How I Bid for Top of Search Without Raising Every Keyword

I pulled my placement report one day and saw something that changed how I bid. Same campaigns, same targets — but split by where the ad actually showed, the numbers weren’t close. Top of Search was converting more than twice as well as the product-page slots. And I was bidding them all the same.

I sell wood-finishing products under Rowdy Rooster Woodworks, so these are my real numbers, last 90 days, split by placement:

• Top of Search — 26.4% conversion rate, 30.1% ACoS, $1.60 per click
• Rest of Search — 13.6% conversion rate, 42.8% ACoS, $1.17 per click
• Product Pages — 10.6% conversion rate, 51.2% ACoS, $1.14 per click

Look at what that means. Top of Search converts 2.5× better than a product-page placement. It also costs the most per click — $1.60 versus $1.14 — and it’s still by far my cheapest customer, because conversion is what actually sets your cost per sale. The cheap clicks down on the product pages are my most expensive sales: a 51% ACoS against Top of Search’s 30%.

A flat bid treats all three the same. Which means a flat bid is quietly overpaying for the worst placement to get a little more of the best one.

The real question: when is the top worth the premium?

The numbers above already make the case — for my catalog, Top of Search is the placement worth paying for. The hard part was never whether it’s valuable. It’s when it’s worth the premium to be there, and how to get there without handing Amazon a blank check for every other placement too. That’s the whole game with placement bidding: be aggressive exactly where you convert, and nowhere else.

The wrong fix: just raise your bid

The obvious move is to raise your bid so you show up at the top more often. Don’t. Your base bid lifts you in every placement at once. Raise it, and yes, you bid more for Top of Search — but you also bid more for the product-page slot that’s already running a 51% ACoS. You’d be paying more for your worst customers to get more of your best ones. That’s backwards, and it’s exactly the problem Amazon’s placement modifiers exist to solve.

What I actually do: isolate the placement

Here’s the play that works on my account. I take my proven targets — the keywords and products I already know convert — and pull them into their own campaign. Then I set the base bid low and put a high Top-of-Search modifier on top of it.

The low base bid is the trick. Amazon runs a separate auction for each placement, and your effective bid is the base times the modifier:

Down in Rest of Search and on product pages, 80 cents is too low to win much — so I mostly don’t show there, which is fine, because those are my weak placements anyway. But with the +100% modifier, that same target bids $1.60 for the top slot — right where it needs to be to win the placement that converts at 26%. The campaign ends up showing almost only at the top, on targets I’ve already proven, at a price I set deliberately. I’m not raising bids everywhere. I’m concentrating them on the one placement that earns it. For my product targets, the rules engine handles most of that decision for me.

That’s the answer to “bid more for Top of Search without raising everything”: you don’t raise your bid, you raise your modifier, and you keep the base low enough to stay out of the placements you don’t want.

Product pages aren’t the enemy — they’re defensive

Notice I don’t bid up product pages at all — every one of my product-page modifiers sits at zero. That doesn’t make product pages worthless. For me they’re a defensive play, same as product targeting: I want to show on competitor and related product pages to defend my turf and catch the occasional browse-to-buy. I just won’t pay a Top-of-Search premium for it, because the shopper on a product page is in a different mindset — browsing, scrolling, comparing, not typing a high-intent search and clicking the first result. Different placement, different job. Show up there cheaply and defensively; spend your real money up top.

The blank-check problem — read this part twice

Now the warning, and it’s the most important paragraph here. A placement modifier doesn’t replace dynamic bidding — it stacks on top of it. Your real bid at the top of search isn’t your base bid, and it isn’t even your base bid plus the modifier. It’s a multiplication:

If you’re running dynamic “up and down” bids and an aggressive Top-of-Search modifier, those multipliers compound. Amazon raises your bid for an impression it thinks will convert, and then the placement modifier doubles that. A bid you set at 80 cents can quietly clear three dollars — and you never typed that number anywhere. Most sellers never see it coming: they set a base bid, set a placement modifier, and assume the modifier is the ceiling. It isn’t.

Dynamic bidding is the multiplier that’s hardest to predict, which is exactly why I run most of my campaigns dynamic down only, and why isolating with a low base matters — the base is the one number both multipliers are anchored to, so keeping it small caps how far the whole thing can run. Top of Search can blow up your budget as fast as it grows your sales.

What we built, and why

This is exactly the kind of decision RedHen Labs is built to make safely. It reads your placement performance from your own account — your Top of Search versus product-page split, not a generic benchmark — and judges each placement on its own numbers, leaving out the newest 3 days while their sales are still coming in. The down-only caution is the default for the campaigns it builds.

In November 2026 we rebuilt the AI’s placement step around the warning above. The AI now sees every placement’s numbers when it sizes your bids, and in any one run a campaign gets a placement raise or bid raises from it, not both, because the two multiply. And when a placement is losing money at 0% — the product pages in my numbers above — there’s no lower setting to give it, so the AI suggests one rebalance: every bid in the campaign a little lower, with Top of Search and Rest of Search raised just enough to keep their bids. That’s this article’s low-base, high-modifier play, worked out from your numbers, kept inside your limits, and approved by you as one change.

Placement rules on the Professional plan read each placement’s own numbers too, and can raise, lower or set any placement, or a campaign’s audience boosts. The AI’s placement suggestions always wait for your approval; a rule’s wait too unless you switch on Run automatically. Both are set up step by step, with screenshots, in the guides to AI bid management and placement rules. The placement lever is powerful enough to respect — so we wrapped it in guardrails instead of pretending it’s safe to set and forget. You can run both rules and an AI layer over it, because some of this is mechanical and some needs a judgment call.

The takeaway

On my account Top of Search converts more than twice as well as anywhere else — but it’s also the fastest way to torch a budget if you bid it blind. The lever isn’t the strategy. The strategy is: find your real placement split, isolate your proven winners and bid them up at the top only, leave product pages cheap and defensive, leave your base bid alone, and put a rule on the whole thing so it can’t run away. Powerful, used carefully. See your real numbers — free for 14 days.