A brand came to us last year with a 6.2x return on ad spend and no idea why they kept running out of cash. Their previous agency was not lying. The number on the deck was real. It was also almost entirely made of people who had already decided to buy.
Branded search is the cheapest traffic on Amazon. Somebody types your name, you win the auction at forty cents, they buy the thing they came for. That converts at thirty or forty percent and returns eight or nine times what it cost. Blend it into the account total and it will carry an enormous amount of waste on its back.
Pull it apart
Take your search term report for the last ninety days and split it into two buckets: terms containing your brand name or an obvious misspelling of it, and everything else. Then work out return on ad spend for each bucket separately.
Most accounts we look at come back somewhere in this shape. Branded returns 7x to 11x and eats fifteen to thirty percent of the budget. Non brand returns 1.8x to 2.6x. The blended number lands around 4x and looks respectable, which is the problem, because the only number you can actually grow with is the second one.
The uncomfortable follow up question
Once you have the branded figure on its own, ask what it would have done at zero spend. Not all of it disappears. Some of those buyers would have scrolled past a competitor ad and bought you anyway, because they typed your name for a reason.
You can test this. Turn branded search off on one product for two weeks and watch total revenue on that ASIN, not advertising revenue. We have done this on eleven accounts. The usual answer is that somewhere between fifty and seventy percent of branded conversions happen anyway, which means half of that spend was buying customers you already had.
That does not mean turn it off. A competitor bidding on your name while you sit it out is a real risk, and defending the top slot on your own brand is worth something. It means put a ceiling on it, treat it as an insurance premium rather than a growth channel, and stop counting it when you judge whether advertising is working.
What to report instead
Three numbers, weekly:
- Non brand return on ad spend, which tells you whether you can buy new customers profitably
- TACoS, total ad spend over total revenue, which tells you whether advertising is a growing or shrinking share of the business
- Contribution margin per unit on your top five products, which tells you whether any of it matters
The account we mentioned at the top now reports a 2.1x non brand return. That is a much worse looking number and the first one that ever helped them make a decision.
