Basket Incidence: Tracking How Retail Ads Change Habits
Basket incidence answers a direct question: in what share of eligible shopping baskets did the product, brand or category appear?
That makes it one of retail media’s most intuitive transaction outcomes. It is closer to product choice than an impression or click, and it can show whether more trips included the advertised item.
Campaign-period incidence tracks product choice during the measured window. It does not, by itself, show that a durable habit changed. Change the denominator, the store mix, the category definition or the availability conditions and the rate changes. Compare exposed shoppers with people who were never equally eligible and the campaign can claim credit for behavior it selected rather than caused.
Basket incidence is useful when the definition is stable, the campaign job is explicit and the measurement design can separate attribution from incrementality.
The basic definition
For a defined product scope and eligible transaction population:
Basket incidence = eligible baskets containing the product divided by all eligible baskets.
The numerator can represent a SKU, brand, subcategory or category. The denominator can represent all transactions, category transactions, shopper trips, a store group or another pre-defined eligible population.
A complete incidence definition names:
- product or category scope;
- transaction and shopper eligibility;
- store, channel and period;
- denominator;
- treatment of returns, canceled orders and substitutions;
- availability and promotion rules;
- whether repeat transactions from the same shopper are counted separately.
Google Analytics ecommerce events can carry product identifiers, categories, price and quantity through cart and purchase events.[3] That event structure supports incidence analysis, but incidence still depends on retailer taxonomy and transaction rules.
Four incidence views that answer different questions
Absolute incidence
Baskets containing the defined item divided by all eligible baskets. This is useful for trip penetration across a store, channel or audience.
Category-conditional incidence
Baskets containing the item divided by baskets containing its category. This is closer to choice within the category and can separate category growth from brand selection.
Joint and directional incidence
- Joint incidence: baskets containing A and B divided by all eligible baskets.
- A-to-B attachment: baskets containing A and B divided by baskets containing A.
- B-to-A attachment: baskets containing A and B divided by baskets containing B.
The denominator belongs in the metric name. A joint rate and a directional attachment rate answer different questions and should never share an unlabeled “cross-incidence” heading in a report.
Incremental incidence
The difference in incidence between a treatment and credible counterfactual, normally expressed in percentage points. A pre/post change alone is not necessarily incremental because demand, price, promotion, availability and trip mix can change at the same time.
Do not switch between these rates in one report without labeling the denominator. A brand can improve category-conditional incidence while absolute incidence falls if fewer shoppers enter the category. The commercial interpretation is different.
Incidence is one part of the sales equation
Transaction revenue can change because of several components:
- the number of eligible baskets;
- the share of baskets containing the product;
- units purchased when present;
- price and product mix.
Incidence isolates the presence component. That is useful because it distinguishes “more trips chose the product” from “existing buyers purchased more units” or “price increased.”
It is not the same as purchase frequency. Incidence is calculated across baskets. Frequency usually describes how often a shopper purchases across a period. A campaign can increase incidence among one-time triers without changing repeat frequency, or increase repeat purchases among existing buyers without changing the broader basket rate much.
It is also separate from basket size and average basket value. Basket size covers units and breadth per transaction; average basket value covers money per completed transaction.
Design the campaign around eligible baskets
An incidence campaign should begin with the baskets that could reasonably change.
A category-acquisition campaign might use all relevant trips in stores where the product is available. A brand-switching campaign may use category baskets that did not already contain the brand. A cross-sell campaign may use baskets containing a credible trigger category but not the promoted complement.
The eligibility rule should be defined before exposure. Otherwise, the report can quietly remove difficult baskets after the result is known.
Availability is part of eligibility. If the product was not ranged or was out of stock, the shopper could not add it. Promotion also needs explicit treatment because a price reduction can change incidence independently of media.
Use shopping missions when trip context materially changes the baseline. A quick top-up and a weekly stock-up create different category opportunities. Do not call a shift in mission mix an advertising effect.
A worked example
A protein-bar brand runs in-store and offsite media around post-workout occasions. The weak analysis compares incidence among exposed shoppers with all unexposed shoppers.
The stronger design defines eligible stores, times and baskets before launch. It confirms that the product is ranged and available, records concurrent promotions and creates a holdout among similarly eligible shoppers or locations.
The primary outcome might be absolute brand incidence in eligible baskets. Supporting views can include category incidence, category-conditional brand incidence, units per containing basket and new-to-brand purchase.
If category incidence rises but the brand’s conditional incidence does not, the campaign may have expanded the category without improving brand choice. If brand incidence rises while a close substitute falls, part of the result may be switching rather than basket expansion. Both findings can be commercially useful, but they answer different questions.
Measure incremental incidence
The IAB/MRC guidelines require clarity about outcome definitions, data sources, attribution and whether reporting covers paid, organic or both placement types.[1] An incidence report should document who was eligible, who was exposed, what counted as a basket and which product mapping was used.
Use three layers:
- Delivery: Was the media served to an eligible shopper, store, session or trip?
- Observed behavior: Did the product appear in the basket, and how did category, substitute and unit behavior change?
- Causal impact: How did treatment incidence compare with what would have happened without the campaign?
IAB and IAB Europe describe incrementality as a counterfactual question and warn that attribution and ROAS do not establish causality.[2] Randomized holdouts are preferred where feasible. Matched stores, matched shoppers or modeled approaches can be used when randomization is impractical, but the assumptions and uncertainty should be disclosed.
For the exposure-to-purchase data chain, see closed-loop measurement. Closed-loop attribution connects events. The control design determines whether the incidence change is likely incremental.
Report rates and counts together
A percentage without its base can hide instability. Report:
- containing-basket count;
- eligible-basket count;
- incidence rate;
- treatment-control percentage-point difference;
- confidence or uncertainty where available;
- availability and promotion conditions;
- breakdowns that were planned before launch.
Avoid excessive slicing. A store, hour and audience combination with very few baskets can create dramatic rates that will not repeat. Predefine minimum sample and privacy thresholds.
Also show the incidence distribution across stores or audience groups when the mean hides variation. A campaign can produce an overall gain while failing in locations with poor availability.
When incidence can support a habit-change claim
A campaign-period increase is evidence of product choice in that period. A habit claim needs a longitudinal design that tests whether behavior repeats and persists.
Predeclare a post-campaign outcome window and follow the same eligible shoppers, households or comparable cohorts across repeated purchase opportunities. Report repeat-purchase incidence, the share purchasing in multiple windows, time to next purchase and the persistence of the treatment-control difference after promotional pricing and campaign exposure end. The counterfactual must face the same seasonality, assortment, availability and category conditions.
Shopper-level repeat measures should distinguish trial from repetition. One purchase is trial; purchases across separate opportunities may support a repeat pattern. Even then, describe the evidence as sustained behavior unless the study was designed to establish habit formation. Disclose attrition, identity-match coverage and minimum privacy thresholds, and do not infer persistence from only the shoppers who returned and remained observable.
Common mistakes
Using all baskets when only some were eligible. Availability, ranging, geography and channel can change the true denominator.
Confusing percentage points and percent change. State both only when each is calculated correctly.
Calling post-minus-pre incremental. Seasonality, price, promotion and assortment can move incidence without media.
Ignoring category incidence. Brand incidence can rise because the category grew, not because brand choice improved.
Ignoring substitution. A gain may come from another SKU or brand rather than category expansion.
Mixing basket and shopper metrics. Incidence, frequency and penetration use different units.
Cherry-picking segments. Predefine breakdowns and report the total eligible population.
Buyer checklist
Before approving an incidence plan, confirm:
- What is in the numerator?
- What exactly is the denominator?
- Which baskets are eligible to change?
- Is the product ranged and available?
- Is the outcome absolute, category-conditional, cross or incremental incidence?
- How are price, promotion and substitutes handled?
- Are paid and organic exposure reported separately?
- What is the counterfactual?
- Will counts, rates and uncertainty be shown together?
- What decision follows each plausible result?
FAQ
What is basket incidence?
It is the share of defined eligible baskets that contain a specified SKU, brand or category.
Is basket incidence the same as conversion rate?
No. Conversion normally starts from a visit, session, click or exposure. Basket incidence starts from a transaction population.
Is higher incidence always good?
No. The gain may be unprofitable, promotion-driven, caused by substitution or concentrated in baskets that would have bought anyway.
How should incidence lift be expressed?
Show the treatment and comparison rates, the difference in percentage points and, if useful, the relative change. Always include basket counts.
Can incidence prove that advertising changed habits?
Only with a credible counterfactual and an outcome window appropriate to the habit claim. One campaign-period increase usually proves less than a sustained repeat pattern.
Bottom line
Basket incidence turns product choice into a clear transaction rate. Its strength comes from disciplined definitions, not from a simple formula alone. Fix the numerator and eligible denominator before launch, protect for price and availability, report counts with rates, and use a counterfactual before calling the change incremental.
Sources
[1] Interactive Advertising Bureau and Media Rating Council, *IAB/MRC Retail Media Measurement Guidelines*, 2024.
[2] IAB and IAB Europe, *Guidelines for Incremental Measurement in Commerce Media*, 2025.
[3] Google Analytics, *Measure ecommerce*.
Ready to see how this works in practice?
Footprints AI helps brands and retailers measure what matters. See our customer success stories or get in touch to discuss your retail media strategy.




