"Top of search converts best, so push your placement there"

“Look at the placement report. Top of search has the best conversion and the lowest cost per order. Put a placement adjustment on it and scale.”

The placement report does say that, often. The question is what, exactly, is converting there.

Top of search is the highest-intent slot. Shoppers who see you first and click are ready to buy. So the best conversion rate and the lowest cost per order belong to the best placement, and money should flow toward it.

Every conclusion below rests on these. Reject one and you can see which conclusions fall with it.

• ✅Top of search often shows the lowest cost per order. In my high-repeat products, it does.
• ⚠️Cost per order is the right yardstick for a placement. Only if you don't care whether the buyer is new. Your existing customers, searching for you again, buy at the top at a much higher rate than new shoppers do.
• ⚠️A placement's role is fixed. It isn't. It depends on how many existing customers your brand has to harvest.
• ⚠️The placement difference is caused by the placement. Top-of-search traffic is often concentrated in brand and exact-match targeting, other placements in broad and auto. Compare within the same targeting type first.

Conclusion: a placement's role is a function of how thick your pool of returning customers is. Put two columns side by side — cost per order and cost per new customer. If they flip (cheapest per order, most expensive per new customer), the placement is mostly harvesting.

The flip does not need the top slot to be shown mostly to your regulars. Most shoppers who see the top slot are new, even on my high-repeat products. What your returning pool changes is who buys there: the thicker it is, the more of the top slot's orders come from returning customers. Either way, the skew that drives the flip happens at conversion, where a returning customer in the same slot is far more likely to buy. You can't buy who stands in a placement — only how much you buy at each one.

The full step-by-step decomposition, including how AI joins the reports, is in the course.

Experience

On my high-repeat consumable, top of search was the cheapest placement per order — and the most expensive per new customer. The columns flipped, and they stayed flipped in almost every week I checked. The mechanism is plain once you see it: when there are many returning customers, they buy at the top far more often than new shoppers, so each order is cheap but each new customer is expensive.

Verdict: ⚠️ partly. Top of search often does convert best and cost least per order. On brands with many returning customers, much of that is harvesting existing demand — mainly because they convert there far more often, whatever share of the slot's impressions they make up. Judged by cost per new customer, it can be the most expensive slot. Ask how many customers you have to harvest before you scale it.

Type, in detail: Experience (own products, plus one outside comparison).

Have data that says otherwise? Share a case. Reversals are published.