At a glance
A cross-channel gauge that estimates how much of your Criteo retargeting spend is going to shoppers who already know your brand and would very likely have returned on their own through direct or organic channels. Criteo retargets from a cookie pool of recent on-site visitors, and a meaningful slice of that pool is loyal, brand-aware buyers who were coming back anyway. Paying to retarget them is margin you could reclaim. This card joins Criteo’s retargeting clicks against your commerce-platform and analytics signal for repeat and direct-intent customers, and expresses the overlap as a percentage. It is not an argument to switch retargeting off, it is a tool to right-size the retargeting cycle so you are paying for incremental return, not re-buying customers you already had.
| What it counts | The estimated share of Criteo branded retargeting clicks attributable to shoppers who show strong direct or organic return intent, expressed as a percentage gauge. Built by joining Criteo retargeting click signal with commerce-platform repeat-customer and direct-session signal. |
| Cost basis | CPC-dominant. The reclaim opportunity is the CPC paid on clicks that would likely have arrived free. |
| Currency | The gauge value is a percentage. The implied reclaim is in advertiser-account currency when sized against retargeting spend. |
| Conversion attribution | Criteo claims these conversions under 30-day click + 7-day view; the cross-channel question is how many were incremental versus inevitable. This card estimates the inevitable share. |
| Attribution window | 30D click + 7D view default on the Criteo side. The cannibalisation estimate is a 30-day rolling read. |
| Bot / invalid traffic | Excluded from both sides of the join where filtering identifies it. |
| iOS 14.5+ ATT impact on the card | Moderate. ATT shrinks Criteo’s identifiable cookie pool, which can concentrate retargeting on the most loyal, most identifiable users, exactly the cohort most likely to return anyway, so post-ATT this percentage can drift upward. |
| Catalogue-feed dependency | Indirect. Feed health affects which products get retargeted but not the brand-loyalty overlap this card measures. |
| Time window | 30D (rolling 30 days). A monthly read smooths daily noise and matches the cadence of budget decisions. |
| Alert trigger | >30% of branded clicks would have been free organic. An illustrative threshold; the gauge flags when the estimated cannibalised share crosses the ceiling, signalling room to reclaim margin. |
| Roles | owner, marketing |
Calculation
Calculated automatically by joining your Criteo data with your connected commerce platform and analytics signal. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
A UK speciality-coffee DTC retailer with a strong repeat-purchase base runs Criteo retargeting alongside a healthy direct and organic channel. The gauge estimates how much of branded retargeting click volume overlaps with high direct-intent customers. Account currency GBP. Window is the rolling 30 days to 20 Jun 26.| Cohort | Retargeting clicks | Conversions | Notes |
|---|---|---|---|
| First-time / low direct-intent | 12,400 | 540 | Genuinely incremental retargeting value |
| Repeat buyers with strong direct/organic habit | 7,600 | 690 | Likely to have returned without the ad |
| Total branded retargeting | 20,000 | 1,230 | Gauge reads ~38% cannibalised |
- Roughly 38% of branded retargeting clicks fall in the high-direct-intent cohort. These are loyal buyers who visit directly, search the brand name, and reorder on their own rhythm. Paying a CPC to retarget them is largely re-buying a return that was coming anyway. The gauge crossing the illustrative 30% line is the flag to act.
- The loyal cohort converts harder, which is misleading. Notice the repeat cohort converts at a much higher rate (690 on 7,600 clicks) than the first-timers. That high conversion rate is exactly why naive ROAS makes retargeting these users look brilliant, the platform takes credit for a sale that was always going to happen. Incrementality, not raw ROAS, is the right lens here.
- The reclaim is margin, not a kill switch. The action is not to switch retargeting off. It is to cap frequency on the loyal cohort, exclude very recent purchasers, and shift the saved budget toward the genuinely incremental first-time and low-intent pool where the ad changes the outcome.
- Post-ATT this number tends to rise. As Criteo’s identifiable pool shrinks, retargeting concentrates on the most cookie-stable users, who skew loyal. So a rising gauge can partly reflect identity attrition rather than a new strategy fault. Read it alongside the tracking-decay card.
- The honest test is a holdout. The cleanest way to confirm cannibalisation is an incrementality holdout on the loyal cohort: suppress retargeting for a matched group and compare their return rate to the targeted group. If they return at nearly the same rate, the spend on that cohort was cannibalised.
- Gauge high + strong direct/organic channel = real cannibalisation, cap frequency on loyalists.
- Gauge high + weak organic channel = suspect, the overlap may be over-estimated, validate with a holdout.
- Gauge rising post-ATT = partly identity attrition concentrating on loyal users, not only strategy.
- Gauge low = retargeting is mostly reaching genuinely incremental users, leave it alone.
- Holdout return rate near targeted rate = confirmed cannibalisation on that cohort.
Sibling cards merchants should reference together
| Card | Why it matters next to Branded Paid Clicks Cannibalising Organic | What the combination tells you |
|---|---|---|
| ROAS | The headline efficiency reading this card cautions against trusting alone. | A high ROAS with a high cannibalisation gauge means the ROAS is inflated by inevitable conversions. |
| Conversions by Campaign | Shows where the loyal-cohort conversions concentrate. | Helps target frequency caps and recent-purchaser exclusions to the right campaigns. |
| Conversion Rate by Campaign | High conversion rates flag the loyal cohort. | Unusually high conversion rate on a retargeting line item is a cannibalisation tell, not always a success. |
| Total Revenue | Criteo-attributed revenue. | The slice attributable to cannibalised clicks is revenue Criteo claims but did not create. |
| Total Spend | Sizes the reclaim opportunity. | The cannibalised percentage applied to branded retargeting spend estimates the recoverable margin. |
| Criteo One Tag / First-Party Audience Decay | The identity-attrition context. | A rising gauge alongside tag decay means retargeting is concentrating on the few stable loyal cookies. |
Reconciling against Criteo
Where to look in Criteo’s own dashboard:Criteo Management Centre → Reporting → Performance Report for branded retargeting clicks and conversions, alongside your analytics direct and organic channel reports (for example Google Analytics → Acquisition → Traffic acquisition) and your commerce platform’s repeat-customer view.Criteo cannot show this card on its own. Criteo sees the clicks and the conversions it served; it has no view of how those same shoppers behave in your direct and organic channels, which is the other half of the cannibalisation question. That cross-channel join, Criteo retargeting against commerce repeat-customer and analytics direct-intent signal, is what Vortex IQ assembles. To sanity-check the gauge manually, compare Criteo’s branded retargeting conversion volume against your analytics direct and branded-organic conversion volume for the same period; a large overlap of the same customers is the cannibalisation this card estimates. Why our number is an estimate, not an exact figure:
| Reason | Direction | Why |
|---|---|---|
| Incrementality is inferred | Either direction | True cannibalisation can only be proven with a holdout; this gauge is a modelled estimate from observable direct and organic return signal, so treat it as directional. |
| Attribution overlap | Ours conservative | Where Criteo and organic both claim the same conversion, the cross-channel join apportions rather than double-counts, which can read lower than a naive overlap. |
| ATT identity attrition | Pushes ours up | A shrinking identifiable pool concentrates retargeting on loyal users, raising the estimate for reasons of measurement as much as behaviour. |
| Brand strength | Varies by store | Stores with strong direct and organic demand will legitimately show higher cannibalisation than low-awareness brands. |
| Card | Expected relationship | What causes legitimate divergence |
|---|---|---|
google_analytics.ga_direct_sessions / google_analytics.ga_organic_sessions | Strong direct and organic return signal is the basis for the cannibalisation estimate; the more of it, the higher the likely overlap with retargeting. | Analytics attribution and Criteo attribution use different models, so the overlap is estimated rather than summed. |
shopify.repeat_customer_rate / bigcommerce.repeat_customer_rate | A high repeat-customer rate raises the prior that retargeting is reaching buyers who would return anyway. | A high repeat rate driven by subscription or replenishment behaviour cannibalises differently from one driven by discretionary reorders. |