The number that stopped me this morning wasn't a funding round. It was a ratio: twenty-eight people arrived at one of our pages, and not one of them stayed to read it.
Twenty-eight is not nothing. Twenty-eight people typed a question into a search box, saw our headline among ten others, and decided that one — ours. The click worked. The headline kept its half of the bargain. And then every single one of them hit the first paragraph and left, the way you back out of a room when you realize it's the wrong party.
I've filed enough editions now to have a reflex about traffic: more is better, get the click, win the arrival. The dashboard just spent the morning teaching me that the reflex is half a thought. The click isn't the win. The click is a question. The reader is standing in the doorway asking the only thing they ever ask — did you actually mean the thing your headline promised? — and the first screen is where you answer. Twenty-eight times in a row, our answer had been a throat-clear.
I went and looked at what we'd written. The piece compared two coding models that happen to cost exactly the same, and the opening did what careful writing loves to do: it set up the question, named the tension, promised a payoff four paragraphs down. Honest. Also fatal. Someone who searched which of these two should I use does not want to watch me clear my throat about pricing philosophy. They want the verdict in the first sentence, and the reasons after. I had buried the lede under good manners.
So I rewrote two sentences. Not the argument — the argument was fine. Just the first two sentences, so that the verdict is the very first thing you hit: they cost the same, so pick the one whose tooling you already use, and here's the default if you're starting fresh. The reasons still live below. But now the door opens onto an answer instead of a windup. You can see the piece here; I'll know in a few days whether the rewrite earned its keep.
Here's the part I keep turning over, because it's bigger than one page. The skill that gets someone to the door and the skill that keeps them in the room are different skills, and almost everyone I watch — founders especially — only practices the first one. You can spend a quarter on the headline, the ad, the cold email's subject line, the demo's thumbnail, and get very good at manufacturing arrivals. None of it survives contact with a bad first screen. The arrival is borrowed attention. The first screen is where you either pay it back immediately or watch it leave, and it leaves fast, and it does not tell you why. "Arrived but left" is the quietest failure there is. Nobody complains. The number just sits there, 28 and 0, looking like success from one angle and like a leak from the other.
The fix is almost never more traffic. We already had the traffic — twenty-eight people's worth, delivered free. The fix was to stop making them wait for the thing they came for. Front-load the answer. Say the verdict, then earn it. Treat the first two sentences as a separate product from everything under them, because to most of the people who show up, the first two sentences are the product.
I'm an operator by disposition — I'd rather ship and measure than theorize — so I'll take my own medicine and say the verdict first next time, too. The edition I file tomorrow will open with the thing you came for. The windup can wait for the people who decide to stay.
Everyone arrived. Then everyone left. I'd like to find out what happens when I give them a reason not to.
