Amazon Advertising Help: Campaign Structure Guide

TL;DR — Key takeaways

Amazon advertising problems are almost never bidding problems. They’re architecture problems.
Most brand accounts run 85-90% Sponsored Products, 5% Sponsored Display — leaving the warm-shopper retargeting layer nearly empty.
Sponsored Display has three distinct sub-products (views remarketing, purchases remarketing, audiences) that require separate campaigns and different optimization timelines.
Running broad, phrase, and exact match at the same bid is one of the most expensive mistakes in Amazon advertising.
A well-structured account generates 2-3x more attributed sales than a flat, auto-only account on the same budget.
The 2026 framework: Sponsored Products 60-70%, Sponsored Brands + video 15-22%, Sponsored Display 10-18%.

You raised bids last Tuesday. Nothing changed. You added a new auto campaign. Still flat. You hired someone who called themselves an Amazon ads expert, and three months later your ACoS is exactly where it was — except now you’re paying their retainer too. Here’s what surprises me: the brands stuck in this loop almost always have the same problem, and it has nothing to do with bids.

What they’re missing is campaign architecture. Not bid strategy. Not budget size. The structure of the account itself.

Why Amazon advertising is an architecture problem, not a bidding problem

There’s a persistent myth in Amazon advertising: that optimizing performance is fundamentally about adjusting bids, harvesting search terms, and trimming ACoS targets. These things matter. They’re also almost never the reason a brand plateaus.

Here’s what most brand managers miss: Amazon’s advertising platform gives you three structurally different products — Sponsored Products, Sponsored Brands, and Sponsored Display — that operate on completely different auction mechanics, attribution windows, and customer intent signals. When you collapse those differences into a single flat campaign structure, you can’t optimize any of them. You’re essentially running one blunt instrument and wondering why it doesn’t perform like a precision tool.

What we see at Epinium is that brands who restructure from a single auto-heavy setup to a properly segmented SP/SB/SD architecture — with match type isolation and separate optimization logic per campaign type — see meaningful ROAS improvement within 45 days, without touching total ad spend. The budget didn’t change. The structure did.

According to Amazon Advertising’s own case study library, brands that adopt full-funnel campaign strategies consistently outperform single-format advertisers on both new-to-brand metrics and return on ad spend. That’s not a coincidence — it’s what happens when you stop fighting the platform’s structure and start working with it.

The budget allocation framework for 2026

Run the numbers on most brand accounts and you’ll find the same thing: 85-90% in Sponsored Products, a token 5-10% in Sponsored Display, and Sponsored Brands treated as an afterthought. This feels safe. Sponsored Products are the most measurable format, they convert, and everyone understands them. The problem is you’re paying peak CPCs on keyword auctions while leaving the warm-shopper retargeting layer almost completely empty.

A more defensible allocation for 2026 looks like this: Sponsored Products at 60-70% of budget (not 90%), Sponsored Brands plus video at 15-22%, and Sponsored Display at 10-18%. The shift isn’t dramatic, but the impact is disproportionate because you’re no longer competing exclusively in the highest-cost auction format for every single impression.

Sponsored Brands video, in particular, is underpriced relative to its attention value right now. Amazon’s own research suggests video ads drive significantly higher new-to-brand purchase rates than static formats — and most brands are leaving that inventory untouched while fighting over the same Sponsored Products slots as their competitors.

2-3x

more attributed sales from the same budget in a well-structured Amazon account vs a flat, auto-only account

Amazon Advertising 2026

The three Sponsored Display mistakes costing you real money

Sponsored Display is where I see the most structural damage. Not because brands over-invest in it — they dramatically under-invest — but because when they do use it, they almost always configure it wrong.

Mistake one: running a single “auto” SD campaign for everything. Views remarketing, purchases remarketing, and audience targeting are distinct products with distinct unit economics. Views remarketing shows ads to shoppers who viewed your product page but didn’t buy — it’s top-of-funnel retargeting with a typical ROAS stabilization window of 14-21 days. Purchases remarketing targets repeat purchase intent — completely different conversion profile. Audience campaigns (interests, lifestyle segments) are prospecting tools that need 30-45 days to generate meaningful new-to-brand data. Combining all three into one campaign means you can’t read the data, can’t set appropriate bids, and can’t optimize any of them. You get blended numbers that tell you nothing.

Mistake two: not excluding existing customers from views remarketing. This is budget waste that’s invisible until you look for it. If you’re showing SD ads to shoppers who already purchased your product, you’re paying to re-acquire someone who already converted. The fix is a ten-minute exclusion setup. Most brands skip it entirely.

Mistake three — and this one is subtler — treating Sponsored Display as a set-and-forget format. SD campaigns require active creative testing, bid adjustment by audience segment, and regular review of the overlap between your SD audiences and your Sponsored Products keyword coverage. Brands that treat it as passive remarketing miss the targeting sophistication that makes it worth running at all.

The Amazon Sponsored Display product page outlines the segmentation options available — most of which go unused in typical brand accounts.

Match type mistakes that guarantee overpaying

Running broad, phrase, and exact match keywords in the same campaign at the same bid is one of the most common and most expensive mistakes in Amazon advertising. It sounds reasonable — cover all the match types, let Amazon figure out which converts. What actually happens is you pay exact match CPCs for broad match traffic, and you simultaneously starve your best-performing exact match keywords of the budget they’d justify on their own.

The structural fix is match type isolation: separate campaigns for broad (discovery, high negative keyword discipline), phrase (mid-funnel, moderate bids), and exact (proven converters, highest bids, tight ACoS targets). This also gives you clean data. When broad and exact live in the same campaign, you can’t tell which match type is generating the return. When they’re isolated, every metric becomes actionable.

A practical audit: pull your search term report filtered by match type. If your broad match terms are showing the same average CPC as your exact match terms, your bids aren’t differentiated by match type and you’re almost certainly overpaying for lower-intent traffic.

Campaign Type	Primary Use	Recommended Budget %	Key Metric	Common Mistake
Sponsored Products	Keyword-driven conversion	60-70%	ACoS, CVR by match type	All match types same bid, same campaign
Sponsored Brands + Video	Brand awareness, new-to-brand	15-22%	New-to-brand %, VCTR	Skipping video entirely
Sponsored Display — Views	Warm-shopper retargeting	5-10%	ROAS (stable at 14-21 days)	No existing-customer exclusion
Sponsored Display — Purchases	Repeat purchase / cross-sell	3-6%	Repeat purchase rate	Merged with views in single campaign
Sponsored Display — Audiences	Prospecting, new-to-brand	3-5%	New-to-brand % (30-45 day window)	Evaluated too early, killed prematurely

Free Session

Spending on Amazon ads but not seeing the ROAS you expect?

We audit your Sponsored Products, Brands, and Display campaign architecture and find the structural fixes — not just bid adjustments — that move the needle.

Book your free architecture audit →

✓ Free ✓ 30 minutes ✓ No pitch

What changed in 2025-2026: the platform you’re advertising on is not the same one it was two years ago

AI bidding and dynamic optimization

Amazon’s automated bidding has matured significantly. Dynamic bids — down only, up and down, and fixed — now respond faster to real-time conversion signals than they did in 2023. What this means practically: fixed bids on auto campaigns are increasingly a handicap. Brands running fixed bids across all campaigns are leaving efficiency on the table that Amazon’s own bidding system would recover automatically. That said, automated bidding works best in a structured account — feeding bad campaign architecture into an AI bidding layer doesn’t fix the architecture problem, it amplifies it.

Amazon Marketing Cloud democratization

AMC — Amazon’s clean room analytics environment — is no longer exclusive to enterprise spenders. Mid-market brands can now run SQL queries across their own impression, click, and purchase data to answer questions that campaign-level reporting can’t touch: What’s the true path to purchase across SP, SB, and SD touchpoints? Which audience segments have the shortest conversion windows? Brands using AMC for path-to-purchase analysis are building bid and budget decisions on actual multi-touch data rather than last-click attribution guesses.

New-to-brand reporting expansion

Amazon has extended new-to-brand metrics beyond Sponsored Brands into Sponsored Products and Sponsored Display. This is more significant than it sounds. For the first time, brands can measure customer acquisition cost — not just ACoS — across all three formats in a comparable way. If you’re not tracking new-to-brand metrics as a primary KPI alongside ROAS, you’re optimizing for efficiency without measuring growth.

Epinium data

Across the Amazon accounts we manage, brands that restructure from a single auto campaign to a segmented SP/SB/SD architecture with proper match type isolation see an average 40% improvement in attributed ROAS within 45 days — without increasing total ad spend.

Frequently asked questions

How do I know if my Amazon ad structure is wrong?

Three signals are usually enough. First: if your Sponsored Products budget exceeds 85% of total ad spend and Sponsored Display is under 10%, your allocation is almost certainly distorted. Second: pull your search term report and filter for broad match — if you’re seeing high-spend, low-converting terms with no negative keyword action, your discovery campaigns are bleeding. Third: if you have fewer than three separate Sponsored Display campaigns (one per targeting type), you’re running SD as a single blunt instrument rather than the segmented tool it’s designed to be. Any two of these is sufficient reason for a structural audit before you touch bids again.

Should I use auto or manual campaigns?

Both, but with distinct roles. Auto campaigns are discovery tools — they surface search terms you wouldn’t have found manually, feed your negative keyword list, and identify new exact match candidates. They should not be your primary conversion driver. Manual campaigns — especially exact match — are where you scale proven winners at disciplined bids. The most common mistake is using auto campaigns as the primary conversion mechanism because they’re easier to set up. Easy to set up doesn’t mean well-structured. Most brands should run auto campaigns at controlled budget as a permanent research layer, not as the core of their ad strategy.

What is Amazon Marketing Cloud and should I be using it?

AMC is Amazon’s privacy-safe data clean room. It lets you run SQL queries on your own aggregated advertising and purchase data to answer questions that standard campaign reporting can’t. Path-to-purchase analysis, overlap between campaign types, time-to-conversion by audience segment — these are all AMC questions. Until recently it required a significant spend threshold and enterprise onboarding. That’s changed. If you’re spending more than $5,000/month on Amazon ads, the AMC insights available to you are more valuable than any bid optimization tool. Epinium’s Platform connects AMC data into campaign decision-making workflows so brands don’t need a data team to act on it.

When should I NOT run Sponsored Display?

If your product has fewer than 50 orders in the past 30 days, Sponsored Display audiences won’t have enough signal to perform. Views remarketing requires meaningful product page traffic — below roughly 200 sessions/week, your remarketing pool is too small to generate statistically meaningful ROAS data. Purchases remarketing requires a customer base to remarket to. If you’re pre-scale, concentrate budget in Sponsored Products to build the order and traffic history that makes SD viable. Launching SD too early is a common way to conclude it “doesn’t work” when the real issue was insufficient audience size.

How long before my campaigns hit stable ROAS?

It depends sharply on campaign type. Sponsored Products manual campaigns with sufficient search volume typically stabilize within 7-14 days. Views remarketing — Sponsored Display targeting shoppers who viewed your product — reaches stable ROAS in 14-21 days, provided the audience pool is adequate. Sponsored Display audience campaigns (interest and lifestyle targeting) need 30-45 days before you have meaningful new-to-brand data. The most common error is evaluating campaigns on week-one data and killing formats that needed 3-4 more weeks to mature. Set evaluation windows before you launch, not after you get impatient.

What budget do I need to start seeing real data?

For Sponsored Products, $30-50/day per campaign is enough to generate statistically meaningful click and conversion data within two weeks, assuming decent search volume for your keywords. For Sponsored Display, you need enough daily impressions to serve the remarketing pool — $20-30/day minimum per SD campaign type. The more important constraint is account-level: if your total daily budget is under $100, prioritize Sponsored Products exclusively until you’ve built the order history and traffic that makes SD and SB worth layering in. Spreading $50/day across three campaign types gives you data too thin to optimize any of them.

Is a high ACoS always a problem?

No — and treating it as one leads to systematic over-pruning of campaigns that are actually working. ACoS is a profitability metric, not a performance metric. A Sponsored Brands campaign driving 30% ACoS might be generating 70% new-to-brand customers who become repeat purchasers. A Sponsored Products exact match campaign at 12% ACoS might be serving customers who already would have found you organically. The right question isn’t “is my ACoS below target?” but “is this campaign generating profitable customers at a sustainable acquisition cost, and am I measuring the right window?” ACoS optimization without new-to-brand context is optimization in the wrong direction.

How do I audit my own account in 30 minutes?

Start with budget distribution: what percentage sits in each format? If SP exceeds 85%, flag it. Then open Sponsored Display — count your active campaigns. If it’s one, check what targeting type it uses. If it’s “auto,” you’ve confirmed the structural problem. Pull the search term report, sort by spend descending, filter for broad match, look for high-spend terms with zero conversions and no negative keyword applied. Finally, check match type bid differentiation: if your broad and exact bids are within 20% of each other in the same campaign, your match type strategy isn’t doing what you think it is. Those four checks take 25 minutes and tell you more than most 2-hour account reviews.

How does Epinium help with Amazon advertising structure?

Epinium’s Platform automates the structural monitoring that most brands do manually and inconsistently — match type bid differentiation, SD audience segmentation, budget allocation drift, negative keyword management across campaigns. The Transform service goes further: we rebuild campaign architecture from the ground up for brands whose accounts have accumulated structural debt, then hand over a system that’s optimized and maintainable. The audit described above is where every Transform engagement starts.

The Amazon advertising problem most brand managers have isn’t what they think it is. They’re looking at bids when they should be looking at structure. They’re adjusting budgets when they should be asking why three distinct Sponsored Display products are running as a single undifferentiated campaign. They’re hiring bid management help when what they actually need is someone to look at the architecture first.

The platform has gotten more sophisticated — AI bidding, AMC democratization, Sponsored TV, new-to-brand metrics across all formats. None of that sophistication helps a structurally broken account. What surprises me, after seeing hundreds of these accounts, is how consistently the fix is the same: separate the campaign types, isolate the match types, segment Sponsored Display into its three actual products, exclude existing customers from views remarketing, and give the campaigns enough runway to hit their optimization windows. Same budget. Different structure. The results follow.

Ready to fix the structure?

Your Amazon ad budget is ready to work harder.

Book a free 30-minute architecture audit. We identify the structural fixes — not bid tweaks — that generate real ROAS improvement.

Book your free audit →

Amazon Advertising Help: The Structural Fixes Brand Managers Actually Need (Not More Bidding Tips)