Why Instacart Backtracked on an AI-Pricing Experiment

0:00Hello, I'm Stephen Carroll. I'm in Brussels where many of Europe's biggest decisions get made. And I'm Caroline Hepker in London with the hosts of the Bloomberg Daybreak Europe podcast. We're up early every weekday keeping an eye on what's happening across Europe and around the world. We do it early so the news is fresh, not recycled and so you know what actually matters as the day gets going. From Brussels, I'm following the politics, policy and the people shaping the European Union right now. And from London, I'm looking at what all that means for markets, money and the wider economy. We've got reporters across Europe and around the globe feeding in as stories break So whether it's geopolitics, energy, tech or markets you're hearing it while it happens It's smart, calm and to the point And it fits into your morning You can find new episodes of the Bloomberg Daybreak Europe podcast by 7am in Dublin or 8am in Brussels, Berlin and Paris On Apple, Spotify, YouTube or wherever you get your podcasts

1:02Bloomberg Audio Studios. Podcasts. Radio. News. One weekday this fall, 39 shoppers fired up the grocery delivery service, Instacart, and filled their virtual carts. They all put the same 20 items in there from the same Seattle Safeway. Wheat Thins, Heinz Ketchup, Cheerios, Skippy Peanut Butter. But when it came time to check out, the total price of their baskets was different. As much as$9.59 different. One shopper's$2.99 jar of peanut butter was another shopper's$3.69 jar of peanut butter. Two identical boxes of wheat thins could vary by as much as$0.90. Now, a few dollars here or there might not sound like a lot.

1:55But over a year of buying groceries, it can add up. Groceries are a major line item for most families' budgets. The other thing that's a problem for many families is when pricing is so variable, your grocery budget starts to feel really unpredictable. That's Lindsay Owens, the executive director of the Groundwork Collaborative, a progressive economic policy think tank. This month, Groundwork released an investigation with the organization Consumer Reports in partnership with the news nonprofit More Perfect Union that found that Instacart has been using an AI tool to run algorithmic price experiments on shoppers around the country, often without their knowledge.

2:38At a time when inflation has driven grocery costs higher, the revelation that Instacart was charging some consumers more for the same goods struck a chord. A$1 to$2 markup just from shopping online versus in person just sounds crazy to me when you're already doing the tip and having this service fee here. Obviously, this needs legislation, so let your representatives know that this matters to you. But until we have that law, avoid in-store apps. And it did get attention from lawmakers and regulators. Instacart, its stock is falling this morning after Reuters reported that the FTC sent the grocery delivery company a civil investigation demand.

3:19Instacart said that it does not use any personal, demographic, or user-level behavior data to set prices, and that prices weren't changing based on real-time supply and demand. Instead, they said that the price discrepancies were the result of completely randomized A-B testing, the kind of tests retailers often do to figure out how sensitive shoppers are to the price of certain items. Still, the company said, the report raised concerns. Quote, At a time when families are working hard to stretch their grocery budgets, customers should never have to question the prices they see on Instacart, an Instacart spokesperson told us.

3:58And this week, the company made a big 180. Instacart is ending a program where customers saw different prices for the same product ordered at the same time from the same store. Oftentimes, you know, companies get criticized for things all the time. It doesn't necessarily change their strategy. That's Leah Nylan, who covers antitrust for Bloomberg. But Instacart was pretty quick to sort of pull back on this policy, given the significant outcry that there was in response. As more online retailers lean on AI to influence their pricing strategy, the Groundwork Collaborative's Lindsay says Instacart's saga raises new questions.

4:36How far can you take a pricing practice before you touch the stove?

4:44I'm Sarah Holder, and this is The Big Take from Bloomberg News. Today on the show, what happens when an algorithm determines your grocery bill? How volunteer shoppers revealed Instacart's invisible pricing practices? And what came after the backlash?

5:10In 2022, Instacart acquired a company called Eversight that offers AI-powered pricing software. And what this did was it helped retailers, primarily grocery stores like Kroger, Albertsons, Safeway, things like that, set prices for like sort of everyday grocery items. Leah Nylan, an antitrust reporter for Bloomberg, says the tool has been controversial. They could increase the price if they thought, for example, you were somebody who really needed an item or were the sort of person who might not see a price increase and walk away. So they could sort of jack up the prices for certain people or lower them for others.

5:52The sort of problematic thing about this tool is, as a consumer, you have no idea that Instacart or the grocery store is even using it. Earlier this year, researchers and analysts at the Groundwork Collaborative and Consumer Reports decided to look closer at the grocery delivery company's pricing strategy and how it could impact consumers. We were interested in this kind of concept of surveillance pricing, this idea that companies are collecting data on us not just to get better at advertising to us, but also maybe to get better at overcharging us. Lindsay Owens again, the executive director of the Groundwork Collaborative.

6:30We knew a little bit about what Instacart was up to because they have been reasonably transparent with one population, their investors. So they have talked about their pricing strategies and their earnings calls. They have been transparent about the types of technologies they were buying and perfecting in their patent applications. And they have also talked a little bit about this on their website and some of their marketing materials. When they purchased Eversight in 2022, one of the reasons they did so is because Eversight offered better technology for running the types of experiments we found.

7:06So we knew that some sort of pricing experiments were going on with Instacart, but we wanted to understand, was it just a few pennies here or was it something that could really be contributing to the high cost of groceries that so many families in America are frustrated about right now. So we basically turned Instacart's pricing experiments right back around on them. Consumer Reports went to their network of members and asked for volunteers to participate in an Instacart shopping experiment. More than 400 people ended up participating. We said, OK, all of the consumers in the study, you're going to all pick out the exact same 18 to 20 grocery items, wheat thins and Quaker oats and deli lunch meat.

7:52You're going to put them in your cart and then you're going to take screenshots of the prices you're being offered for each of the items. And then we had our data set. What we found was pretty extraordinary. These were not little differences. Sometimes one consumer was offered a price 23 % higher than another consumer for the exact same good at the exact same time. The total price for the consumer's baskets differed by an average of about 7%. So the best way to explain this, if you and I were standing in line at Target right now and we were holding the exact same box of Cheerios, we would expect to pay the same amount, right?

8:29Same store, same day, same item. Yep. Like we're in line together, IRL, right? And you check out, your box costs$4.99. You're still bagging your box. I check out next. My box is$6.99. I look over at you and I'm like, wait, how much did you pay? And you're$4.99. And then we both look at the cashier. Wait, why am I paying$6.99? If you and I stepped out of line, fired up our phones, got on Instacart, got the same box of Cheerios from that same pickup location that we're standing in, there is a very good chance that we will be offered different prices. We found price variation on 75 % of the items in the grocery carts.

9:11And we found that 100 % of our secret shoppers got variable pricing at some point during the experiment. So no one was immune, and 75 % of the items were offered at variable prices were effectively part of this big experiment. Wow. Lindsay also says that this isn't a niche issue. Grocery delivery picked up during the pandemic, and a range of people use the services today. In an Instacart 2025 economic report, the company said that 68 % of its customers considered Instacart essential. App trackers estimate Instacart has over 14 million active users. Some people have kept up with it because they can't get to the store, or maybe it's not accessible through public transportation.

9:57Maybe you, you know, have an injury or something like that and can't sort of pick up your own groceries. And also Instacart allows shoppers to grocery shop at a whole host of different outlets, low end grocers, discount grocers, big box grocers, high end grocers. Right. So it really covers the watershed of different consumer preferences. And so one of the reasons why we felt like this study was, you know, so important, but also we felt like our findings were so problematic is because this isn't sort of a niche market. This really is increasingly a big part of the American grocery experience. And what did you find about why or what could explain whether a customer would get a higher or lower price?

10:41Was there anything about the profiles of shoppers that would inform what prices Instacart was serving them or was it random? So I think we have a couple of suggestive answers to this. The first is we know from Instacart's own public statements that part of the reason they do this is just to test how much they can charge for given items. So they're really conducting pricing sensitivity analyses, figuring out how high they can take Instacart's markup on top of the retailer's price before you take an item out of your cart or close your computer and walk to a brick and mortar grocery store. On our end, we tested a whole host of demographic information to try to see if we could predict which prices you would get based on the information that we had.

11:28We didn't find any statistically significant patterns, so we can't rule out that it was random. However, we did find that people were sorted into distinct price groups. So there were some people who got higher prices pretty much across the board. There were some people who got lower prices more often than not.

11:54According to the report, Instacart confirmed that the findings from the tests accurately reflect its pricing experiments and strategies. In a statement to Groundwork and Consumer Reports, Instacart said, just as retailers have long tested prices in their physical stores to better understand consumer preferences, a subset of only 10 retail partners, ones that already apply markups, do the same online via Instacart. These limited, short-term, and randomized tests help retail partners learn what matters most to consumers and how to keep essential items affordable. But in statements, Instacart also said that there's been reporting that's inaccurately characterized the mechanisms behind some of its pricing experiments.

12:39And Bloomberg's Leah Nylan says Instacart stressed that the retailer is the one that sets the base price of groceries they sell on the app. Retailers often have their own reasons and strategies for engaging in pricing experiments, as groceries have notoriously narrow profit margins. They felt that this study was sort of unfairly blaming them for something that they don't have a lot of control over. Another thing I thought was interesting in the statement that they released was that they really reiterated that they weren't using surveillance pricing. In other words, they weren't using customers' personal information to set prices, that it was like random A-B testing.

13:15Can you explain further the difference between algorithmic pricing and surveillance pricing? And why do you think Instacart was so clear in making that distinction? Yeah. So when people think of the term surveillance pricing, they generally think of a company that is taking a lot of information about them and sort of coming up with a price that is very personalized to them. So they'll have all of these different characteristics and they will use those particular things about you to set the price. Instacart was saying we weren't doing that. It's not like we were like deciding that you as Jane Doe in here, you know, we have all this information about you.

13:53and so we were tailoring the price to you. We were just like using various algorithms that sort of like are testing various things about our app to see if maybe, you know, like if we highlight this type of deal, you're more likely to click on it. And the way they do that is this thing called A-B testing where they'll show one customer one thing and they'll show another customer something slightly different. And then they see which one is more effective for whatever it is that they're testing.

14:23Coming up, how price experiments started creeping into more industries, and how lawmakers and regulators are responding.

14:37Bloomberg Invest returns to New York on March 3rd and 4th, where the sharpest voices across banking, asset management, and private capital will discuss the forces reshaping finance. Powered by Bloomberg's Global Newsroom and data from the Bloomberg Terminal, this flagship summit will cover everything from AI-driven disruption and central bank policy shifts to the emerging risks and opportunities in private credit. Join the conversation and register today at bloomberglive.com slash invest.

15:14Like it or not, consumers are used to being experimented on, at least in some arenas. You know, you sort of think when you're buying airline tickets, you know, or maybe concert tickets or something, you know that it's a good that there is a limited supply, right? Bloomberg's Leon Island. There are only so many seats on a flight. So, yeah, maybe like the price might go up when there's like fewer seats left. You know, concert tickets are sort of notoriously, you know, get expensive the more close to the date that you get. But with groceries, you know, there's sort of always been this idea that you go into a grocery store and everybody gets charged the same price.

15:48Increasingly, like grocery companies and other companies along these lines are trying to change that. And they're very interested in finding more and more details about people to try and tailor the pricing towards them. They've been introducing what they call electronic shelf labels at grocery stores so that they might even have an opportunity to change that shelf price throughout the day. And I think that is a surprise to just the regular average consumer because we're not really expecting that. We are used to going into the store and the price is the price. When the Groundwork Collaborative and Consumer Reports investigation revealed that the grocery company Instacart was running its own pricing experiments, Many members of the public were taken aback.

16:31Yeah, from lawmakers and regulators, there was sort of a big hue and outcry because this is something, at least from the regulatory perspective, that they've been concerned about for a while. They know that in this day and age when there is a lot of data about consumers out there, that retailers are trying to find out this information about consumers and do this sort of personalized pricing. But it raises a lot of red flags because you don't know exactly how various factors are increasing or decreasing the price. There are various federal laws that prohibit you from raising or lowering prices based on particular characteristics.

17:07So, for example, you're not supposed to charge people different prices based on race or sex or religion or things like that. But when it's a black box like this, like you don't know exactly what factors the company is using to make the decision about pricing. Among lawmakers, it was also sort of a red flag because food is an essential thing. You know, everyone has to eat. And especially at a time right now when, like, grocery prices have been such a touchpoint with voters, it really did get a lot of lawmaker reaction. After the Instacart report came out, members of Congress, including Senator Chuck Schumer and Senator Amy Klobuchar, wrote letters to Federal Trade Commission officials expressing their concerns.

17:48Reuters reported that the FTC has sent inquiries to Instacart regarding its use of AI-enabled pricing tools. FTC chair Andrew Ferguson posted on X this week that the FTC, quote, began investigating these potentially unlawful practices in the spring. With all this new scrutiny on the role of technology in price setting, I asked Leah about how long the issue has been on regulators' radars. This is something that the Federal Trade Commission, which is the U.S.'s primary consumer protection agency, has been concerned about for a while. During the Biden administration, the FTC had started an investigation into a lot of retailers.

18:25They have this special power that they can ask retailers to sort of turn over information for the purposes of a study. So they had asked a bunch of major retailers to do that. And then they were looking essentially at how frequently they use these sort of algorithms to help determine pricing. And that study is still underway. The Trump folks didn't kill it upon coming into office. But we still don't really have the results of that yet. And the FTC in particular had already had an investigation into Instacart over whether it's Instacart Plus, this is its membership program, was sufficiently transparent enough about its pricing and cancellation policies.

19:04And Instacart has to pay$60 million back to consumers as a result of that investigation? Yes, they're refunding consumers millions of dollars as a result of this investigation because the FTC found that they weren't being transparent enough about the terms of this membership program and how people could cancel it. Because the FTC found that frequently people sort of didn't realize that they were enrolled in this program and they would try and cancel. And instead of giving them their money back, Instacart would give them a credit, which violates some consumer protection laws about memberships. On Monday, Instacart announced that retailers would no longer be able to use Eversight's software to run item price tests, effective immediately.

19:47Now, the company said in a statement, if two families are shopping for the same items at the same time from the same store location on Instacart, they see the same prices. Period. So Instacart, you know, announced they're stopping this tactic, which marks one sort of resolution to this story. But could what happened here with Instacart set a precedent for other industries that are experimenting with this kind of price setting? I do think that if anybody else is experimenting with this price setting, they probably have seen like what the sort of criticism that Instacart got and are probably going to be a little bit more reluctant to at least admit that they are doing this.

20:27And or, you know, when they get blowback, be a little bit more willing to pull back on it. So it will be interesting to see, you know, sort of as we go forward, how many other companies sort of try these types of tactics, how transparent they are about that's what they're doing and sort of what the reaction is going forward. The basics of consumer protection law is that you have to be very transparent about what you are selling and at what price. If you don't do that, the regulators will come after you.

21:03This is The Big Take from Bloomberg News. I'm Sarah Holder. To get more from The Big Take and unlimited access to all of Bloomberg.com, subscribe today at Bloomberg.com slash podcast offer. Thanks for listening. We'll be back tomorrow.

21:32for the Masters in Business podcast. Every week, we bring you fascinating conversations with the people who shape markets, investing, and business. CEOs, fund managers, billionaires, Nobel laureates, traders, analysts, economists, everybody that affects what's going on in the market, whether you own stocks, bonds, real estate, commodities, crypto, you really need to hear these conversations. Sometimes it's behaviorists like Dick Thaler or Bob Schiller. Sometimes it's fund managers like Peter Lynch, Bill Miller, Ray Dalio. Sometimes it's authors, Michael Lewis, author of The Big Short, and Moneyball.

22:13Regardless of the conversation, these are the folks that move markets each week. That's the Masters in Business podcast with me, Barry Ritholtz. Listen on Apple, Spotify, or wherever you get your podcasts.