Conversion Can Go Negative

Every team that sends a campaign measures the same thing: how many people clicked, how many bought, how much revenue showed up in the attribution window. That number is always positive, or close enough, so the campaign gets a thumbs-up and goes into the playbook.

Almost nobody measures the other side of the ledger — the people who saw the message, decided you don’t respect their inbox, and quietly stepped back. They don’t bounce in a way the dashboard flags. They mute, they skim past, they stop opening. Some of them were going to buy this month anyway. The blast cost you that, and it never shows up as a cost. Conversion isn’t a floor at zero. It’s a number that can go negative.

When you send to the whole list, you’re paying a tax on every contact who didn’t need the message. Most of them feel nothing and forget it. But a slice of them — often your best, most engaged people, the ones most likely to read — feel pestered. Engagement is a renewable resource right up until it isn’t, and the heaviest senders are the ones who burn it fastest.

I’ve watched the math on this in production at scale — roughly two million communications a month across seventeen countries. The reflex is to reach for volume because volume is the lever you can see. But on any time horizon long enough to matter, the cost of the irritated buyer compounds faster than the win from the marginal click. You’re trading a durable relationship for a one-day spike, and calling the spike a success.

Part of what I’m paid for is the send that never happens. A client comes in wanting a "big push" — everyone, today, same offer. Often the honest answer is: this will cost you more in goodwill than it returns in orders, and here’s the model that shows it. Saying no to that send is not me withholding work. It’s the work.

Indiscriminate discounting is the loudest version of the same mistake — a margin crime dressed up as a marketing win. You hand money to people who would have paid full price and train everyone else to wait for the next blast. The campaign you don’t run protects the P&L exactly as much as the one you do.

The alternative isn’t sending less for its own sake. It’s sending into readiness. Trigger on behavior, not the calendar — the message fires because a person did something that says they’re close, not because it’s Tuesday. Sequence across channels so the same intent is met where the customer actually is: email here, push there, WhatsApp when it earns the room. At Uber Eats, tighter targeting moved click-through from 1.53% to 2.28% and open rate from 43.9% to 52.7% — fewer messages, landing on people who wanted them.

Volume is a strategy you can copy in an afternoon. Timing is the one that compounds. Fewer, better-timed messages beat the blast on any curve long enough to matter — and they keep the buyers who would have come to you on their own, which is the conversion nobody counts and everybody loses.