What the Wall Street Journal reported about AI and order value
The Wall Street Journal reported that Albertsons said online shoppers using its AI-powered shopping assistants had higher average order value (WSJ report). This is a company-reported association: it does not, by itself, show that using the assistant caused shoppers to spend more.
A comparison of average order value (AOV) describes the average dollar amount of orders in the groups being compared. As analysis, not a claim about the WSJ report’s methodology, there are several possible reasons one group’s orders might be larger. For example, shoppers who choose to use an assistant could differ from other shoppers in their engagement, shopping needs, or purchase patterns. The available source does not establish whether such factors were controlled.
The reported observation is a potentially useful signal, but the source available to this article does not establish the size or statistical reliability of the difference, how it was measured, or whether it resulted in incremental sales, profit, or return on investment. It also does not establish whether the observation would apply to other grocers. These are open questions, not evidence that the reported association is meaningless.
Why a bigger basket is not yet an economic win
As analysis, AOV is a useful early indicator, but it is not a complete measure of retail economics. A larger order could have different margins, discounts, and fulfillment costs from a smaller one. It could increase sales without increasing incremental profit if, for example, it relies on costly incentives or adds operational expense. Conversely, a modest increase could be valuable if it brings profitable, genuinely incremental items into an order without adding much service cost.
A retailer evaluating an assistant could connect shopper metrics to contribution economics. Recommended measures include gross margin after discounts and promotions, and the incremental costs associated with the feature. Those costs might include technology, customer support, merchandising work, or operational friction. The reported AOV association does not establish these outcomes and should not be presented as an ROI estimate.
There is also a measurement consideration in the word “average.” A small number of large orders can move an arithmetic average even if the typical shopper changes little. As a proposed analysis, retailers could examine medians and the distribution of order sizes alongside the mean, distinguish first-time use from sustained behavior, and track order frequency and total spend over a defined period. If shoppers consolidate purchases into fewer, larger orders, order value per transaction could rise without a corresponding increase in their overall spending or the retailer’s sales.
What a stronger test could look like
As a recommendation, a randomized experiment could help estimate whether an assistant causes a change. Eligible shoppers could be assigned to an experience with the assistant or a comparable control experience, with assignment and exposure documented before outcomes are measured. The analysis could define its primary outcome and time horizon in advance and include all assigned shoppers, rather than only people who voluntarily engaged deeply with the feature. This approach can help address selection bias.
A test could report the absolute and relative difference in AOV, the number of shoppers and orders, and uncertainty around the estimate. It could specify whether the figure is per order or per customer and how cancellations, returns, substitutions, discounts, and taxes are treated. Segmenting results by customer type, shopping mission, or product category may help explain where a tool works, but those comparisons should be interpreted cautiously when sample sizes are small or many comparisons are made.
A recommended economic scorecard would extend beyond basket size. Possible measures include incremental gross margin, repeat purchase and retention, conversion, items per order, fulfillment cost, discount dependence, returns or cancellations, and customer satisfaction. These measures could reveal trade-offs—for example, if an assistant increases conversion but lowers margin, or if an initial basket increase fades over time. Follow-up across multiple shopping cycles could help assess whether behavior persists.
The limits of the reported observation
The WSJ report supports the narrow claim that Albertsons said online shoppers using its AI-powered shopping assistants had higher AOV. The sources available to this article do not establish the measurement details or whether the association reflects a causal effect. A broad market benchmark would not, on its own, answer those questions or validate a retailer-specific result, so no such benchmark is used here.
Nor should an observation at one company automatically be generalized to grocery retail as a whole. As analysis, grocery shopping tasks such as finding a substitute, planning a meal, or replenishing staples could have different effects on basket size. The available source does not establish which, if any, of these tasks accounts for the reported association.
A signal to investigate, not proof of impact
The reported higher AOV among assistant users is worth investigating, but it does not prove that the assistant caused larger baskets or generated incremental sales, profit, or ROI. Retailers and investors would need a clearly described comparison and evidence connecting shopper behavior to incremental contribution to assess the commercial impact.
Further evidence could include a quantified result, a defined measurement period, and a description of the comparison method and shopper groups. A controlled test and measures of margin or repeat behavior could help establish whether the feature creates value rather than merely correlating with larger baskets. Until such evidence is available, the reported observation should be treated as an encouraging but incomplete data point.
Source
- The Wall Street Journal, “Albertsons Says Its Shoppers Are Buying More Thanks to AI”: https://www.wsj.com/cio-journal/albertsons-says-its-shoppers-are-buying-morethanks-to-ai-81793975
the wall street journal reported about ai
Grocery baskets with fresh produce and packaged goods arranged in
Grocery baskets with fresh produce and packaged goods arranged in
Grocery baskets with fresh produce and packaged goods arranged in