stakeholders. Executive alignment is critical. Stakeholders that you need — who is going to use this — it's critical that they give input as part of this process. They see where their input was applied, because then they have skin in the game. They see how this fits, and it's a partnership versus marketing throwing something on them and then they don't do it. So those would be the three things.

Drew: Well, Drake, we really appreciate you being here with us. One of the things that is just so clear to me is that if you know that having a better ICP could increase the likelihood of your win rate by three to five times, that would be a worthwhile expenditure of your time — maybe a top priority — because the minute you can define your ICP tighter, the more focused your marketing can be, the more focused your customer success can be, the more focused your salespeople can be, the more focused your battle cards can be. It all sort of trickles down in a very positive way, and we saw that here. And the opposite is true — the broader the target, the broader everything is. And I'm glad Noreen is here, because she did this at Sada, where several years ago they said they're only going to target one of the cloud of three. They went all in, so they narrowed their addressable market by about 50% — maybe more — but saw their sales double. So that's a gross sort of way of doing it. What you've defined is a far more — I'll call it a much tighter — way of doing it. I also like the key insight about the micro verticals or sub-verticals. We see people say, "Oh well, cybersecurity is a target." No, that's not a target. That was really interesting to me, Drake. So I appreciate that.

 

If you're a B2B CMO and you want to hear more conversations like this one, find out if you qualify to join our community of sharing, caring, and daring CMOs at cmohuddles.com.

Show Credits

Renegade Marketers Unite is written and directed by Drew Neisser. Hey, that's me! This show is produced by Melissa Caffrey, Laura Parkyn, and Ishar Cuevas. The music is by the amazing Burns Twins and the intro Voice Over is Linda Cornelius. To find the transcripts of all episodes, suggest future guests, or learn more about B2B branding, CMO Huddles, or my CMO coaching service, check out renegade.com. I'm your host, Drew Neisser. And until next time, keep those Renegade thinking caps on and strong!