Most B2B companies define their ICP using just firmographics: industry, headcount, geography. Some can't even tell the difference between B2B and B2C customers in their CRM.
We don't do that. We use our own product to figure out who gets the most value from Unstuck Engine—and it's way more nuanced than "SaaS companies with 50-200 employees."
We build for B2B GTM teams. Not just SaaS - any company selling to other businesses. Software, services, agencies, consultancies. What matters isn't your industry. What matters is that you need precision in your go-to-market.
The common thread? You've realized that volume doesn't work anymore. Sending 10,000 cold emails to get 5 meetings isn't efficient—it's expensive. You're looking for a better way.
Most companies build one ICP and call it done. We don't.
We run multiple ICPs simultaneously—up to 8 at once—because different customer segments need different approaches. And we practice what we preach: we use Unstuck Engine to score our own prospects.
Here's what one of our ICP configurations looks like:
This isn't arbitrary. We build these scoring models based on two key factors: win rates and net dollar retention. The companies that close fastest and stay longest? Those patterns become our ICP definitions.
But unlike traditional single-dimension scoring, we're looking at dozens of attributes simultaneously, each weighted based on actual outcomes. That's the difference between "they're a SaaS company, close enough" and "they're a 51-200 person privately-held SaaS company founded in 2019-2022, using subscription pricing, with a sales-led motion, and they use Clay—this is an 87/100 fit."
Our ICPs aren't defined by industry or company size alone. They're defined by use case—how much of the platform you actually need.
Each successive ICP uses more of our capabilities. This isn't a hierarchy of "better" customers. It's a map of different needs.
What they use: CSV export/import, webhooks, our scoring engine
Who they are: Teams that already have sophisticated automation workflows in Clay or n8n. They take our multi-dimensional ICP scoring—the part that's genuinely hard to build—and plug it into their existing systems.
Why this works: They get the best of what we do (precision scoring across multiple ICPs) without changing their entire workflow. They're technical enough to handle integration themselves.