1
0 Comments

People signed up, saw their leads, and did nothing. That part was on me.

For months I obsessed over the wrong half of my product.

I kept making the "finding" better — sharper intent scoring, cleaner signals, more channels. And it worked. The tool surfaces people describing the problem you solve, in real time, as they type it. Reddit replies run 20.6% for us. Cold email sits near 1%. The top of the funnel was never the problem.

Here's what was.

A new user connects their account, watches the tool pull up real buyers — someone who posted an hour ago, "i'm doing outreach by hand and it's eating my whole day" — and then does nothing. Looks at the list. Leaves.

I assumed that if people could see exactly who to message, they'd message them. They didn't.

I missed the reason for weeks.

The first send is the scariest thing in the entire flow. It goes out under your name, to a stranger, in public. Finding a lead costs you nothing emotionally. Sending to them does. I'd spent months shaving effort off the part that was already easy.

A tool can remove 90% of the work and still fail — if the last 10% is the frightening 10%.

So I stopped improving the finding and rebuilt the first action.

The old first step was basically "go set up your outreach." Now the first thing you do is approve one draft — to one person who already raised their hand, written in your own voice, so hitting send doesn't feel like gambling your reputation. One yes. Not a campaign. Something you can undo.

Activation stopped being "learn the product." It became "get one real reply."

If you're building anything, here's the portable part: find the single action that means someone is actually activated — then make that exact action the smallest, safest, most reversible thing in the whole app. Don't onboard people to your features. Onboard them to one outcome.

You can try the finding part yourself — first leads are free, no card: https://www.leadsynthai.app

What's the one action that counts as "activated" for your product — and how long does it take a new user to hit it? Curious where mine lands.

on July 21, 2026
Trending on Indie Hackers
I built a web-based vector editor from scratch and integrated an AI Agent. Need just ONE beta tester! User Avatar 65 comments I built an AI that turns an idea into a live business in under 10 minutes. Here’s what 1,000 launches taught me User Avatar 35 comments "Looks Good to Me" Is Quietly Killing Your Feedback Loop User Avatar 24 comments Building Noodle, a keyboard-first REST client for the terminal User Avatar 24 comments Launched 580 landing pages in 1 week. Solo. No team. User Avatar 18 comments What Building a Local Dumpster Rental Business Taught Me About Customer Intent User Avatar 7 comments