When Your Groceries Become a Gamble: How AI Pricing on Instacart Could Be Costing You More
Imagine ordering the same dozen eggs as your friend — from the same store, at the same time — and discovering you’re being charged more. That’s the startling reality revealed by a new investigation into Instacart’s AI‑powered pricing practices.
📉 What the Study Found
An independent collaborative study by Consumer Reports, Groundwork Collaborative, and More Perfect Union — dubbed “Same Cart, Different Price” — analysed live shopping data from 437 Instacart users across four U.S. cities. The results were eye‑opening. (Consumer Reports)
- Nearly 75% of the grocery items tested appeared with multiple different prices for the same product. Some had as many as five distinct price points. (Groundwork Collaborative)
- For everyday staples like cereal, peanut butter, and eggs, price differences between shoppers ranged on average 13%, with some individual items marked up to 23%. (Consumer Reports)
- For a typical household of four, these price variations could translate into as much as US $1,200 extra per year in grocery bills. (Consumer Reports)
What’s more concerning: these discrepancies occurred even when shoppers selected the exact same items from the same store location at the same time. (sfist.com)
🧠 How Instacart’s AI Pricing Works
The culprit behind this pricing roulette is an AI-driven pricing tool called Eversight — a software which Instacart acquired in 2022. The tool runs what the company calls “pricing experiments”: subsets of retail partners are enlisted to test varying prices across different user cohorts, all without notifying customers about who’s in the experiment or what criteria determine the pricing differences. (Groundwork Collaborative)
In theory, such experiments are meant to help retailers discover how sensitive shoppers are to price changes and optimise sales. But in practice, many customers end up paying more — sometimes significantly — simply because they were randomly placed into a pricier cohort. (Truthout)
Interestingly, not all items were affected. For example, staples like certain pasta or condiments in the safety‑net test didn’t show price variation, suggesting the experiments target only selected categories or retailers. (Consumer Reports)
The companies behind the report argue this practice constitutes “hidden algorithmic pricing,” eroding pricing transparency and fairness for consumers — especially when there’s no clear disclosure at checkout. (Consumer Reports)
⚠️ What This Means for Everyday Shoppers
- Erosion of trust and fairness. The idea that two shoppers can pay different prices for identical items undermines the expectation of fairness and transparency in retail — an expectation consumers usually associate with physical stores.
- Budgeting becomes harder. For households keeping a tight watch on monthly groceries, unexpected variances of 13‑23% per item can make budgeting unpredictable.
- Consumer protection concerns. Without explicit disclosure that prices might vary depending on invisible factors like your user cohort, many argue this crosses into deceptive or unfair pricing.
- Precedent for wider algorithmic pricing. Grocery is just the start — as AI pricing becomes more common, other everyday purchases could be similarly affected if regulation or disclosure doesn’t catch up.
✅ What Instacart Says — and the Pushback
In response to the findings, Instacart says only a small number of retail partners (around 10) use its AI pricing experiments, and the price differences seen by individual shoppers are “negligible.” The company also asserts that the retail partners themselves set the item prices. (Tom’s Guide)
But researchers contest this explanation: some major stores like Target — which supposedly don’t partner with Instacart — still displayed varying prices in the tests. That undermines the claim that only certain partners engage in these experiments. (Consumer Reports)
Experts and consumer advocates are calling for clearer disclosure practices — arguing that when prices are algorithmically tuned behind the scenes, shoppers deserve transparency: no surprises, no hidden “AI tax.” (Groundwork Collaborative)
🔑 Glossary
| Term | Meaning |
|---|---|
| Algorithmic pricing / Dynamic pricing | The practice of using algorithms (often powered by AI) to set or adjust product prices in real time, based on factors like demand, inventory, or customer segmentation. |
| Pricing experiment (A/B pricing) | A controlled test run by retailers or platforms where different groups of customers are shown different prices for the same item — typically used to assess willingness to pay or price sensitivity. |
| Eversight | The AI-driven pricing software acquired by Instacart in 2022, used to enable algorithmic pricing experiments across retail partners. |
| Cohort | A subgroup of users chosen (often randomly) to receive a specific variant in an experiment — e.g., a higher or lower price — for comparison. |
🧭 Final Thoughts
The shift toward AI-driven pricing in sectors like grocery delivery blurs a once-clear boundary: the expectation that equal products carry equal prices. While retail price‑testing isn’t new, embedding it behind opaque algorithms and hiding it from consumers raises ethical and regulatory questions. As AI continues to shape commerce, the call for transparency — and perhaps regulation — grows ever more important.
For now, if you use Instacart (or similar services), it may be worth double-checking your cart — or even comparing prices elsewhere. Because sometimes, the “same” item isn’t quite the same.
Source: Tech in Asia article “AI pricing on Instacart leads to cost variations for users: study” (techinasia.com)
news1
