Amazon PPC Strategy for Thin Margins (2026)

Most Amazon PPC strategy advice assumes you have room to be wrong. Test broadly, let campaigns gather data, accept a high ACoS while you "buy rank," optimise later.

That advice is written for sellers keeping $12 a unit. If you keep $2.02, none of it survives contact with your bank account.

This is what strategy looks like when your margin is a royalty — Merch on Demand, KDP, or any product where Amazon pays you a fixed cut rather than leaving you a margin to manage.

Start from the ceiling, not the benchmark

The standard approach picks a target ACoS from a benchmark and works down. Skip that. Work out the number above which every sale loses money, and treat it as a wall.

Break-even ACoS = royalty ÷ sale price

A $16.99 shirt paying $2.02 breaks even at 11.9%. Not the 25–30% that general Amazon PPC guides quote, because those numbers come from private-label economics where a $30 product carries $9 of margin.

Then subtract returns. A returned unit takes its royalty back; the click that sold it does not come back. Across one real Merch account over 60 days we measured 149 returns against 1,645 units — a 9.1% return rate, with individual designs past 20%. That turns $2.02 into an effective $1.84 and the real break-even into 10.8%.

Convert that to something you can actually type into Amazon:

Max bid = effective royalty × conversion rate
        = $1.84 × 10% = $0.18

Every decision below is downstream of that number. If you take nothing else from this, take the habit of computing it before you touch a bid — the free max bid calculator does it in about ten seconds.

Strategy 1: fewer designs advertised, not more

The instinct with a large catalogue is to advertise everything at a low bid and let the winners emerge. On thin royalties that instinct is expensive.

At $0.18 a click and a 10% conversion rate, a design needs roughly 10 clicks to produce one sale that breaks even. Spread $100 across 300 designs and each one gets a third of a click. You learn nothing about any of them and you have spent $100.

Concentrate instead. Advertise your top 10–20 sellers properly — enough budget that each accumulates 40–50 clicks — and leave the long tail to organic. You will get statistically meaningful data on the designs that can actually carry ad spend, which is the only data worth having.

Strategy 2: one ASIN per ad group

If several designs share an ad group, they share a bid. A design converting at 15% and one converting at 2% get identical treatment, and the good one subsidises the bad one indefinitely.

Splitting each ASIN into its own ad group is the single highest-leverage structural change available to a Merch seller. It's tedious to do by hand across a catalogue, which is why most people don't — we wrote up the mechanics in The Merch Lottery Campaign Setup.

Strategy 3: treat the four targeting types as four campaigns

Automatic campaigns bid on close match, loose match, substitutes and complements. Amazon reports them together. They behave nothing alike.

Substitutes puts your design on a competitor's product page, where a shopper is already looking at a shirt. It usually does most of the converting. Complements bids you into a basket someone has already assembled — reasonable for the second book in a series, poor for an unknown design.

Blended into one campaign ACoS, a targeting type can burn budget for months without ever converting and you would not see it. Split them out before you conclude a campaign is failing; often three types are fine and one is the problem.

Strategy 4: move in small steps, and know why

A bid raise resets nothing, but a campaign restructure does. Moving a converting design into a new campaign throws away the conversion history Amazon's auction was using to give it cheap placements — and a better structure will underperform the sloppy one it replaced for three to four weeks while Amazon re-learns.

So: adjust bids often, restructure rarely, and never restructure and re-bid in the same week. You will not know which one caused what.

When you do adjust, cap the step. A 10% raise per run compounds to 33% over three runs, and three runs is a fortnight. Without a ceiling tied to your break-even, a rule that raises bids on low-impression targets will walk you past profitability one reasonable-looking increase at a time.

Strategy 5: judge on profit, not ACoS

ACoS tells you what percentage of ad revenue went to ads. It does not know your royalty, it does not know about returns, and it does not know the organic sales your ads pulled along.

A campaign at 9% ACoS on a design with a 25% return rate can be losing money. A campaign at 40% ACoS on a design where every ad sale brings two organic sales can be your best performer. Judged on ACoS alone, you would pause the second and scale the first.

What you actually want is royalty minus ad spend, per campaign, with returns subtracted. That number is not in Amazon's reports — it needs your sales report joined to your ad data by ASIN.

What a first month looks like

Compute break-even ACoS and max bid. Write both down.

Pick 15 designs with existing organic sales. Existing sales mean the listing converts; you are adding fuel, not testing whether it burns.

One auto campaign per design, one ASIN per ad group, bid at roughly 60% of your max bid.

Run three weeks untouched. Resist the urge. You are buying data.

Pull the bulk file and split performance by targeting type and by design. Pause what has clicks and no orders. Raise bids — by no more than 10% — only where profit after returns is positive.

Promote proven search terms into manual campaigns and negate them in the auto campaign so the two stop bidding against each other.

Most sellers do step 6 first and wonder why nothing works.

The part that is genuinely hard

None of the above is intellectually difficult. It is difficult because Amazon does not give you the numbers in a usable shape. Royalty lives in your Merch sales report, ad spend lives in the Ads console, returns live in a third place, and nothing joins them by ASIN.

That join is the whole job, and it's what PPC Optimizer Pro does: it reads your Amazon Ads bulk file alongside your Merch or KDP sales report, attributes real royalty per design, subtracts returns, shows organic halo separately, caps bid changes against your break-even, and hands back a bulk file ready to upload.

If you would rather do it by hand, the formulas above are all of it. The tool just saves you the afternoon.

Amazon PPC Strategy When Your Margin Is a Royalty