The ICP mistake I made at Apollo and how to avoid it

When I was at Apollo.io scaling from $800K to $50M, we had what I thought was a solid ICP. Industry targeting. Company size filters. Revenue range criteria. All the standard stuff.

It worked well enough to get us to $50M. But looking back, we left a lot of pipeline on the table because our ICP described companies, not people. We knew which accounts to target but we did not deeply understand the specific problems our buyers dealt with every day.

Kevin Dorsey (KD) articulated this perfectly. Most companies have an ICP document. Almost none of them actually understand their buyers.

The practical difference is huge. When your ICP is firmographic, your outbound sounds like every other vendor: "Hi, I noticed your company does X. We help with Y." When your ICP is problem-based, your outbound describes reality: "I keep seeing VPs of RevOps struggling with lead response times above 24 hours because their routing logic has not been updated in 2 years. Is that you?"

The second one gets replies because it demonstrates understanding. The first one gets deleted because it demonstrates nothing.

If you are building a B2B product and doing outbound, here is the exercise. Pull 20 conversations with customers who closed. Write down the exact phrases they used to describe their problems before they knew your product existed. Build sequences around those phrases. That is your real ICP.

Full article: artemisgtm.ai/blog/why-most-b2b-companies-get-icp-wrong