Amazon PPC Campaign Structure: The Framework That Scales
A scalable Amazon PPC structure separates match types, splits branded from non-branded traffic, and uses portfolios to control budget by product group. Auto campaigns run for discovery, not performance. Without this separation, spend data becomes unreadable and optimization becomes guesswork.
Campaign structure: The organizational hierarchy of your Amazon advertising account — portfolios, campaigns, ad groups, targeting. How you build this hierarchy determines how granularly you control bids and budgets, and whether your performance data is readable enough to optimize from.
Why Most Campaign Structures Break Under Scale
The most common structure for a seller with 20–50 ASINs is some variation of this: one or two broad auto campaigns containing most of the catalog, occasionally a manual campaign with a handful of keywords pulled from the top-performers report. The campaigns have names like “Auto Campaign 1” and “Manual - Main,” and the ad groups inside them were never separated by match type.
This structure has one characteristic quality: it works well enough to hide that it is not working.
Sales come in. The account appears active. But when you pull the search term report, you cannot tell which keywords are driving revenue for which products. When you want to suppress a keyword that is converting poorly for Product A but well for Product B, you have no mechanism — they share the same ad group. When your quarterly planning asks you to allocate budget between product lines, the campaigns do not map to product lines.
The structure was never designed for the account it is now running. It was built when the catalog was small and never rebuilt. The cost of not rebuilding it compounds every week in wasted spend, suboptimal bids, and decisions made without sufficient data. Fixing it starts at the top of the organizational hierarchy.
The Portfolio Layer: Organizing by Business Unit
The hierarchy starts at the portfolio level. Amazon Portfolios allow you to group campaigns together for budget management, shared reporting, and organization.
The most useful portfolio structure maps to how you think about the business — by product line, by brand, or by margin tier. A seller with three product categories should have three portfolios. A seller with a high-margin flagship product and a catalog of accessories should separate them because the acceptable ACoS for each is different, and mixing them in a single portfolio makes budget allocation impossible to optimize.
Portfolios also enable budget caps at the group level — you can set a monthly budget for a portfolio that prevents individual campaigns from cannibalizing each other’s spend. For sellers managing multiple product lines with different margin profiles, portfolio-level budget control is the mechanism that stops a high-volume, low-margin product from absorbing budget that should be going to a lower-volume, high-margin one.
Name conventions matter at this layer and cascade downward. A naming convention like [Brand/Line] | [Campaign Type] | [Match Type] makes the account readable without opening individual campaigns to understand what they contain. Below the portfolio layer, the most structurally important decision is how traffic divides between branded and non-branded terms.
Separating Branded and Non-Branded Traffic
The most consequential structural separation in an Amazon PPC account is between branded and non-branded campaigns. They are fundamentally different acquisition channels and should never share a campaign.
Branded campaigns target keywords that include your brand name. Shoppers who search your brand name are already aware of you — they have lower acquisition costs, higher conversion rates, and represent a different decision stage than shoppers searching for a product category. The economics of branded keywords are typically much better than category keywords, which means mixing them into the same campaigns as non-branded search terms hides how well you are actually performing on each.
A non-branded campaign targeting “stainless steel travel mug” is competing for awareness and consideration. A branded campaign targeting “[Your Brand] travel mug” is defending against competitors who bid on your brand name and capturing customers already looking for you. The bid strategies, ACoS targets, and optimization logic are different for each. Running them in the same campaign makes neither optimizable.
Non-branded traffic also carries the keyword discovery value that informs your entire strategy. The search terms that convert on non-branded campaigns are the keywords you eventually want to move to exact match and bid aggressively. That intelligence is obscured when branded and non-branded terms share reporting. The next separation happens within each campaign, at the match type level.
The Match Type Layer: Where Optimization Actually Happens
Within each campaign, separating keywords by match type is the difference between having control over bids and not having control.
The standard structure runs three campaigns per product group per targeting type: one for exact match, one for phrase match, one for broad match. The rationale is bid control: exact match keywords that are proven converters deserve higher bids than broad match terms that are still being tested. If they share a campaign, Amazon’s auction dynamics mean the algorithm decides which match type to apply — you lose the ability to make that decision yourself.
Exact match campaigns carry your top-performing, highest-confidence keywords. Bids are set aggressively because these are proven terms. Phrase match campaigns capture variations of those terms without requiring you to enumerate every possible search query. Broad match campaigns are the testing ground — wider reach, lower bids, high negative keyword discipline to prevent irrelevant traffic from burning budget.
Auto campaigns sit alongside this manual structure, not inside it. Their role is not performance — it is discovery. An auto campaign allowed to run with reasonable daily budget and aggressive negative keyword harvesting surfaces new search terms you did not know to target. The search term report from auto campaigns is the primary input to your manual campaign keyword strategy. When a term converts in auto, it graduates to exact match in a manual campaign with a controlled bid. The auto campaign continues discovering while the manual campaigns optimize.
Negative Keywords: The Underbuilt Half of the Structure
Structure without negative keyword discipline is not structure — it is organized waste. Negative keywords are what make match type separation work by preventing the same query from triggering multiple campaigns simultaneously.
At the portfolio level, negative keywords block irrelevant search categories from entering any campaign in the group. If your product has nothing to do with a specific competitor’s brand name but broad match terms keep triggering on it, a portfolio-level negative resolves it once rather than duplicating the negative across every campaign.
At the campaign level, cross-campaign negatives prevent cannibalization. When a keyword is added to your exact match campaign, add it as a phrase match negative to your phrase and broad campaigns. This forces Amazon to serve the exact match version — the one you control — when the query is exact. Without this step, Amazon may choose to serve a broad match ad for a query that should be served by your exact match campaign at a lower bid.
The negative keyword audit is a weekly task, not a setup task. The search term report from auto and broad campaigns produces new terms to evaluate every week. A term with ten or more clicks and no conversions is a negative keyword candidate. A term irrelevant to your product is an immediate negative regardless of clicks. Skipping this audit is the single fastest way to watch ad spend erode without visible cause. What that erosion looks like when it has been running unchecked is the subject of the scenario below.
Operational Scenario: Rebuilding from Flat Structure
A seller with 50 ASINs in the outdoor furniture category had been running Amazon ads for 14 months. The account had four campaigns: one auto, one manual broad, one manual phrase, and one for Sponsored Brands. All 50 ASINs were in every campaign.
The symptoms were clear in retrospect. The account showed a 28% overall ACoS — acceptable on the surface — but no way to tell which products were driving that number. Three products in the catalog were running at ACoS above 60%, subsidized by ten products running at ACoS below 18%, all averaged into a number that looked fine. Monthly ad spend was $9,000, but the distribution across products had never been intentionally set.
The restructure took three weeks. Portfolios were created for two product lines (seating and storage). Within each portfolio, campaigns were separated by match type. Branded terms were pulled out into dedicated campaigns. The auto campaigns were narrowed to one per product line, with aggressive negatives applied based on 14 months of search term data.
Six weeks after the restructure, the 60% ACoS products were visible as problems, not hidden in averages. Two were paused for listing issues unrelated to PPC. One was restructured with different targeting. The average account ACoS dropped — not because bids changed dramatically, but because budget was redirected toward the products that could actually use it.
The structure did not improve the products. It made the data readable, and readable data made decisions possible that had been invisible before. For sellers at this stage, the Amazon PPC advertising service covers what a full account review looks like in practice.
FAQ
How many Amazon PPC campaigns do I need? The number depends on your catalog size and how many product lines you have. A minimal but functional structure for a single product line includes: one auto campaign for discovery, one exact match manual campaign, one phrase match manual campaign, one broad match manual campaign, and one branded campaign. Multiply by product line if you have multiple categories with different margin profiles. More campaigns than this is often justified; fewer is usually a sign the structure has not been built for optimization.
Should I have separate campaigns for Sponsored Products and Sponsored Brands? Yes — they are different ad formats with different placements, cost structures, and conversion behaviors. Sponsored Products campaigns appear on search results pages and product detail pages; Sponsored Brands campaigns appear at the top of search with brand imagery. They serve different functions in the funnel and require separate budget allocation, not a shared campaign with mixed ad types.
What is a portfolio in Amazon PPC and do I need one? An Amazon Portfolio is a grouping layer above campaigns that allows shared budget caps, combined reporting, and account organization. You need portfolios once you have more than one meaningful campaign group — typically when you have multiple product lines, multiple brands, or need to enforce budget limits by category. A seller with a single product and three match-type campaigns does not need a portfolio. A seller with 30 ASINs across two product categories should have two portfolios with campaigns nested inside each.
Why do I need separate campaigns for each match type? Match type separation gives you bid control. Exact match keywords that are proven converters should have higher bids than broad match terms being tested — but if they share a campaign, you cannot apply different bid strategies to each. Separating them lets you bid aggressively on what works and conservatively on what you are testing. It also clarifies performance data: you can see exactly how your proven keywords are performing versus your discovery terms, rather than having them averaged together.
How often should I add negative keywords? The negative keyword review should happen weekly, using the search term report from auto and broad match campaigns as the primary input. Any term with ten or more clicks and zero conversions is a candidate. Any term clearly irrelevant to your product is an immediate negative regardless of click count. Sellers who run a negative keyword audit less frequently than biweekly are typically accumulating spend on irrelevant queries without visibility into where it is going.
Can I start with a simpler structure and rebuild later? Technically yes — but the rebuild cost compounds over time. Campaigns that have run for months with mixed products and match types accumulate historical data (bids, quality scores, impression history) that cannot be transferred cleanly to a new structure. The disruption of rebuilding after 12 months of mixed structure is substantially higher than building a clean structure from the start and expanding it incrementally. If you are setting up ads for the first time, build the structure correctly now — the [Amazon PPC fundamentals guide]({PENDING: C05-001}) covers the foundational setup before structure becomes the optimization priority. If you are rebuilding, plan for 4–6 weeks before the new structure produces comparable performance data to what you had.
If your current campaign structure has grown beyond your ability to see which products are actually profitable on ads, request a PPC account audit.