Launch day gave me 14 signups. A forum reply gave me 96

Launch day sent me 14 signups. A forum answer I wrote three weeks later sent 96. That looks like an easy conclusion and it's the wrong one, because if you measure it the obvious way the launch wins for nearly three years.

Here's the honest version of the comparison, and then the reason content still wins.

Cumulative traffic, which is the measure everyone reaches for

A launch spike is a decaying exponential. Mine was around 800 visitors on the day, then roughly halving each week. Add that up over all time and you get about 8500 visitors, and you have essentially all of them inside the first fortnight.

A page that ranks is the opposite shape. It starts tiny and grows. Mine began at about 30 visitors a month and grew around twelve percent monthly as it climbed the rankings and picked up links.

Both lines are the same product in the same year. The crossover is month 32.

Thirty two months. Nearly three years before the compounding page has delivered more total visitors than one day of launch traffic.

If you've read a hundred posts telling you that content compounds and launches are vanity, that number should annoy you a bit. It annoyed me. It's also correct, and pretending otherwise is how people end up disappointed eight months into a content plan.

So why is the answer still content

Because total visitors is a terrible measure and I only used it to make the point.

Same product, same month, roughly a seven to one gap in signups and a twelve to one gap in customers who stayed. The visitors weren't equivalent, and that's the entire point.

The signup is identical on both routes. Everything either side of it differs.

Launch traffic arrives with no problem in their head. They're browsing new products, which is a hobby, not a job. Search traffic arrives having spent an hour of their own evening being annoyed by the exact thing you fix. You aren't competing for their attention. You're turning up at the moment they've already decided they want this solved.

The bit that actually decides it

The launch is done. There's no version of next month where I do it again.

The page is an asset that still runs. It has produced something between two and six signups every month for over a year with no maintenance, and the traffic model above is generous to the launch because it assumes the page stops growing, which it hasn't.

Cumulative visitors compares two flows. What you should be comparing is one flow against one asset, and the sensible question isn't which delivered more by now. It's what you own at the end.

Where I wasted six months

I tried to industrialise it. Twenty more forum answers in a week, most of them thin. Two signups from the lot.

The one that worked took forty minutes and covered a problem I'd personally lost an afternoon to. It contained a sentence about what happens when the bank posts a fee on the last day of the month and your import runs at midnight. Nobody who hasn't been bitten by that writes that sentence, and that sentence is the whole reason anybody believed the rest of the answer.

That's the constraint people miss. You can't batch this, because the raw material isn't writing time. It's having genuinely solved something.

The output rate is set by the first box, not by how much time you set aside for writing.

I keep a file. Every time I lose more than an hour to something inside my own product or a customer account, I write down what broke and what fixed it while it's fresh. About thirty entries so far. When somebody asks a question that touches one, I have a real answer ready and it takes forty minutes instead of an afternoon. When nobody asks, I don't write anything that week, and that's fine.

Choosing where to write

The general startup forums were a waste. Everyone there is a founder, and founders aren't your customers unless you sell to founders.

What worked was small, tied to one job title, and slightly unpleasant. People complain there. Complaining is the signal. A forum where everybody is supportive is a forum where nobody has an urgent problem.

Three checks:

• Do these people have the job title I sell to
• Are they describing problems rather than sharing links
• Could I answer one question a week here without researching it first

If the third is no, that room isn't for me yet, and showing up anyway just produces the thin answers. There are two places in the world where I can be useful. Two is enough.

One more test. Search for the problem you solve, in the words a customer would use. If page one is all vendor blogs, that market has no honest room to talk in. If page one has a forum thread from 2019 with forty replies, go there.

The self promotion question

I worried about this far more than it deserved.

The rule is that the answer has to be complete without the product. Somebody should be able to read it, do the thing by hand, never click through, and still feel they got what they came for. Then one line at the bottom, under my own name, saying what I build.

Two years, zero complaints. People who object to self promotion are reacting to a pitch with filler on top, and everyone can spot the difference in a sentence.

There's a second reason it works. The manual instructions are usually annoying enough to sell the product by themselves. My best answer explains how to do the thing by hand in eleven steps. I never had to point out that my tool does it in one, because anybody who just read eleven steps has already worked that out.

What this costs

Forty minutes, once or twice a month, plus twenty minutes a week reading the forum to catch the right question. Call it three hours a month.

The launch cost two months of part time preparation and produced one stranger who stayed. Not nothing, and worth doing once, because it feels good and a few people who congratulated me have since sent customers my way. Just don't call it a channel and don't build a quarter around it.

And if you're starting the content thing this week, hold the month 32 number in your head. Not to put you off. So that in month eight, when the numbers are still small and every post you read says this should be working by now, you know that's exactly where the curve says you'd be.