170 cold emails. 0 replies. Here's the ICP mistake I made -- and what finally worked.

I've been building Velor -- an AI customer support agent for small B2B SaaS teams. The product works. Getting first customers has been harder.

Here's what I got wrong.

The mistake: my ICP was too big

My initial target was "B2B SaaS companies with a support problem." Sent 170 cold emails to companies with 30-200 employees. Zero replies. Not even an out-of-office.

I spent weeks blaming my copy. Then my subject lines. Then my timing.

It wasn't any of those.

Companies with 30+ employees already have a support process. Buying new software means a team discussion, a budget line, and a champion who can push it through. Even founders at those companies aren't feeling the ticket pain personally -- someone else is handling it.

What actually worked: going much smaller

I narrowed to 2-15 person founder-run B2B SaaS teams. The ones where the founder IS the support person. Replying to tickets at midnight between shipping features.

The pain is immediate and personal. There's no approval process. The decision is "does this save me time?" -- and the answer is obvious.

First customer: a 2-person fintech team. Week one: 70% ticket deflection. They felt it immediately. One call to close.

The ROI math becomes simple at this size: if you're handling 100 tickets/month at ~$15/interaction in founder time, a $500/mo tool that cuts that by 70% pays back in the first week.

What I'd tell myself a year ago

Don't target buyers who need a budget cycle. Target founders who feel the pain themselves and can say yes today.

Small company, personal pain, fast decision = the only early customer profile that matters.

If you're a small SaaS team drowning in support and shipping has slowed down because of it -- happy to show you what we built: velorplatforms.com