All about subscription #2

In fall 2002, I watched a stranger win $5,000 on live television for answering trivia questions, and I thought I was watching my company's breakthrough.

We had put our regional final on the air. The winner had climbed a leaderboard for weeks to get there. Our shortcode sat at the bottom of the screen the whole broadcast. Somewhere in a server room, our message counter was spinning like a slot machine.

I was watching our breakthrough. I was also watching our diagnosis. It took me embarrassingly long to tell the difference.

Per-use revenue is a fire. It burns exactly as long as you feed it, and not one day longer.

That's my whole argument. The rest is the story of how I learned it, ten cents at a time.

Nobody texted machines

Text messaging in 2002 was enormous, and it was entirely person-to-person. You texted your friends. You texted your crush. Nobody texted a machine, because no machine had anything to say.

That was our founding idea: person-to-machine. Build something on the other end of a text message that was actually worth ten cents.

Our first product was a trivia game modeled on Who Wants to Be a Millionaire. You texted WIN to 3030 and the game texted you back a question. We built a dynamic question bank organized by category. We ran weekly and monthly leaderboards — regional winners first, then national. And once, for one regional final, we went on live TV with a $5,000 prize.

It worked. At peak, more than three million text messages had been sent to that game. At ten cents a text, I felt like a genius roughly twice a month.

The problem was the other twenty-eight days.

Two charts, one line

Here is what our revenue actually looked like. The TV broadcast: a spike. A newspaper ad: a spike. Radio spots: a spike. A wave of word-of-mouth buzz: a spike. And after every single one — a slide. Not a plateau. A slide, and a fast one.

One day I put the revenue chart next to our marketing calendar. They were the same chart.

That's when the fire metaphor stopped being a metaphor. The TV broadcast was a big log. The newspaper ads were kindling. Word of mouth was a lucky gust of wind. And every Monday morning we showed up to a pile of warm ash and started rubbing sticks together again.

We hadn't built a money machine. We had built a very expensive campfire, and we were the only ones gathering wood.

This is not a 2002 problem

It would be comforting to file this under "SMS era, ancient history." I don't think you can.

Pay-per-view has this wiring. Credits have this wiring. Tokens, top-ups, usage-based pricing — same wiring. When customers pay per use, revenue tracks usage, usage tracks attention, and attention tracks whatever you spent to get it. The names change. The physics don't.

And to be clear: some businesses are built for fire. Event promoters live spike to spike on purpose. If that's your business, this post owes you nothing.

But if you're staring at a spiky revenue chart and telling yourself the next campaign will be the one that finally makes it permanent — I have been you. The campaign works. Permanence is not what it produces.

Again, the honest part

Per-use was still the right way to start. Ten cents was a nothing-commitment that let millions of people try us without thinking. It proved demand fast. I'd probably start the same way again.

And I can't fully untangle, two decades later, how much of our volatility was the pricing and how much was the product. Trivia is bursty entertainment by nature. Maybe demand was the fire and pricing just refused to hide it. What I know is that pricing per use guaranteed our revenue could never be steadier than our customers' attention — and attention is the least steady thing anyone has ever tried to build on.

What we did next

We stopped charging for texts and started charging for months. It was a far less clever idea than the trivia game, and it mattered far more. Within two years it got the company acquired.

That story is the next post.

Until then: pull up your revenue chart. Now pull up your marketing calendar. If they're the same chart, you don't have a revenue problem. You have a wood-gathering problem.

Where does this argument break? Tell me — I answer.