ICP, Knowing Your Customer in 2025
Defining your Ideal Customer Profile (ICP) has never been more crucial or more complicated. As buying teams expand, and AI reshapes targeting strategies, understanding who your best-fit customers are is a nuanced task. In this roundtable hosted by Joaquin Dominguez, Head of Marketing at Adzact, a group of senior B2B and product marketers came together to explore how organisations are rethinking ICPs in 2025.
Host
Joaquin Dominguez , Head of Marketing at Adzact
Guests
Sophia Martin, Group Product Marketing Manager at Confluent
Antony Chang, Regional Marketing Leader at Brightcove
Jonah Youdeem, Product Marketing Director at Eightfold
Sush Koka, Director of Product Marketing at Bamboo Rose
Vivian Jordan, Senior Product Marketing Manager at Cornerstone on Demand
The Shape-Shifting Definition of an ICP
The discussion opened with a deceptively simple question: What does ICP mean to you and your company? It quickly became clear that the definition of an ICP is anything but simple.
Some organisations approach ICPs as a set of clear, firmographic boundaries. In verticalised industries—like oil and gas or supply chain tech—ICPs can be cleanly defined by sector, company size, and even CapEx patterns. But for others, especially those with broad horizontal solutions or multi-product portfolios, ICP is less of a static list and more of a strategic framework for continuous learning and segmentation.
Sophia framed ICPs as a blend of end-users and economic buyers, requiring both product-centric and ROI-centric messaging. Sush spoke of moving from clear-cut industry-based targeting to more abstract criteria, such as sales enablement maturity. For Brightcove, Antony described a dual ICP model—one clean and established for media clients, another sprawling and evolving for enterprise sectors.
"When you're targeting based on something like sales enablement maturity, it’s not a data point you can pull from ZoomInfo or an annual report. It forces you to get creative with proxies, signals, and sometimes just your own judgment." Sush Koka
In every case, ICPs were tightly bound to the company’s go-to-market model, maturity stage, and product mix. Jonah and Vivian emphasised that ICPs are not just labels but hypotheses. They're informed by data, yes, but also by shifting market opportunities, product readiness, and company strategy.
"When you're targeting based on something like sales enablement maturity, it’s not a data point you can pull from ZoomInfo or an annual report. It forces you to get creative with proxies, signals, and sometimes just your own judgment." Sush Koka
Beyond Eligibility: The Best Fit vs. the Just-Eligible
One critical insight surfaced repeatedly: Eligibility is not the same as fit. Vivian captured this tension, describing how many companies feel pressure to chase every eligible buyer, even those who may be poor matches in terms of retention, expansion, or strategic value.
Several participants drew attention to the importance of defining what not to target—creating an "anti-ICP" as a practical tool for disqualification. Jonah noted that distinguishing bad-fit customers not only protects NPS and retention metrics, it sharpens sales and marketing focus.
"Having an anti-ICP is just as important as knowing who to target. It keeps everyone aligned on where not to waste time and helps you avoid signing deals that look good on paper but turn into churn risks post-sale." Jonah Youdeem
Vivian introduced a helpful three-part evaluation model used in her organisation: Volume, Performance, and Potential. The team maps win rates, ARR growth, and customer lifetime value against TAM saturation to uncover high-opportunity segments. This more rigorous ICP framework supports internal alignment and guides resourcing—particularly important for large enterprises navigating regional and vertical complexity.
"For a while, we weren’t factoring in potential—and that quickly became a problem. We’d identify banking as our top-performing industry, only for a regional leader to say, ‘Yes, we already know that—and there’s only one prospect left in Italy.’ Without accounting for potential, we were optimising for past success rather than future opportunity." Vivian Jordan
"For a while, we weren’t factoring in potential—and that quickly became a problem. We’d identify banking as our top-performing industry, only for a regional leader to say, ‘Yes, we already know that—and there’s only one prospect left in our region’. Without accounting for potential, we were optimising for past success rather than future opportunity." Vivian Jordan
Operational Realities and Data Frustrations
Once defined, the next challenge is operationalising the ICP. Most participants expressed frustration with available data sources—particularly third-party providers like ZoomInfo.
Sush shared a cautionary tale about relying on NAIC codes to segment retail sub-sectors, only to find widespread misclassification and unworkable targeting lists. Antony echoed this issue, explaining that poor data hygiene can cause breakdowns in CRM systems and misalignments with sales teams. The consensus was that while third-party data may provide a useful starting point, it rarely survives scrutiny without heavy manual validation.
"You look at the list and think, ‘This can’t be right’—but unless you manually review it, that bad data flows straight into Salesforce, confuses your targeting, and frustrates your sales team. The time spent cleaning it up often outweighs the time it was supposed to save." Antony Chang
There were also practical lessons about timing. ICPs are often revised too frequently, without allowing enough time for sales cycles to play out and produce meaningful feedback. Jonah pointed out that even with well-structured data, it’s critical to clarify why an ICP is being defined in the first place—is it to expand market share? Enter a new vertical? Drive upsell in an existing base?
Without this clarity, ICP exercises risk becoming academic rather than actionable.
"You look at the list and think, ‘This can’t be right’—but unless you manually review it, that bad data flows straight into Salesforce, confuses your targeting, and frustrates your sales team. The time spent cleaning it up often outweighs the time it was supposed to save." Antony Chang
Organizational Alignment: The Hardest part
Multiple participants described the internal politics and cross-functional alignment challenges of executing against an ICP.
Vivian highlighted that the organisational design of a company—dedicated teams by sector, region, or product line—can either support or undermine a strategic ICP. She described situations where high-potential segments lacked internal champions or dedicated sales teams, while legacy segments with poor performance continued to receive disproportionate attention due to existing team structures.
Sophia added that introducing new personas presents its own risks. Even when a new market seems attractive, if the product doesn’t fully meet their needs, credibility is quickly lost. She questioned when a new persona becomes a true ICP—before or after product-market fit?
Jonah and Sush reflected on similar experiences: sometimes the surface-level signals—like firmographics or similar workflows—suggest fit, but fail to account for critical ecosystem factors such as shared supplier networks or user expectations.
Testing, Learning, and Moving Fast
A key theme running through the conversation was the need to treat ICPs as iterative, not fixed. Jonah advised placing a flag in the ground, testing the hypothesis, and being willing to shift based on what’s learned.
Sophia advocated for product marketers to sit in on first sales calls to hear objections directly, ensuring that messaging aligns with reality. Antony reminded marketers to lean on sales for insight, especially where data alone can’t provide clarity.
Everyone agreed that while rigour is necessary in the ICP definition process, agility is essential in execution. Vivian captured it best: be rigorous in the definition, agile in the operation.
"As an ex-SDR, I can’t stress enough the value of sitting in on first sales calls—not just hearing the pitch, but listening to how prospects actually respond. You catch the real objections, the hesitations, and whether your messaging is resonating. It’s the fastest way to keep your ICP grounded in reality." Sophia Martin
"As an ex-SDR, I can’t stress enough the value of sitting in on first sales calls—not just hearing the pitch, but listening to how prospects actually respond. You catch the real objections, the hesitations, and whether your messaging is resonating. It’s the fastest way to keep your ICP grounded in reality." Sophia Martin
Final Advice for 2025 ICP Refinement
Each speaker shared advice for B2B marketers refining ICPs in 2025:
Don’t let perfect be the enemy of progress. Move fast, test ideas, and refine them over time.
Use data, but don’t rely on it blindly. Historical performance is useful, but potential matters too.
Align early with sales. The best ICP is one your revenue team believes in.
Beware of data hygiene. The best segmentation strategy fails if your lists are broken.
Define who not to sell to. Knowing your anti-ICP protects your resources—and your brand.
Conclusion
Ideal Customer Profiles are no longer just strategy decks or CRM tags—they are living hypotheses, tools for organisational focus, and often, sources of strategic tension. As this conversation made clear, refining and operationalising an ICP in 2025 demands more than firmographics or buyer personas. It requires cross-functional collaboration, clean data, clarity of purpose, and above all, a willingness to test, fail, and learn.
As B2B marketers look to the future, the companies that best understand their customers will not be those with the most detailed ICP documents—but those that use them as a springboard for shared learning, smart experimentation, and strategic growth.
Join the Precision Punks Community
Access to the slack community of serious B2B advertisers
Chatham House rules ‘Adzact confidential’ events with peers
Private in-person events in London and New York
Exclusive data & learning resources to build your career
