Amazon PPC in 2026: Structure Beats Spend

Audit a hundred Amazon ad accounts and you'll find the same disease in ninety: not too little spend, not even too much — spend with no architecture. Campaigns named "SP_Auto_2", brand terms cannibalising organic rank, and no one able to answer the only question that matters: what is this rupee of ad spend supposed to do?

Every rupee gets a job

We structure accounts so every campaign has exactly one of four jobs:

1. Defend — brand terms and PDP placements protecting your shelf from conquesting. Judged on share-of-shelf, not ROAS (its ROAS will look amazing; that's not the point).
2. Win rank — aggressive exact-match pushes on the 5–15 keywords that decide your category. Judged on organic rank movement. This campaign is supposed to look expensive.
3. Harvest — exact-match on proven converters from search-term reports. Judged on ROAS. This is where efficiency lives.
4. Discover — autos and broads feeding the harvest loop. Small, capped, always-on.

When every campaign has one job, budget meetings become strategy meetings. "Cut the expensive campaign" stops being a sentence anyone says, because the expensive campaign is buying rank that pays organic dividends for months.

Govern TACoS, not ACoS

ACoS judges ads in isolation; TACoS judges the system. A healthy growth account often shows TACoS falling while ad spend rises — because rank campaigns are converting paid demand into organic demand. Set TACoS corridors by SKU lifecycle:

• Launch: 20–35% — you're buying rank and reviews.
• Growth: 10–18% — rank campaigns still running, harvest scaling.
• Defend: 5–10% — defence and harvest only.

The weekly loop

Search-term harvest every week. Negative-match hygiene every week. Rank tracking on the deciding keywords every week. Bid changes from data, not vibes. It's unglamorous, which is exactly why it compounds — most accounts simply don't do it consistently.

What this looks like at scale

Across the portfolio we operate, structure-first rebuilds typically cut TACoS 25–40% within two quarters while growing revenue — the Harley-Davidson merchandise program saw a 42% TACoS reduction during a 3.1x revenue scale-up. Architecture did that, not budget.

Spend is a commodity. Structure is the moat.