Average Basket Value: How Retail Media Should Read Value, Not Just Visits

A campaign can drive more store visits while producing smaller transactions. Another can reach fewer shoppers but shift them toward higher-value products. Visit counts alone cannot distinguish the two.

Average basket value adds a commercial lens: how much transaction revenue was recorded per eligible completed basket?

The metric is easy to calculate and easy to misuse. It can rise because shoppers bought more items, chose premium products, faced higher prices, used fewer discounts or simply came from a naturally high-spending audience. None of those explanations automatically proves that media created incremental value.

Use average basket value as a diagnostic outcome, then decompose the drivers and test the causal claim.

Define the metric and denominator

For a defined eligible transaction population:

Average basket value = transaction revenue divided by completed eligible transactions.

Depending on the system, the metric may be called average order value, average transaction value or average basket value. The name matters less than the rules behind it.

For example, Shopify defines average order value as gross sales minus discounts, divided by orders, while excluding post-order adjustments.[5] That is one valid specification, not a universal rule; every report still needs to state its own revenue and order treatment.

Document:

• gross or net revenue;
• treatment of tax, delivery fees and tips;
• discounts, coupons and loyalty rewards;
• refunds, cancellations and partial fulfillment;
• currency and market;
• online orders versus store transactions;
• which transactions are eligible;
• shopper, store and time scope.

Google Analytics ecommerce documentation calculates purchase value from item price and quantity and supports purchase and refund events.[3] Google Cloud analytics uses purchase-complete events for total revenue and average order value, and notes that revenue accuracy depends on populated revenue and quantity fields.[4]

If one channel includes delivery fees and another excludes them, the averages are not comparable. If refunds are ignored, high-return products can inflate reported value.

Decompose value before interpreting it

Average basket value is produced by several mechanisms.

At the arithmetic level, keep the decomposition non-overlapping:

Average basket value = items per basket x average realized revenue per item.

Realized revenue per item reflects SKU and pack mix, list prices and discounts. Trip mix is not another component inside an individual basket; it changes which transactions enter the population being averaged.

Quantity

Shoppers may buy more units or more distinct products. Report basket depth and breadth separately when each matters.

Product mix

The basket may shift toward premium tiers, larger packs, different categories or higher-margin products.

Price

List-price changes can raise value with no behavioral change.

Promotion

Discounts can reduce value per unit while increasing units, incidence or total revenue.

Trip mix

The campaign may reach more stock-up trips, high-value stores, loyal shoppers or planned occasions.

A useful report shows which component moved. “Average basket value increased” is an observation, not an explanation.

Mean, median and distribution

The arithmetic mean is sensitive to extreme transactions. A small number of large stock-up baskets can move the average while the typical transaction remains unchanged.

Report the mean because it connects total revenue and transaction count. Add the median, planned value bands or another distribution view when skew is material. Show the transaction count behind every segment.

Do not remove large baskets merely because they weaken the story. Define outlier rules before the campaign and apply them consistently to treatment and control.

A worked example

A premium skincare brand runs offsite and onsite media. Exposed shoppers show a higher average basket value than unexposed shoppers.

That comparison is not enough. Premium-skincare audiences may already spend more, shop in different stores and buy more categories. The campaign may also coincide with a price promotion or gift-with-purchase.

A stronger analysis asks:

• Did the brand enter more eligible baskets?
• Did buyers add a product or replace another skincare item?
• Did pack, tier or category mix change?
• Were prices and promotions balanced?
• Did total basket value rise because of skincare or unrelated categories?
• What happened in a comparable holdout?

The campaign might increase the target brand’s incidence while total basket value remains flat because shoppers substitute within a fixed budget. That can still be a successful share-shift campaign. Alternatively, total value may rise while target incidence stays flat because the audience had more large stock-up trips. That is weak evidence for the brand campaign.

The metric must be interpreted against the job.

Segment without cherry-picking

Average basket value can be cut by channel, store, audience, mission, new-to-brand status, product tier and many other dimensions. Every extra cut creates another chance to find an attractive result by luck.

Predefine the most decision-relevant segments. Require minimum transaction counts and privacy thresholds. Report the total eligible population before highlighting a subgroup.

Useful planned cuts can include shopping missions when trip context is defined before exposure. Other planned cuts can include:

• new versus existing brand buyers;
• promoted versus non-promoted transactions;
• store format and channel;
• mission or trip type;
• category-containing versus all eligible baskets;
• treatment intensity when assigned independently of behavior.

Avoid segmenting on outcomes that happen after exposure. For example, comparing recommendation clickers with non-clickers selects shoppers by response and cannot estimate campaign impact.

Measure incremental basket value

The IAB/MRC guidelines emphasize consistent definitions, transparent data sources and clear attribution approaches.[1] An average-basket-value report should disclose transaction eligibility, revenue treatment, currency, product scope and whether paid and organic exposure are combined.

Use three layers:

1. Delivery: Which eligible shoppers, stores or sessions received the campaign?
2. Observed outcome: What were transaction count, revenue and average value, and which components changed?
3. Incremental outcome: How did value compare with the credible no-campaign counterfactual?

IAB and IAB Europe state that attribution and ROAS do not prove causality and call for credible counterfactuals and bias control.[2] This is crucial for value-based targeting because high spenders are likely to have high values with or without media.

Average basket value is a ratio, so analyze it with conversion rather than in isolation:

Revenue per eligible session = conversion rate x average basket value.

Report eligible sessions or shoppers, conversion rate, completed transactions, total revenue and average basket value together. A higher average may reflect the loss of low-value transactions, a change in conversion or a change in revenue among completed baskets. Causal analysis must estimate both conversion and value components.

Randomized holdouts are preferred when feasible. If randomization is not possible, matching should include pre-period spend, frequency, store, channel, mission, category behavior, promotion and availability. Report uncertainty and sensitivity to the method.

For the event connection from exposure to transaction, see closed-loop measurement. Closed-loop attribution is necessary for observation. It is not sufficient for causality.

Connect average value to total and incremental revenue

Average value can rise while total revenue falls if transaction count falls enough. It can fall while total revenue rises if the campaign creates many additional smaller transactions.

Always read value with:

• eligible shopper or session count;
• transaction count and conversion;
• total revenue;
• average value;
• target-product revenue;
• margin where available;
• media cost;
• the incremental estimate.

Do not calculate incremental revenue by multiplying an observational value difference by all exposed transactions unless the value difference and transaction effect are causally identified. The transaction count itself may have changed because of the campaign.

Average basket value is also not ROAS. ROAS compares attributed or incremental revenue with media spend, while incremental sales estimates total causal revenue change. Basket value is one transaction-level component of those broader outcomes.

Price, promotion and inflation controls

A value metric is especially sensitive to price. If the same basket costs more, average value rises without additional units, breadth or shopper benefit.

Choose and document a reproducible constant-price method. One option is to revalue treatment and comparison baskets with a fixed SKU-level reference-price table from a declared base period, while reporting observed paid value separately. Define the fallback for new SKUs and exclude or flag products with no valid reference price rather than silently assigning one.

Separate:

• constant-price value where feasible;
• observed paid value;
• units and distinct products;
• promotion depth and redemption;
• premium or pack mix;
• category and store price changes.

For multi-market reporting, keep currencies separate or convert explicitly using a documented rate and date. Do not average monetary values across currencies without conversion.

Common mistakes

Calling basket size basket value. Quantity and money are different metrics.

Ignoring transaction count. A higher average can accompany lower total revenue.

Targeting high spenders and claiming lift. Selection creates high attributed value.

Ignoring price and promotion. Behavioral and price effects can move in opposite directions.

Using gross revenue inconsistently. Tax, fees, refunds and discounts must follow one rule.

Reporting only the mean. Extreme transactions can hide the typical result.

Mixing currencies. Compare or convert explicitly.

Treating attribution as incrementality. Use a counterfactual.

Assuming higher is always better. Profit, repeat behavior, returns and shopper value may tell a different story.

Buyer checklist

Before approving average basket value as a KPI, ask:

• What revenue components are included?
• Which completed transactions form the denominator?
• How are refunds, discounts and cancellations handled?
• Are channels and currencies comparable?
• Will mean, median and transaction counts be shown?
• How will quantity, mix, price and promotion be separated?
• Is total revenue reported beside the average?
• Is the campaign trying to add, trade up, switch or retain?
• What counterfactual supports the incremental claim?
• How will margin and media cost enter the final decision?

FAQ

Is average basket value the same as average order value?

Often, yes, but confirm the transaction, fee, tax, refund and channel rules. Different systems can use the labels differently.

Is higher average basket value always better?

No. It may reflect price increases, fewer low-value transactions, unprofitable promotions or a shift toward high-return products.

Should the metric use all baskets or category baskets?

Use the denominator that matches the campaign job and report it explicitly. Category baskets are useful for within-category value; all eligible baskets show the broader transaction effect.

How is average basket value different from basket size?

Basket size measures units or breadth. Basket value measures money per transaction.

Can average basket value prove retail media impact?

Not alone. It is an observed outcome. Incremental impact requires a credible comparison with what would have happened without the campaign.

Bottom line

Average basket value makes retail media commercially legible, but the average is only the start. Fix the transaction and revenue definitions, show the distribution, separate quantity, mix, price and promotion, read the metric with total revenue and margin, and use a counterfactual before calling the value 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*.

[4] Google Cloud, *View analytics*.

[5] Shopify Help Center, *Sales reports: Average order value*, accessed September 2026.

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.