Multi-Axis Revenue Segmentation
Plot customers on a graph of size plus the attributes that actually drive your revenue, not size alone
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 4
- Confidence
- 92%
A method for carving the universe of companies into segments that buy differently. Size (small/medium/large) is only the X-axis; you add axes for the attributes that most correlate with revenue in your specific business — growth potential, business model, traffic volume, workload type. You derive those axes by regressing on what actually drives revenue and where you repeatedly win.
Origin
Practiced by Jeanne DeWitt Grosser at both Stripe and Vercel. At Vercel she built it in her first 30 days by sitting with the head of data science to find which attributes predicted a $100k vs $1M customer.
Core principles
- 01Size alone is necessary but insufficient — it captures buying complexity, not buying likelihood or value
- 02Add axes that are specific to what drives YOUR revenue (consumption businesses care about growth rate; Stripe cared about business model)
- 03Segmentation is a company-wide input, not just a GTM artifact — it should shape what product managers build
- 04Cap it at about three attributes; beyond that you can't reason about it or staff against it
How to run it
- 1
Start with the size axis
Lay out small / medium / large (SMB / mid-market / enterprise). This captures how buying complexity changes: single decision maker at the low end, economic + technical buyers in the middle, procurement and committees at enterprise.
Watch out If you stop at size, you miss what within your offering changes how something gets sold.
- 2
Add the axis that predicts customer value in your model
For a consumption business, add growth potential — a company growing 200% year-on-year is worth far more than one growing 8%, so you spend more time and money on the fast growers. This directs targeting, not just messaging.
- 3
Add a 'how they buy' axis specific to your product
At Stripe this was business model (B2B needs wires + billing; B2C needs Apple Pay; marketplaces need Connect). At Vercel it's workload type (e-commerce sale uses different language and back-end than a crypto company) and traffic (a small company with top-25 web traffic like OpenAI gets promoted to enterprise treatment).
Pro tip Use observable third-party signals — Vercel uses Google's CrUX score to read a site's traffic and promote high-traffic small companies.
- 4
Derive the axes by regressing on wins
Sit with data science and ask: what attributes predict a $100k vs a $1M customer, and what clusters where we repeatedly win? That produced Vercel's insight that CrUX rank and workload type mattered, and that they under-penetrated enterprise SaaS (legacy code too costly to migrate).
Pro tip Teach the framework company-wide — Grosser delivers a 'Know Your Customer' session to every new hire's first week so PMs build with a segment in mind.
In the wild
Regressing on wins showed strong e-commerce penetration (fast sites matter for ecom) but weak enterprise-SaaS penetration despite being a developer-oriented front-end cloud — because those apps were built before Vercel existed and migrating millions of lines of code is a huge lift.
→ Clarified where to target; later revealed enterprise SaaS interest in the AI cloud as an easier wedge (adding AI-native features to existing apps).
Common mistakes
Segmenting on size alone
Size only tells you buying complexity; it misses value and buying behavior, so you spread effort evenly instead of concentrating on the segments that actually drive revenue.
Using more than about three attributes
Beyond three you can't reason about the segments or staff them — with five sellers you can't put one seller in five different segments, so over-slicing is unactionable.
Is it for you?
Best for
Founders defining an ICP and revenue leaders entering a new market who need to decide where to concentrate scarce selling capacity
Not ideal for
Very early companies still hunting for product-market fit whose customer base looks nothing alike yet
From the transcript
“x-axis was size um so small,”
“medium, large. Yaxis was growth potential”
“if you were going to grow at 200% yearon-year, you were more valuable to Stripe than if you were going to grow at 8%”
“think of three attributes that narrow them down”
“I actually deliver and every new hire's first week one of our company values is KYC know your customer”
From the episode
What world-class GTM looks like in 2026
Jeanne DeWitt Grosser (Vercel, Stripe, Google)