Most underperforming Amazon ad accounts we audit don’t have a bidding problem. They have a structure problem. The bids are getting blamed for what the architecture broke.
After managing hundreds of brands, the pattern is consistent: when an account is built as a pile of campaigns thrown together over two years — a campaign for every launch, every promo, every “let’s test this” — you can’t read it, you can’t control budget by intent, and you can’t tell which dollars are defending revenue versus chasing it. The reports lie because the structure makes them lie. You see a blended 14% ACOS and feel fine while branded defense quietly subsidizes a non-branded campaign hemorrhaging money.
This is the Amazon PPC campaign structure that fixes that — the account architecture we rebuild toward, why each layer exists, and the budget-control logic that makes the whole thing readable.
Why structure beats bidding
Here’s the uncomfortable truth: you cannot optimize what you can’t segment. Bid strategy, dayparting, budget reallocation — every lever you have only works if campaigns are organized so each one answers a single question.
A campaign that mixes your branded terms, three non-branded keywords, and two competitor ASINs is unmanageable. Branded converts at 25%+ and costs pennies. Non-branded converts at 8% and costs real money. Competitor conquest converts at 4% and costs the most. Blend them in one campaign and the average tells you nothing — you’ll never know whether to push, pull, or pause, because the campaign is doing three jobs with one budget and one bid.
The first principle of account architecture: one campaign, one intent, one job. Everything else follows from that.
The 4-tier intent architecture
We organize Sponsored Products around four intent tiers, kept in separate campaigns and separate portfolios. Each tier has a different economic purpose, a different ACOS expectation, and a different budget logic.
Tier 1: Branded defense
Purpose: own your own name. When someone searches your brand, you appear, and you don’t let a competitor’s conquest ad sit above you on your own term.
ACOS expectation: very low — often 5–10%. This is the cheapest, highest-converting traffic you have.
The argument you’ll hear: “I’d rank organically for my own brand anyway, so this is wasted spend.” Sometimes true for a dominant brand on a clean SERP. But branded terms are exactly where competitors bid to steal high-intent shoppers who already want you. Branded defense is cheap insurance, and the right move is usually Sponsored Brands for branded terms (lower effective CPC at the brand level) with a thin Sponsored Products layer underneath. Keep it isolated so its rock-bottom ACOS never masks problems elsewhere.
Tier 2: Non-branded harvest (your bread and butter)
Purpose: capture shoppers searching the category — “magnesium glycinate,” “stainless steel dog bowl” — who don’t know you yet. This is where incremental, net-new revenue lives.
ACOS expectation: at or near your break-even target. This tier is the actual growth engine and where most of your manageable budget should sit.
Structure: segment by match type and by performance. Proven converters get their own exact-match campaigns with tighter bids; discovery happens in broad/phrase and auto campaigns that feed the exacts. Which brings us to the harvest loop.
Tier 3: Discovery / auto (the feeder)
Purpose: find new converting search terms you didn’t know to target.
ACOS expectation: looser — you’re paying for data, not just sales.
The loop: auto and broad campaigns surface search terms → winners get promoted (“harvested”) into Tier 2 exact campaigns → the discovery campaign gets a negative for that term so it stops paying twice. Without this loop, auto campaigns cannibalize your exacts and you pay for the same term in two places. The harvest-and-negate cycle is the single highest-ROI habit in a well-structured account.
Tier 4: Competitor conquest
Purpose: show up on competitor ASINs and competitor brand terms (where TOS allows) to steal consideration.
ACOS expectation: highest of the four — you’re converting at 3–5% and paying a premium. That’s fine if it’s capped and watched.
The danger: conquest is where budgets quietly bleed. It’s seductive — “let’s take their customers” — and it’s the tier most likely to run a 40% ACOS while feeling strategic. Keep it in its own portfolio with a hard budget cap so it can never grow past the role you assigned it.
Portfolios are your budget control panel
Campaigns do the targeting. Portfolios do the governing. This is the layer most accounts skip, and it’s the one that makes the architecture actually controllable.
Group campaigns into portfolios by intent tier (or by product line within tier for bigger catalogs), then use portfolio-level budget caps to enforce the role of each tier:
- Branded portfolio: capped low. It should never consume a big share of spend; if it’s ballooning, a competitor is attacking your brand term and you want to know.
- Non-branded portfolio: your biggest allocation. This is the engine.
- Discovery portfolio: a deliberate, fixed “research budget” you’re comfortable spending to learn.
- Conquest portfolio: hard-capped. This is the one most likely to run away from you.
The payoff: you can answer “how much am I spending to defend versus to grow versus to learn versus to attack?” in one screen. In a pile-of-campaigns account, that question is unanswerable — which is exactly why those accounts drift to a 15%+ TACoS without anyone noticing the mix went wrong.
How granular should campaigns get?
The eternal debate: single-keyword campaigns (SKAGs) versus themed campaigns. Our position in 2026:
- Go granular where the money is. Your top revenue keywords deserve their own exact-match campaigns so you can control bid and budget at the term level. These are worth the management overhead.
- Theme the long tail. Hundreds of low-volume keywords in individual campaigns is unmanageable and starves each one of the data it needs to optimize. Group them into themed campaigns.
- Let volume decide. A keyword earning real spend and sales graduates to its own campaign. One earning a click a week stays in a themed bucket. Granularity should follow revenue, not ideology.
Over-granular accounts fail a different way than messy ones: 300 single-keyword campaigns each with three clicks a week never accumulate enough signal to optimize, and the management time explodes. Structure for readability and control, not for maximum fragmentation.
Don’t forget the ad types in the structure
Account architecture isn’t only Sponsored Products. Slot the other formats into the intent tiers deliberately:
- Sponsored Brands / Sponsored Brands Video — strongest on branded defense (Tier 1) and category/non-branded discovery (Tier 2). SBV in particular carries far higher CTR than static placements and belongs in your top-of-funnel and brand-defense layers.
- Sponsored Display — retargeting (own-ASIN views, cart abandoners) and defensive/conquest placements on product pages. Keep it capped; it’s easy to overspend on broad audiences.
The point isn’t to run every format everywhere. It’s to assign each format to the tier where its economics work, inside the same four-tier logic.
Migrating without torching performance
If your account is currently a mess, don’t blow it up overnight — you’ll reset learning and tank performance for weeks. Migrate deliberately:
Avoid restructuring in the two weeks before a major event like Prime Day — freeze the structure and let it run.
FAQ
How many campaigns should an Amazon account have?
There’s no magic number — it scales with catalog size and revenue. The right test isn’t count, it’s readability: can you look at any single campaign and state its one job, its intent tier, and its ACOS target? If yes, you have enough structure. If campaigns are doing multiple jobs, you have too few (or messy ones).
What’s the difference between portfolios and campaigns for budget control?
Campaigns set targeting and per-campaign budgets; portfolios let you cap and govern spend across a group of campaigns by intent. Portfolios are how you enforce “branded stays small, conquest stays capped, non-branded gets the bulk” without babysitting every campaign individually.
Should branded and non-branded be in separate campaigns?
Always. They convert and cost so differently that blending them makes both unreadable. Separation is the single most important structural decision in the account — it’s what lets you see whether branded defense is quietly subsidizing a losing non-branded or conquest campaign.
Are single-keyword campaigns (SKAGs) still worth it in 2026?
For your top revenue keywords, yes — granular control pays off where the spend is real. For the long tail, no — themed campaigns aggregate enough data to optimize and save you from managing hundreds of starved campaigns. Granularity should follow revenue.
How often should I restructure my PPC account?
Restructure is a once-in-a-while reset, not a routine. Build it right, then maintain it with the weekly harvest-and-negate loop and monthly budget reallocation. If you’re restructuring often, the architecture wasn’t built to be readable in the first place.
—
A clean account structure won’t lower your bids for you — but it’s the thing that lets every other lever work. When each campaign has one job, each portfolio enforces one role, and the harvest loop keeps discovery feeding your exacts, you can finally read your own account and act on what it tells you.
If you’re looking for a team that manages every lever — creative, advertising, and operations — Velocity Sellers works with brands doing $100K+/month on Amazon. Contact us for a free account audit.
