Most founders make the same mistake with their ICP, they define it way too broadly. “Anyone with X problem” or “any company in Y industry”. These aren’t real targets, they’re just guesses. A real ICP needs to pass what we call the “list test”, can you actually sit down and generate a finite list of prospects? You should be able to plug your criteria into Apollo or LinkedIn and get back actual names and companies. If you can’t do that, you don’t have a strategy, you’ve just got a wish.
The key is getting specific. Instead of broad stuff like industry or location, focus on concrete attributes: exact job titles, employee ratios (like engineers versus salespeople), physical assets (warehouses, refrigerated trucks), or tech stack details (Salesforce users versus HubSpot users). These are the filters that actually help you find real prospects.
Here’s what most people don’t realize, your ICP is usually something you discover, not something you invent upfront. Take Intelichain, they started targeting SMBs and found them way too price-sensitive. Through actual sales conversations, they realized large enterprises were much more receptive. They shifted focus, and revenue took off. Same story with Metaverse. They went after enterprise metaverse clients initially, but their real buyers turned out to be L&D departments and content marketing agencies who needed those long, engaging sessions.
Since your total addressable market is usually too big to tackle all at once, start with a Market Beachhead, a small, uniform segment that’s easiest to break into. Focus there, get profitable, then expand.
Also, forget predictive analytics for figuring out your ICP. Most companies already know who they want to sell to. The real problem isn’t defining the profile, it’s getting the actual list of contacts.
