How To Spot Branded Traffic Cannibalization in SQP

A 6% ACoS on your own brand name is not proof of efficiency. It is often proof that you are buying clicks you would have won for free.
‍
That gap between how good branded ads look and how little they actually add is where branded traffic cannibalization in SQP hides.

The fix is not another opinion about brand bidding. It is a data check you can run inside Search Query Performance (SQP), the report already sitting in your Brand Analytics dashboard.
‍

This guide is for Amazon sellers, brand owners, in-house pay-per-click (PPC) managers, and agency analysts who suspect their branded spend is padding the numbers.

You will get the exact metrics to read, a decision rule with thresholds, a worked example, a cheap incrementality test, and a weekly checklist.
‍

Here is the short version up front. When your organic listing already ranks 1 to 3 for your brand name and your paid purchase share is very high, your ads are probably moving organic sales into the paid column, not creating new ones.
‍

What is branded traffic cannibalization in SQP?

Branded traffic cannibalization is when your Sponsored Products or Sponsored Brands ads capture sales on your own brand terms that your organic listing would have won anyway.

The ad gets the credit. The sale was never at risk.

SQP is where this shows up, because it reports your share of impressions, clicks, cart adds, and purchases for each search term against the rest of the market. You find it in Seller Central under Brands, then Brand Analytics, then Search Query Performance. Amazon Brand Registry enrollment is required.

The trap is the math behind ACoS. A shopper who types your exact brand name already decided to buy. They convert fast and cheap, so Advertising Cost of Sale looks excellent.

But a low ACoS measures correlation, not cause. It cannot tell you whether the ad earned the sale or just stood in front of a buyer already walking to the register.

That distinction has a name: incrementality. An incremental sale would not have happened without the ad.

A cannibalized sale would have happened regardless. SQP gives you the raw material to tell them apart.
‍

Which SQP metrics expose branded cannibalization?

Four share metrics matter, and you read them as overlap signals, not vanity stats. SQP breaks the funnel into impression share, click share, cart add share, and purchase share, each shown as your slice of the total market for that exact query.
‍

• Impression share. Your percentage of total market impressions for the brand term. On your own brand name this should already be high from organic placement.
  ‍
• Click share. Your percentage of total clicks. If click share sits far above impression share, your main image and title are doing the work, not the ad.
  ‍
• Cart add share. Your slice of add-to-carts. Note Amazon only counts a cart add when the shopper adds the exact product they clicked.
  ‍
• Purchase share. Your percentage of total purchases on that query. This is the headline. The attribution window is 24 hours.
  ‍

The tell is simple. On a branded query where you already own organic rank 1 to 3, a purchase share near total market means almost every buyer was yours before the ad ran. Layer in New-to-Brand (NTB) data from your Sponsored campaigns.

If NTB share on the brand term is low, the ad is mostly serving existing brand shoppers, which is the signature of cannibalization.

Compare that to a non-branded query, where ads genuinely expand reach and NTB share runs higher. The contrast between the two is your fastest read.
‍

The decision rule: rank plus purchase share

Stop guessing and use a threshold. The table below is an Amplivus diagnostic rule of thumb, not an official Amazon benchmark.

Pair your organic rank for the brand term with your branded purchase share, then read the verdict off it.
‍

The logic is the relationship between rank and share. The higher you already rank organically and the larger your purchase share, the smaller the pool of sales an ad could possibly add.

When you own the top slot and nearly all purchases, the ad has almost nothing left to win that you would not win for free.

Treat the table as a trigger, not a final answer. It tells you where to run the real test, which comes next. A high reading is a reason to investigate, not an automatic pause.
‍

A worked example (anonymized)

Numbers make this concrete. The figures below are a hypothetical, illustrative SQP row for a mid-size supplement brand bidding on its exact brand name, not a specific client account.
‍

• Brand query weekly search volume: about 4,200
  ‍
• Organic rank for the brand term: 1
  ‍
• Brand impression share: 88%
  ‍
• Brand click share: 84%
  ‍
• Brand purchase share: 93%
  ‍
• Branded campaign ACoS: 5.8%
  ‍
• NTB share on the branded campaign: 7%

On the surface, a 5.8% ACoS looks like a win nobody should touch. Read it through the rule and the picture flips. Organic rank is 1, purchase share is 93%, and only 7% of branded orders are new to the brand.

That means roughly 93% of buyers already belonged to the brand, and the listing already sat first organically.

The likely reality: most of those paid sales would have landed organically.

The ad is renting a sale you already owned. This row lands squarely in the top line of the table, so the next move is a controlled test, not an immediate cut.
‍

How do you test branded ad incrementality without AMC?

You do not need Amazon Marketing Cloud (AMC) to test this. The cheapest reliable method is a bid-down holdout, and any seller can run it in a week.

Run it like this:
‍

1. Record a baseline week: total branded units, total branded revenue, branded ad sales, and organic sales for the brand term.
   ‍
2. Cut your branded campaign bids by 50% to 80%, or pause the campaign entirely, for a full week.
   ‍
3. Hold every other variable steady. No price changes, no new creative, no major stock-outs.
   ‍
4. Measure total brand-term sales again, paid plus organic combined.
   ‍
5. Compare total sales, not ad sales. If total sales barely move while ad spend drops, the spend was non-incremental. If total sales fall meaningfully, the ads were pulling real demand.

The reason this works is that organic backfills cannibalized clicks. When you pull the ad, shoppers who would have clicked the sponsored slot simply click the organic listing instead, and the sale still closes.

This is not a fringe theory. A landmark field experiment in paid search, run at eBay by economists Blake, Nosko, and Tadelis and published in Econometrica, found brand-keyword ads produced no measurable short-term lift because those buyers were already arriving through organic search.

In one test, 99.5% of brand-search traffic reached the site even with the ads switched off. The mechanism on Amazon is the same. Your brand name is your highest-intent query, which is exactly where cannibalization risk peaks.

One caution. Run the test for at least a week, ideally two, so a slow day does not skew the read. And watch whether a competitor jumps into your top slot the moment you pull back, which changes the calculation.
‍

When is bidding on your own brand still worth it?

Cannibalization is common, but "always cut branded ads" is the wrong lesson. Brand bidding earns its keep in specific, provable situations.
‍

• Active conquesting. A competitor is bidding on your brand name and showing above your organic listing. Pull your ad and your bid-down test shows real sales loss. That is incremental defense, and Bing Ads studies on brand-term bidding show this is where brand ads add the most clicks.
  ‍
• New product launches. A new ASIN has no organic rank yet for the brand term. Ads carry the demand until organic catches up.
  ‍
• Variation steering. You want brand searchers landing on a specific hero variation or bundle, not whichever child Amazon surfaces organically.
  ‍
• Top-of-search real estate. You want to own both the sponsored and organic slots to push competitors below the fold during a launch or promotion.
  ‍
• Thin organic rank. If you sit at rank 4 to 10 for your own brand, the ad is buying visibility you do not yet own organically.

The test in every case is the same holdout. If total brand-term sales drop when you pull the ad, the spend is incremental and worth it. If they hold steady, you found waste.

Defense is a tactic with a measurable trigger, not a permanent line item.
‍

Common mistakes when reading SQP for cannibalization

The report is powerful, and it is easy to misread. A few traps to avoid.
‍

• Trusting ACoS alone. A low branded ACoS is the symptom that masks the problem, not evidence of health. Always pair it with purchase share and an incrementality test.
  ‍
• Ignoring the 24-hour window. SQP attributes impressions, clicks, cart adds, and purchases within a 24-hour period. Longer consideration cycles will not show cleanly, so read trends week over week.
  ‍
• Forgetting the denominator. Purchase share is your slice of the total market for that query. A new competitor entering the term can move your share without any change on your end.
  ‍
• The parent and child cart-add quirk. Amazon only counts a cart add when the shopper adds the exact product clicked. Variation listings can understate cart add share, so do not over-read that one column.
  ‍
• Cutting without a holdout. Never pause a campaign on the table reading alone. The threshold flags risk. The bid-down test confirms it.
  ‍
• Reading one week in isolation. Promotions, competitor launches, and seasonality all move shares. Look at three to four weeks before you commit to a verdict.
  ‍

A weekly SQP cannibalization checklist

Make this a habit, not a one-time audit. A short weekly routine keeps branded waste from creeping back.
‍

1. Export the latest SQP week from Brand Analytics in Seller Central.
   ‍
2. Filter to your branded queries and note organic rank for each.
   ‍
3. Record impression share, click share, and purchase share per term.
   ‍
4. Pull NTB share from the matching Sponsored campaigns.
   ‍
5. Run each term through the rank plus purchase share table.
   ‍
6. Queue any "likely cannibalization" term for a bid-down holdout test.
   ‍
7. Log results and compare week over week to catch competitor moves early.

Ten minutes a week beats one panicked audit a quarter.

Over time you build a clean record of which brand terms are incremental and which are passengers, which is exactly the evidence you need to defend a budget decision.

Branded cannibalization is not a reason to fear your own ads. It is a reason to measure them.

The sellers who win the long game treat the brand name as a hypothesis to test, not a slot to defend on reflex. Read the shares, run the holdout, keep what proves incremental, and redirect the rest toward queries where ads actually grow the brand.

If you would rather have it diagnosed for you, the Amplivus team runs this exact SQP and incrementality analysis as part of a branded campaign audit.