Table of contents
The Hierarchy: Portfolio → Campaign → Ad Group
Every Amazon PPC campaign structure follows four levels, each with one job:
- Portfolio — a budget cap and reporting folder for a product line. Optional, but essential once you run 10+ campaigns.
- Campaign — where the budget lives, plus targeting type (keyword vs product), ad format, and start/end dates.
- Ad group — a themed cluster of keywords or targets sharing a default bid (e.g. “exact – core terms”).
- Keyword / target — where the bid lives, with its match type.
The golden rule: each campaign does exactly one job. Discovery, exact-match winners, brand defense, competitor conquest, retargeting — one per campaign. The moment a campaign has two jobs, you can’t tell which one the budget is serving.
Key takeaway: Budgets live at campaign level, bids at keyword level. Structure campaigns by job , ad groups by theme .
The 4 Amazon PPC Campaign Types (and When to Use Each)
| Type | Best for | Watch out for |
|---|---|---|
| Sponsored Products | Core sales driver; keyword and product targeting on listings and search | Where most waste happens — needs the tightest structure |
| Sponsored Brands | Brand defense, new product launches, owning branded search | Requires Brand Registry; creative matters more here |
| Sponsored Display | Retargeting viewers, competitor conquest on product pages | Different metrics — judge on reach and return visits, not just ACoS |
| Sponsored TV / DSP | Scale-stage awareness (advanced) | Not for beginners — start with the three above |
New sellers: start with Sponsored Products only — one auto discovery campaign plus one manual campaign. Add Sponsored Brands once you’re brand-registered and have converting keywords to defend. Add Sponsored Display when you have traffic worth retargeting.
Match-Type Architecture: Auto, Broad, Phrase, Exact
Match types are your precision dials. The standard architecture separates them so each can be budgeted and judged on its own terms:
- Auto campaign — Amazon’s discovery net. Low bids, tight budget. Its job is finding search terms, not scaling them. Mine its search term report weekly.
- Broad / phrase manual — the testing ground for keyword ideas. Capped bids; every converter graduates upward.
- Exact manual — the profit engine. Only proven converters live here, each with a bid tuned to its own ACoS. This campaign deserves the biggest budget.
Keep match types in separate campaigns (not just separate ad groups): budgets differ wildly between discovery and scaling, and campaign-level negatives let you funnel traffic cleanly — e.g. negate harvested exact terms in the auto campaign so impressions consolidate on the winner.
Try it: Campaign Builder Answer a few questions about your product and goals — it generates this exact structure with names, budgets, and match types. Free, in your browser.
Build my structureNaming Conventions That Keep You Sane
You will manage hundreds of campaigns. Future-you needs names that sort, filter, and explain themselves at a glance. A proven pattern:
Examples: YogaMat – SP – Exact – Core , YogaMat – SP – Auto – Discovery , YogaMat – SB – BrandDefense . Fixed segments in fixed order mean alphabetical sorting groups everything correctly, and anyone — including a future agency or VA — can read the account. Our campaign name builder generates consistent names from your taxonomy automatically.
Budget Allocation Across the Structure
A starting split for a single product line:
- 50–60% — exact winners campaign. Proven converters get the lion’s share; never let winners go out of budget.
- 15–20% — auto discovery. Capped deliberately — discovery is R&D, not the sales team.
- 15–20% — broad/phrase testing. The pipeline feeding future winners.
- 10–15% — brand defense + retargeting. Cheap clicks, high intent.
Rebalance monthly: shift budget from campaigns missing their ACoS target toward campaigns beating it. And set a portfolio budget cap as the circuit breaker — the one number that guarantees you never wake up to a surprise spend spike.
Scaling: When to Split and When to Consolidate
Split a campaign when segments need different budgets or targets — branded vs non-branded, hero ASIN vs long tail, US vs EU. Consolidate when campaigns are too thin to learn: an ad group with 3 keywords and 40 impressions a month will never produce a statistically meaningful ACoS.
The test is data density: any keyword making bid decisions needs enough clicks to judge. If your structure has outgrown your spend, merge ad groups until each one earns decisions weekly. Structure serves the data — not the other way around.
Key takeaway: One job per campaign, themes per ad group, density before granularity. Name everything consistently, cap the portfolio, and let the winners eat.
Related reading
- Amazon Sponsored Products Campaign: 2026 Setup Guide — Covers amazon sponsored products campaign.
- Campaign Naming Convention: Amazon PPC Best Practices — Covers campaign naming convention.
- Amazon Advertise: The 2026 Seller's PPC Guide — Covers amazon advertise.
- How Much Does Amazon PPC Cost? 2026 Guide — Covers how much does amazon ppc.
- Campaign Targeting: The 2026 Amazon Seller Guide — Covers campaign targeting.
- Automatic vs Manual Campaign Amazon: Which Wins in 2026 — Covers automatic vs manual campaign amazon.
- Branded Campaign Guide: Build One That Converts in 2026 — Covers branded campaign guide.
Amazon PPC Campaign Structure FAQ
Start lean: one auto discovery campaign (low bids, capped budget) plus one manual campaign with a handful of researched exact keywords. Mine the auto campaign’s search terms weekly, harvest converters into the manual campaign, and only add Sponsored Brands/Display once the basics convert. Complexity should follow data — not precede it.
5–15 tightly themed keywords per ad group is the practical range. Fewer and the ad group can’t gather meaningful data; many more and bids can’t be tuned to individual performance. Split ad groups by theme (e.g. “core product terms” vs “use-case terms”), not by arbitrary count.
Yes — always. Auto is discovery (low bids, capped budget, mined for search terms); manual exact is scaling (proven converters, full budget, tuned bids). Mixing them in one campaign means one budget serves two opposite jobs, and you can’t use campaign-level negatives to funnel traffic cleanly between them.
A portfolio is a folder for grouping campaigns — typically one per product line — with an optional budget cap across everything inside it. Portfolios don’t change how ads serve; they’re an organization and spend-control tool. Use one per product line once you run more than a handful of campaigns.
A single product can thrive on 3–5 campaigns : auto discovery, manual exact winners, broad/phrase testing, plus brand defense and retargeting as you grow. Add campaigns when a segment needs its own budget or ACoS target — not before. More campaigns than your spend can feed just starves every one of data.