Growth is where most build-in-public posts start showing charts.
I don't have a chart worth showing yet. What I have is a clear idea of what I'm measuring and why, written down before the results come in. I think that's the more useful post to publish first, because once the numbers exist, it's too easy to pick whichever ones look good.
What happened
Before posting anything on X, I wrote a distribution plan with a six-stage funnel. Each stage has one metric:
| Stage | What it means | Metric |
|---|---|---|
| Attention | Someone sees HairOver | Impressions |
| Curiosity | They want to know more | Profile visits |
| Trial | They click through or install | Clicks / installs |
| Activation | They upload a photo and get a first look | First generation rate |
| Retention | They come back and generate again | 7-day / 30-day retention |
| Advocacy | They share a result | UGC shares |
The website also tracks a small set of events tied to the same funnel: google_play_click, feature_click, hairstyle_category_click, faq_open, cta_click. I explain how the site is built in the landing page post.
Why I made these decisions
Activation is the stage that matters most right now
Impressions are exciting and mostly useless on their own. A thousand people seeing a post tells me the post was seen. It doesn't tell me the product works.
Activation is the first moment the product delivers on its promise. For HairOver, that moment is simple: you see a haircut on your face. Everything before it is marketing. Everything after it depends on it.
So the number I care about most is: of the people who upload a photo, how many reach their first generated look?
If that number is low, I don't have a marketing problem. I have a product problem, and more traffic would only make it more expensive.
Retention is the honest weak spot
I wrote about this in the Entrepreneurship Masterclass: people don't get haircuts every day. HairOver's core problem shows up every few weeks at best, and a real style change much less often.
That means people probably won't come back on their own. If HairOver is going to retain people, it has to give them a reason: new styles, colors, trying something bolder, or coming back right before the next haircut.
I don't know yet which of those works. That's exactly why retention is in the funnel from day one instead of being added later.
Conversion depends on showing the result first
The pricing analysis I did pointed to one idea as the most important paywall decision: let people see a result before asking them to pay. A blurred or watermarked look first, then "unlock your look."
I'm careful about how I say this. The analysis came with benchmark claims about how much this lifts conversion. I haven't verified those numbers, so I'm treating the whole thing as a hypothesis to test, not a fact. The full reasoning is in the pricing post.
What I built or tested
- The funnel and metric list above
- Website events that map onto the trial stage
- A rule for my X content: judge formats by installs, not likes, with UTM links on everything so X → site → install can be traced
- A weekly review for the product metrics and a monthly one for revenue
What happened next
This is the honest part: I'm not publishing funnel numbers in this post. They'll come in their own posts, with the same stages in the same order, so you can compare them against this plan.
What worked
The main thing that worked is having a single metric per stage. It stops me from reading "lots of impressions" as progress when nobody is activating.
What didn't work (yet)
I can't say yet. Ask me again when the first month of data is in.
What I learned
Growth isn't one number, it's a chain. The weakest link sets the limit. More attention doesn't help if activation is broken, and great activation doesn't help if nobody comes back.
For an infrequent-use product like HairOver, I think the chain is most likely to break at retention. But I'd rather measure it than assume it.
What I'm testing next
- First generation rate. Where do people drop off between upload and first look?
- Paywall placement. Result first, then paywall, compared with paywall first.
- Reasons to return. Which ones actually bring people back: new styles, colors, or reminders before a haircut?
- Channel quality. Do X, search, and creators bring people who activate, or just people who look?
Key takeaways
- Six funnel stages, one metric each. Activation (first generation rate) matters most.
- HairOver's likely weak spot is retention, because haircuts are infrequent.
- Showing the result before the paywall is a hypothesis, not a proven fact.
- The funnel is written down before the numbers so the numbers can't rewrite the plan.