A retrospective on a failed launch

I have a friend — and this is one of those "I have a friend" pieces, with names and identifying details lightly altered — who spent ten months building a productivity tool for sales teams. They launched in February of this year. The launch did not work.

The post-mortem we did over coffee three weeks later is, I think, more honest than the public retrospectives founders usually write, so I asked if I could publish a version of it. They said yes, on the condition that no one could tell which product it was. So this piece is deliberately vague about the product and very specific about the lessons.

What the launch looked like

The plan was reasonable. Product Hunt on a Tuesday. A coordinated thread on X. A small podcast tour booked two months in advance. Two paid newsletter sponsorships landing in the same week. A direct outreach campaign to a list of 400 hand-picked sales leaders.

The infrastructure held up. The product did not crash. The onboarding worked. The pricing page rendered correctly.

What did not happen was conversion. A few hundred trial signups. A handful converted to paid. A non-trivial number of "this is nice but" emails. By the end of week two, it was clear that this was not the breakout launch we had hoped for. By the end of week six, it was clear that the trickle was not going to compound.

The diagnosis my friend reached

My friend's first instinct, which I think is the wrong instinct, was to blame the channels. The Product Hunt audience was the wrong audience. The newsletter sponsorships were too broad. The podcast hosts didn't tee up the pitch well enough.

After two weeks of staring at the funnel data, the actual diagnosis turned out to be much harder to swallow. The launch worked exactly as well as the product deserved. The product was good. It was not, however, urgently better than the tools the target audience was already using.

Most launches that "fail" fail this way. Not because the channels were wrong, but because the product was a 7 in a category where the incumbents were 8s. Launches do not lift a 7 to an 8. They reveal what was already true.

The specific mistakes

A few mistakes my friend was willing to name on the record, with my notes:

• The product was built for the wrong persona within the right ICP. Sales teams have at least four kinds of buyer — IC reps, managers, RevOps, and CROs. The product was built for managers but pitched at ICs. The result was a demo that managers loved and ICs ignored, with no path through the org chart.
• The "build in public" content was retrospectively useless. Ten months of weekly updates on Twitter generated maybe forty real customer conversations. Most of the audience was other founders, who are not the ICP for a sales product.
• The pricing was the same as the incumbents. Matching the incumbent price tells the buyer you are equivalent. You probably need to be either cheaper or much more expensive, with a clear reason.
• There was no post-launch plan. The team had ten weeks of plans for the launch and zero weeks of plans for what to do if the launch did not land. They spent the first week of "did not land" rerunning the same playbook.

What happened next

My friend has not shut the product down. They have stopped marketing it. They are spending the next quarter inside two customer organizations as a contractor, watching the actual workflow the product was meant to support. They expect to come out of it with either a sharper v2 or a graceful shutdown.

That is, I think, the most underrated move after a failed launch. Not relaunching. Not pivoting. Just watching the customer for ninety days with the product set aside, so the next round of decisions is based on real observation rather than launch-day adrenaline.

What this kind of retrospective is for

The reason failed-launch retrospectives are usually unhelpful is that they get written too soon and too publicly. The founder wants to extract a lesson before they have one. The audience wants a moral. The result is usually a list of "things we learned" that read like LinkedIn truisms.

The honest version, three weeks in, sounds like this: "we don't know yet whether this product can work. We know the launch did not. We have a quarter of work ahead of us to figure out which of those is the real story."

That is not a tweetable conclusion. It is, in my friend's case, the right one.