How to Build a B2B Ideal Customer Profile Narrow Enough to Use
We sell to anyone who will buy. Why does growth still feel so hard?
Not Another CEO · David Politis
Published Updated
A customer profile broad enough to include everyone drains sales time, blurs messaging, and pulls the roadmap toward edge cases. David Politis builds it from data instead. List current, churned, and lost customers, ask your best people who the best customers are, then find the traits the best fits share and the traits the worst fits share. Then use the profile to decide what to stop doing.
Every deal looks like a good deal
A product that works for one kind of customer usually works, at least somewhat, for others. So a founder says yes. A small account here, a large one there, a buyer from a department the product was never built for. Each one brings revenue. Each one also brings its own requests, its own sales conversation, and its own reasons to leave.
David Politis, who founded BetterCloud and now writes Not Another CEO, calls this one of the biggest mistakes founders make early on. His point is not that the extra customers are bad. It is that a team trying to sell to everyone ends up selling to no one well. He also argues the problem does not fade with size. It can hurt a company at any stage.
What a broad profile costs
Politis lists the damage. Sales spends time on prospects who were never going to buy. Messaging has to cover so many buyers that it speaks to none of them. The customer base gets too varied to understand. Product builds features that solve the wrong problems for the wrong customers.
He tells a story from Tony Eales, an operating partner he worked with at Warburg Pincus, about a fleet tracking company Eales took over as CEO. The team was proud of a large pipeline. It held everyone from a single van owner to fleets of hundreds, sold to fleet managers and owners alike. When Eales looked closely, the company only won with businesses running a small, tight band of vans. Politis recalls it as three to ten. Those customers were big enough to need tracking and too small for an enterprise system. The buyer was always the owner. Once sales and marketing aimed only there, sales became much more predictable.
Build the profile from your own data
Politis is firm that the profile comes from data, not instinct. His method has three parts.
First, list every current customer, every churned customer, and every deal you lost. Add what you know about each. Location, size, industry, revenue, the tools they use, who championed the purchase, how long the sale took, what it was worth, and whether it renewed.
Second, ask the people closest to customers. Your best salespeople know which accounts were easiest to win, and why. Whoever looks after customers knows who is happiest, easiest to grow, and least likely to leave.
Third, look for patterns at both ends. Which customers stay, grow, and close fastest, and what do they share? Which ones take longest to close, churn, and complain, and what do they share? Politis aims for seven to ten defining traits of the best fit, plus as many traits of the worst fits as the data shows. He also defines the buyer inside the account. Selling to mid-level managers when the CFO makes the decision wastes the effort.
The worst-fit list does as much work as the best-fit list. It is the one that tells a team when to walk away.
Use it to say no
A profile only helps if it changes decisions, and Politis expects it to change several. Sales works fewer accounts, each one chosen. He warns this makes salespeople anxious, and argues they do better with fewer accounts because they are the right ones. Marketing puts its effort into the profile. Product drops features planned for other segments and stops work that does not serve the profile, even when that ends a whole roadmap.
That last change is the hardest, and it is where edge cases do the most damage. A request from a customer outside the profile looks like demand. Politis's view is that the team should not build anything that does not serve the ideal customer.
He cites one company he advised that had 10,000 accounts in its CRM. After the exercise, only 2,000 of them really mattered.
Keep it current
Tom Buiocchi, former CEO of ServiceChannel, gave the rule its shortest form: "Segment until it hurts." His company had been selling to any business in retail, and its close rate was poor. It narrowed to large food and apparel companies, based in the US, with more than 100 company-owned locations spread across the country.
Politis ends with one caution. The profile is not a one-time exercise. As the business changes, the best customers change too, so the work gets done again.