Hey,

Quick one today.

I was digging through the promo history of the hobby brand I mentioned on Tuesday, and one pattern jumped out.

Their short flash sales, the 3 to 4 day ones, beat their week-long promos on return on ad spend. Consistently.

Their best promo ROAS of the entire year came from one of the shortest sales they ran. Same brand, same audience, comparable offers.

The difference was the length.

Here's what a promo actually looks like when you plot it by day.

Day one spikes, because the announcement hits every channel at once. The final day spikes harder, because nothing moves people like a deadline.

And the middle sags.

On a 4-day sale, the middle is two days. On a 14-day sale, the middle is a week and a half of discounted margin with no urgency doing any work.

That sagging middle costs you twice. You're selling at reduced margin to people who would have bought anyway, and your "deadline" is so far away that it creates zero pressure.

A discount without urgency is a price cut wearing a costume.

Their data made this embarrassingly visible. Plotted daily, every long promo looked like a valley between two hills, and each day inside that valley was margin given away after the deadline stopped meaning anything.

Long sales damage the brand too. The longer they run, the more they feel like the store's natural state, and the faster your customers learn the most expensive lesson you can teach them: never pay full price, a code is always coming.

I've written about where that training ends, and it's ugly.

The Fix

Three rules for the next promo you run:

1. Halve the length. Whatever you ran last time, run the same offer in half the days. You keep both spikes, you delete most of the sag, and you protect margin on every mid-sale day that no longer exists.

2. Make the deadline real and specific. "Ends Sunday 11:59pm" beats "this week only." And end it when you said you would, because every extended deadline teaches your list that your deadlines are fiction.

3. Judge the promo on gross profit per order, not revenue. A 14-day sale will usually show more total revenue than a 4-day one. The 4-day version routinely wins on profit, because you gave away margin for fewer days.

Revenue is the vanity metric here. Gross profit per day of discount is the honest score.

One founder pushback I hear: "but the long sale made more money overall."

Run the math on what the extra days actually earned versus what they cost in margin on orders that would have arrived anyway. In every account where we've run this check, the middle days cost more margin than any other revenue in the quarter.

Short, sharp, rare. That's the promo philosophy in three words.

Talk soon,
Andrej

PS: Planning something for Q4 already? Reply "promo" with what you ran last time (mechanic + length), and I'll tell you the first thing I'd change about it.