Clear acrylic organizer with three compartments holding a serum bottle, cream jar and tube, tagged structure

How K-Beauty Brands Should Structure Amazon Campaigns in 2026: Category, Product, Funnel, Negatives and Auto Campaigns

9 min
Step into my digital universe
Jeff

Open the Amazon ad console of almost any K-beauty brand that has been selling for a year and you'll find the same thing: dozens of campaigns named "Serum Test 2", "New Auto", "Q3 Push – Copy". The same keyword lives in five of them. Nobody remembers why. Spend keeps growing, and nobody can say which campaign is actually doing the work.

This isn't a bidding problem. It's a structure problem. A clean campaign structure is what makes every other optimization possible: bids, budgets, negatives and reporting all depend on it. This post shows how to organize Amazon campaigns by category, product and funnel stage, how to stop your own campaigns from competing for the same searches, and how to use negatives and auto campaigns as a system instead of a mess.

The Problem: Overlapping Campaigns Hide Waste and Break Your Data

When the same search term is targeted by an exact campaign, a broad campaign and an auto campaign, Amazon still has to choose one of your ads for that auction. You don't decide which one wins. In practice the higher bid or the more "relevant" ad tends to serve, which means your carefully set exact-match bid gets overridden by a discovery campaign you meant to keep cheap.

The damage goes beyond CPC. Performance data for one search term gets split across three or four campaigns, so no single report shows the true picture. Budgets run out in the wrong places. And when someone lowers a bid in one campaign, nothing changes, because another campaign is still picking up the traffic. In accounts we audit, 15–30% of Sponsored Products spend typically goes to search terms that are targeted in more than one place, or to terms that should have been excluded long ago.

In our last post on the logic behind Amazon CPC bidding, we showed how to calculate the right bid for a keyword. That math only works if each keyword has exactly one home. This post is about building that home.

What a Good Amazon Campaign Structure Actually Is

Think of the account in three layers. Each layer answers one question, and each has its own job in reporting.

Portfolios by category. Group campaigns into portfolios like Serums & Ampoules, Sunscreen, Cleansers and Toner Pads. Portfolios let you set budget caps per category and see category-level ACoS and TACoS at a glance. This matches how most K-beauty brands already think about their line-up and margins.

Campaigns by product. Inside each portfolio, build campaigns around a hero ASIN or a tight product family (for example, one serum and its size variations). Don't mix a $14 cleanser and a $38 serum in the same campaign. They have different margins, different conversion rates and different max CPCs, so they need different bids and budgets.

Funnel layers within each product. For each hero product, separate campaigns by intent: brand defense, core category terms, discovery, and competitor targeting. Budgets and target ACoS should differ by layer, because a shopper searching your brand name and a shopper browsing "korean skincare" are worth very different amounts.

Tie it together with a naming convention everyone follows, such as SP | Serum | Rice-Serum-30ml | Exact-Core. Ad type, category, product, match type and funnel layer. When names are consistent, you can filter, sort and report across the whole account in seconds, and anyone new on the team can read the account without a walkthrough.

The Four Funnel Layers for Each Hero Product

1. Brand defense (exact match, branded terms): Your brand name, product names and common misspellings. These convert at the highest rate in the account and are cheap to win, but competitors will take them if you don't. Keep this in its own campaign with its own budget so branded sales never mix with, and inflate, your non-brand performance.

2. Core performance (exact match, proven category terms): The 10–30 non-brand keywords you know convert, like "rice toner", "snail mucin serum" or "korean sunscreen spf 50". This is where most of your profitable volume should come from. Bids here follow the max CPC formula, and this campaign gets budget priority.

3. Discovery (auto, broad and phrase): Research campaigns that find new search terms. Bids are lower, the target ACoS is looser, and the budget is capped, usually 15–25% of the product's Sponsored Products spend. Its job isn't to be profitable on its own. Its job is to feed winning search terms into the exact-match campaign.

4. Competitor and category targeting (product targeting): Target specific competitor ASINs and relevant categories so your product shows up on their detail pages. Keep it separate from keyword campaigns, because product-page placements convert and cost differently from search placements.

Sponsored Brands and Sponsored Display follow the same logic in their own campaigns: Sponsored Brands on branded and core terms to own the top of the page, Sponsored Display for retargeting product-page viewers. As we covered in full-funnel Amazon advertising, DSP sits above all of this to create demand that these campaigns then capture.

How to Stop Double Bidding: The Harvest and Exclude Loop

The rule is simple: every search term should have exactly one campaign responsible for it. The way to enforce that is a loop that runs every week.

First, pull the search term report from your discovery campaigns. Any search term with enough orders at an acceptable ACoS (a common bar is two or more orders in 30 days, under target ACoS) gets harvested: added as an exact-match keyword in the core performance campaign, with a bid based on your max CPC.

Second, and this is the step most teams skip, add that same term as a negative exact in the auto, broad and phrase campaigns it came from. Now the discovery campaigns stop paying for it, and the exact campaign, with its controlled bid, owns it completely.

Third, add your brand terms as negative phrase in every non-brand campaign. This keeps branded traffic flowing only to the brand defense campaign, so your non-brand ACoS reflects real new-customer acquisition rather than people who were already looking for you.

Run this loop for a month and the overlap mostly disappears. Bids start meaning something, because a change in one campaign actually changes what you pay for that term.

Using Negatives Without Choking Growth

Sponsored Products gives you negative exact and negative phrase keywords, plus negative product targeting for specific ASINs or brands. Used well, they're the most direct way to cut waste. Used badly, they quietly kill the traffic you need.

1. Exclude what's clearly irrelevant: Terms that will never convert for your product. For a skincare brand that's often things like "dupe", "free sample", competitor brand names you don't want to target in that campaign, or the wrong skin concern ("acne" for an anti-aging cream). Use negative phrase for these so all variations are blocked.

2. Exclude what has failed with enough data: Use the same threshold from the bidding post. If a search term has had two to three times the clicks you'd need to expect a sale, and still has zero orders, add it as a negative exact. Don't negate a term after 5 clicks; that's noise, not a verdict.

3. Exclude what belongs elsewhere: This is the double-bidding fix above. Harvested winners and brand terms get negated in discovery campaigns, not because they're bad, but because another campaign owns them.

4. Exclude low-value placements in product targeting: If certain competitor ASINs keep taking clicks without converting, often much cheaper or much more established products, add them as negative product targets rather than lowering the whole campaign's bid.

Keep a simple log of what was negated and why. Six months from now, when a term starts trending, you'll want to know whether it was blocked because it failed or just because it moved to another campaign.

Getting the Most Out of Auto Campaigns

Auto campaigns aren't lazy. They're Amazon's own read of what your listing is relevant to, and they regularly find terms a human would never have guessed, especially for K-beauty products with Korean ingredient names. The mistake is running one auto campaign with one bid and leaving it alone.

Amazon splits auto targeting into four groups: close match (searches closely related to your product), loose match (loosely related searches), substitutes (shoppers viewing similar products) and complements (shoppers viewing products that go with yours). Each one behaves differently. For most skincare brands, close match converts best, substitutes is a useful way to get on competitor pages, and loose match and complements spend the most for the least return.

So bid each targeting group separately. A practical starting point: close match at about 100% of your discovery bid, substitutes at 70–80%, loose match and complements at 40–50%. Adjust after two to three weeks of data. Keep the auto campaign's budget capped, harvest winners every week, and negate them back out. Treated this way, the auto campaign becomes your cheapest keyword research tool, not a budget leak.

What to Do First

Don't rebuild the whole account in one weekend. Pick your top hero ASIN and rebuild just that product: one brand defense campaign, one exact core campaign, one capped discovery campaign with auto split by targeting group, and one product targeting campaign. Move the proven keywords into exact, add negatives to the discovery campaigns, and pause the old overlapping campaigns once the new ones are spending. Then repeat for the next product.

The K-beauty brands growing efficiently on Amazon in 2026 aren't running more campaigns. They're running fewer, clearer campaigns where every search term has one owner, and a weekly loop that keeps it that way.

At Veilup, we help K-beauty and skincare brands build and run Amazon advertising programs, from campaign architecture and negative keyword strategy to bidding, DSP and listing optimization. If your account has grown into a tangle of overlapping campaigns, the structure to fix it is already here.

Put this into practice

Connect your Amazon account structure, listing priorities and measurement in one scoped plan. Explore Veilup's Amazon PPC and listing management for the service scope and next steps.

Want this applied to your brand?

Veilup runs performance marketing, AI content and retention for K-beauty and DTC beauty brands. Tell us where you are, and we'll show you where to start.

Book a Growth Call
Your brand, rebuilt for the AI era.