How to Tell Whether Your AI Launch Worked
A launch feels like a success or a failure based on the spike, but the spike misleads. The four questions that actually tell you whether a launch worked.
Published 2026-10-10 · 5 min read
In the first 48 hours after a launch, you cannot tell whether it worked. The traffic number is either exhilarating or deflating, and both reactions are usually wrong, because the spike measures attention and attention is not the thing you launched for. The real read comes from four questions, and none of them is "how many visitors".
This is how to evaluate a launch honestly, at the right time, on the metrics that mean something.
Why the spike misleads
- A big spike with no activation is a failure that feels like success. Thousands of visitors who bounce tell you your distribution worked and your product or page did not.
- A small spike with strong activation is a success that feels like failure. Fifty visitors, ten of whom activated and two of whom are now talking to you, is more progress than a thousand bounces.
- The spike decays regardless. Judging a launch by its peak is judging it by the one number certain to fall. The question is what the floor is after it falls.
- Attention is borrowed. A launch borrows an audience for a day. Whether the launch worked is a question about what you kept, not what you borrowed.
The four questions that actually matter
| # | Question | What it tells you | When to ask it |
|---|---|---|---|
| 1 | Did the right people show up? | Whether your targeting and description worked | Day 1-2 |
| 2 | Did they reach the value? | Whether activation held under real traffic | Day 3-7 |
| 3 | Did anything durable get created? | Reviews, links, a page that ranks, relationships | Day 7-30 |
| 4 | Did the floor rise? | Whether the baseline is higher than before | Day 30-60 |
Question 1: did the right people show up
Not how many — who. A launch that brings a thousand people who cannot use your product told you your positioning attracted the wrong audience, which is useful and fixable.
- Check the source and the questions. What did people ask in the comments and support? Questions reveal whether they are your buyers or just curious passers-by.
- Look at where traffic came from. If it is mostly other founders, you launched into a peer audience rather than a buyer one — a positioning or channel problem, not a product one.
- Note the mismatch. If the people who activated differ from the people you expected, that is your real market telling you who it is.
Question 2: did they reach the value
Activation under launch traffic is the real test, because a launch is the most traffic you will get for a while. If the product does not activate people at volume now, it will not later.
- Measure activation, not signups. The share who completed the core action, not the share who registered.
- Find where they dropped. Launch traffic exposes your onboarding leak faster than anything else. This is the single most valuable output of a launch and the one most founders ignore because they are watching the traffic number instead.
- Compare to your pre-launch baseline. If activation held or improved under load, the product is sound. If it collapsed, the launch found a problem you needed to find.
Question 3: did anything durable get created
A launch should leave artefacts behind. If all that remains is a traffic spike that decayed, it was an event rather than an investment.
- A permanent product page with reviews and comments that keeps receiving some traffic and builds trust.
- Links — from the launch platform, from anything that covered it, from listings you submitted around the same time.
- A list of people who engaged, who are the warmest audience you will ever have and whom almost nobody follows up with.
- Feedback you could not have gotten any other way — the questions, the objections, the reasons people gave for not signing up.
- A story. What happened is now content: the numbers, the surprises, the mistakes. This keeps earning after the spike is gone.
Question 4: did the floor rise
This is the real verdict, and it takes thirty to sixty days. Ignore the peak; look at the baseline after the spike has fully decayed.
- Is your weekly traffic higher than before the launch, once the spike is gone? A launch that raised the floor worked, however small the spike looked.
- Did any channel start producing that did not before? A listing that now sends a few visitors a week, a page that started ranking, a mention that keeps bringing people.
- Are people still arriving who mention the launch? Long-tail interest from the event is a sign it created something durable rather than just noise.
Reading the combinations
| Spike | Activation | Floor after | Verdict |
|---|---|---|---|
| High | Low | Unchanged | Distribution worked, product or page did not. Fix activation before launching again |
| High | High | Raised | Clear success. Do more of whatever produced it |
| Low | High | Raised | Quiet success. The audience was small but right — scale the channel |
| Low | Low | Unchanged | Positioning or product problem, not a distribution problem. Go back to the users |
| High | High | Unchanged | A real spike with nothing durable built. The follow-up is what failed |
What to do with the answer
- If activation was the problem, do not launch again until it is fixed. Launching to a leaky product wastes the audience you spent effort assembling.
- If the audience was wrong, reposition and re-target before the next push — the channel was not the problem, the message was.
- If nothing durable was built, that is a follow-up failure, not a launch failure. Build the durable artefacts from the material the launch produced.
- If the floor rose, note exactly what caused it and plan the next launch to amplify that, not to repeat the spike.
A launch is not a result, it is an instrument. It produces information — about your audience, your activation and your positioning — and whether it worked depends on whether you read that information, not on how high the number got on day one.
More in Measuring growth
- How to Run a Weekly Growth Review as a Solo FounderA 30-minute weekly review that turns scattered metrics into one decision: what to look at, what to ignore, and how to keep it from becoming a ritual.
- How to Reuse a Failed Launch as a Second LaunchA launch that flopped leaves feedback, a fixed product and material. How to turn a failed first launch into a stronger second one, not start over.
- The 90-Day AI Product Growth PlanA sequenced 90-day growth plan for a new AI product, tying together cold start, distribution, content and measurement into one ordered path.