Amazon PPC Cannibalization: Are Your Ads Stealing Organic Sales?

Priya's best campaign ran at a 6% ACoS. It was the first thing she showed anyone who asked how her kitchen-storage brand was doing. $4,100 a month in spend, $68,000 in attributed sales, every week the same beautiful number.

Then her ad account got suspended for three days over a billing card that expired. No ads at all. She braced for a sales crash.

Total sales dropped 4%.

That campaign was bidding on her own brand name and on two generic keywords where her listing already sat in organic position one. Almost every sale it "drove" would have happened anyway. The 6% ACoS was real. The value was mostly not.

That is Amazon PPC cannibalization: ad spend that buys sales you would have made organically for free. It is the most flattering line in most ad accounts and one of the least examined. This guide covers the two kinds of cannibalization, how to measure what your ads actually add, when bidding on keywords you already rank for is the right call, and how to set defensive bids instead of making blanket on/off decisions.

What Amazon PPC Cannibalization Actually Means

Amazon attributes a sale to an ad if the shopper clicked that ad and bought within the attribution window (7 days for Sponsored Products, 14 days for Sponsored Brands and Display, for sellers). Attribution answers one question: did an ad touch this sale? It does not answer the question you actually care about: would this sale have happened without the ad?

The gap between those two questions is cannibalization. The share of attributed sales that would have happened anyway is the cannibalized share. The rest is incremental: sales that exist only because you paid for the click.

There are two versions, and they need different fixes.

1. Paid vs organic cannibalization

Your listing ranks #1 organically for "bamboo drawer organizer." You also run an exact-match ad on that keyword, and it wins the top Sponsored slot above your own organic result. A shopper clicks the ad because it is first. You paid $1.40 for a click you would have received for free twelve pixels lower.

Not every one of those clicks is wasted. Some shoppers would have scrolled past your organic listing to a competitor's ad. But a meaningful share are pure transfers from your free column to your paid column.

2. Paid vs paid cannibalization

Three of your campaigns target the same search term: an auto campaign, a broad-match campaign, and an exact-match campaign. Amazon picks one of your ads to enter the auction. You are not bidding against yourself directly, but your reports split the term's performance across three places, your harvesting logic gets confused, and the campaign with the highest bid sets your cost. If you want the structural fix, our Amazon PPC campaign structure guide covers negative-exact isolation in detail.

This guide focuses on the first kind, because it is the one that hides behind good numbers.

Why Cannibalization Hides Behind a Great ACoS

The campaigns most likely to be cannibalizing are the ones that look the best. That is not a coincidence.

Branded and top-ranked keywords have the highest conversion rates in your account because the shopper already wanted your product. High conversion rate means low ACoS. Low ACoS means the campaign gets more budget. More budget means more of your organic demand gets rerouted through paid clicks.

The tell is in the relationship between ACoS and TACoS. If you scale a branded campaign and ACoS stays at 6% while TACoS climbs and total sales barely move, the ads are relabeling sales, not creating them. We cover the mechanics of that split in ACoS vs TACoS, but the short version is this: ACoS measures efficiency on attributed sales, TACoS measures cost against everything. Cannibalization inflates the first and quietly worsens the second.

How to Measure Incrementality on Amazon

You cannot measure cannibalization from a single campaign report. You need a comparison: what happens to total sales when the ad is on versus off. Here are three methods, from rough to rigorous.

Method 1: The TACoS slope check (15 minutes)

Pull 12 weeks of data by ASIN: ad spend, ad sales, total sales. Plot weekly ad spend against weekly total sales.

If a $1,000 increase in weekly spend comes with roughly $1,000 × (1 ÷ ACoS) in extra total sales, your ads are mostly incremental. If total sales rise by a fraction of that, the gap is your cannibalization rate.

Example: Priya's branded campaign ran at 6% ACoS, which implies $16.67 in sales per $1 spent. When she doubled weekly spend from $950 to $1,900, total weekly sales rose from $41,200 to $42,700. That is $1,500 in extra sales for $950 in extra spend, or $1.58 per dollar. Her incremental ROAS was 1.58, not 16.67. Incremental ACoS: roughly 63%.

This method is noisy because seasonality, price changes and stock levels move total sales too. Use it as a smoke detector, not a verdict.

Method 2: The holdout test (2-4 weeks)

This is the one that actually settles the question.

1. Pick the keyword set. Start with your branded terms and the 5-10 generic keywords where you hold organic position 1-3. Use Brand Analytics Search Query Performance to confirm your organic click share on each.
2. Record a baseline. Two weeks of total sales, sessions and conversion rate for the affected ASINs, plus your organic click share on each keyword from SQP.
3. Pause or drop to a floor bid. Turn off the ads on that keyword set for 14 days. Leave everything else untouched: price, coupons, other campaigns, inventory.
4. Compare total sales, not ad sales. If total sales on those ASINs fall by less than the attributed ad sales you removed, the difference was cannibalized.
5. Check who took the top slot. Search the terms yourself during the test. If a competitor's ad now sits above your organic listing and your SQP click share drops, some of that spend was defensive and worth keeping.

The formula:

If your paused ads had $20,000 in attributed sales over two weeks and total sales fell by $5,000, your incremental share is 25%. Three quarters of that campaign's attributed revenue was coming in anyway.

Method 3: AMC path analysis (for brands with volume)

If you have access to Amazon Marketing Cloud (Amazon's overview), you can query how many ad-attributed purchasers had already viewed or searched for your brand before the click. A high share of "already brand-aware" converters on a campaign is strong evidence of cannibalization. This needs meaningful volume and someone who can write SQL, but it avoids pausing anything.

Should You Bid on Your Own Brand Name on Amazon?

Usually yes, but for a different reason and at a very different price than most sellers pay.

The honest case for branded bidding is defense, not growth. When someone searches your brand name, competitors can and do target it. Brand Registry protects your listings, not your search results page. If your ad is not there, Sponsored Products and Sponsored Brands slots above your organic listing go to rivals whose products look similar and cost $2 less. Some of your shoppers will click them.

So the question is not "branded ads, yes or no." It is "what is the minimum spend that keeps competitors from taking my branded shoppers?"

Marco, who sells climbing chalk under his own brand, ran this test. With branded ads off, two competitors took the top Sponsored Products slots on his brand search within 48 hours. His total sales fell 11% over two weeks, versus the 38% of revenue his branded campaign had been "attributing." So branded ads were worth something, just a lot less than the dashboard claimed.

He turned them back on, but cut his branded bid from $1.10 to $0.45. At $0.45 he still won the top slot most of the time, because few competitors bid aggressively on someone else's brand. His branded spend fell 58% and total sales stayed within 1% of the pre-test level.

That is the pattern for most sellers: keep the branded ads, cut the branded bid until you start losing the top slot, then step back up one notch.

A few cases where branded bidding deserves more budget:

• Your brand name is a generic word. If your brand is "Summit" or "Harbor," branded searches are partly unbranded demand, and ads there behave more like generic keywords.
• Competitors are actively conquesting you. If your brand SERP is full of competitor ads, the defensive value rises. Check it weekly.
• Sponsored Brands on your brand SERP. A Sponsored Brands headline ad showcasing three products can move shoppers from your hero ASIN to higher-margin siblings. That is cross-sell, which is incremental even when the shopper was already yours.
• You are launching a new product under the brand. Branded traffic is the cheapest place to introduce a new ASIN to people who already trust you.

When Bidding on Keywords You Already Rank For Makes Sense

Generic keywords where you rank organically are more nuanced than branded ones, because organic rank on generic terms is contested and fragile. Here is a decision matrix.

Two reasons generic-keyword ads can be worth it even when you rank #1:

Owning two slots. If your ad sits at the top and your organic listing sits just below, you occupy more of the screen. Shoppers who scroll past the first result often still see you. On mobile, where the first screen holds two or three results, this can matter. The value is real but smaller than the attribution says. Test it.

Protecting organic rank. Amazon's organic ranking leans heavily on sales velocity for a keyword. Some of the sales your ad brings in reinforce the organic position you are worried about cannibalizing. If you cut ads on a keyword and a competitor steps into the top ad slot, your sales on that term can fall enough to cost you organic rank a few weeks later. That is the argument sellers use to justify everything; the holdout test tells you whether it applies to your keyword. Our Amazon SEO guide explains how ad-driven sales feed organic ranking.

How to Set Defensive Bids Instead of Switching Ads Off

The worst outcome of discovering cannibalization is overreacting. Sellers read about incrementality, pause every campaign with an ACoS under 10%, and then watch competitors take their brand SERP. The better approach is to bid by role.

Step 1: Tag each keyword by role

Split your targets into four roles:

• Defend: your brand name, your product names, your ASINs as targets.
• Hold: generic keywords where you rank organically in positions 1-3.
• Grow: generic keywords where you rank on page 1 below position 3, or not at all.
• Conquest: competitor brands and competitor ASINs.

Step 2: Give each role its own objective

• Defend: lowest bid that keeps the top slot. Ignore ACoS; it will be tiny either way. Measure impression share on top of search.
• Hold: bid to the level your holdout test justifies. If the test showed 25% incrementality, your effective ACoS is 4x the reported one, so set your target accordingly.
• Grow: standard target ACoS based on your break-even. This is where most of your budget should go.
• Conquest: allow a higher ACoS than Grow, because every sale is a customer taken from someone else.

Step 3: Adjust targets by incrementality, not reported ACoS

This is the move almost nobody makes. If a Hold keyword reports 8% ACoS and your holdout says 30% of its sales are incremental, the true incremental ACoS is 8% ÷ 0.30 = 26.7%. Compare that to your break-even, not the 8%.

Look at that table. The branded keyword with the 5% ACoS is roughly as expensive per incremental sale as the competitor conquesting keyword with a 34% ACoS. Most accounts allocate budget the opposite way.

Step 4: Separate defensive campaigns in your structure

Put Defend keywords in their own campaign with their own budget, and add those same keywords as negative exact in your Grow and auto campaigns. This keeps branded performance from inflating the numbers you use to judge your real growth campaigns, and it lets you set a hard daily cap on defense. Our bid strategy guide walks through setting role-based bids and placement modifiers.

Step 5: Watch the placement, not just the bid

For Defend and Hold keywords, the thing you are buying is a position, usually top of search. Use placement reports to check your top-of-search impression share. If you still hold the top slot at a lower bid, the lower bid wins. If you lose it, step back up. Our placements guide covers how to read and adjust top-of-search modifiers.

Five Cannibalization Mistakes That Cost Sellers the Most

1. Judging the account by blended ACoS. A big branded campaign drags blended ACoS down and makes a struggling account look healthy. Always report ACoS with and without branded terms.
2. Pausing branded ads without checking the SERP. You will usually lose some sales to competitor ads. Test, do not guess.
3. Running one holdout during a promotion. A coupon, a Lightning Deal or a stockout during the test makes the result meaningless. Hold everything else still.
4. Treating the test as permanent. Competitors change. Re-run holdouts on your largest keyword sets every quarter, and check your brand SERP weekly.
5. Giving branded wins credit for growth. If your agency or your own reports celebrate ad-attributed revenue growth, split it by role. Growth in branded ad sales is usually just growth in brand demand that would have come anyway.

Frequently Asked Questions

Does Amazon PPC cannibalize organic sales?

Partly. Ads on keywords where you already rank high organically, and on your brand name, often capture sales that would have come through the organic listing. Ads on keywords where you do not rank on page 1 are mostly incremental. The only reliable way to know the split for your account is a holdout test that compares total sales with ads on and off.

What is a good incrementality rate for Amazon ads?

There is no universal number, but as a rough guide: branded keywords often come in at 10-30% incremental, generic keywords where you rank #1 at 20-50%, and generic keywords where you do not rank organically at 80-100%. Use those ranges as priors, then replace them with your own test results.

Should I stop bidding on my brand name?

Usually no. Keep a presence to block competitors, but cut the bid to the lowest level that holds the top slot. Pausing branded ads entirely tends to hand shoppers to rivals conquesting your brand.

How long should an incrementality holdout run?

At least 14 days, so that both weekday patterns and the 7-day attribution window are covered. For low-volume ASINs, 21-28 days gives a cleaner signal.

Does cannibalization affect TACoS?

Yes. Cannibalized ad spend raises TACoS without raising total sales. That is why a rising TACoS alongside a flat or falling ACoS is one of the clearest warning signs.

The Bottom Line

Amazon PPC cannibalization is not a reason to cut ads. It is a reason to stop grading every keyword by the same number. A 5% ACoS on your brand name and a 30% ACoS on a competitor's brand can cost you the same per new customer, and only one of them looks like a problem on the dashboard.

Priya ran the holdout. She kept her branded ads at a floor bid, cut her two top-ranked generic keywords to defensive levels, and moved $2,600 a month into non-branded keywords where she sat on page 2. Her reported ACoS went up from 14% to 19%. Her total sales went up 17% in the next quarter, and TACoS fell from 9.8% to 8.1%.

Measure incrementality, bid by role, and let the dashboard look a little worse while the business gets better.