So you say you hate ‘surveillance pricing’

Spencer Platt/Getty Images

“Surveillance pricing” sounds very, very bad. Consumers think that companies are tracking their search history, location and income to charge more for groceries and other purchases.

Naturally, lawmakers are trying to crack down on this supposed nightmare.

In Colorado, the legislature passed a bill this summer to ban companies from using customers’ personal data to set individualized prices. The concern is that two people could be charged different amounts for milk based on what the data spits out about their willingness to pay.

But Colorado Gov. Jared Polis saw a big problem with the bill.

“Many Coloradans won’t get discounts on items they buy if I were to sign this,” Polis said in a June letter. “It would punish differentially lower prices, not just higher prices.”

Americans are channeling their rage over high prices, AI and giant companies gathering their data into the issue of “surveillance pricing.” A growing number of cities and states are responding with legislation to reel in companies’ use of customers’ personal data to set prices, especially at large grocery stores and fast-food chains. Some bans also target electronic shelf labels at stores. Unions have backed these restrictions on electronic labels because they fear the technology will eliminate the need for workers to change thousands of paper tags in stores each week.

But the tech is more neutral, and Polis’ veto exposed a contradiction with the growing backlash to the practice: the same algorithm that companies use to charge more also can deliver lower prices, reduce food waste and create other benefits for consumers. Broad state and local bans cut off the possibility of lower prices and other tech that delivers benefits, economists, attorneys and marketing experts say.

“Some of the surveillance pricing laws are not only unlikely to protect consumers, they may even harm them,” said Jean-Pierre Dubé, a marketing professor at the University of Chicago’s Booth School of Business.

Evidence also shows that neither “surveillance pricing” nor electronic shelf labels increase overall prices.

“There is essentially no change in pricing behavior after these digital labels are installed,” a study tracking more than 180 million price records at grocery stores last year found .

The debate around “surveillance prices” often blurs two different practices: dynamic pricing, which adjusts prices based on supply and demand and other market conditions, and personalized pricing.

Dynamic pricing has existed for decades. Happy hour, lunch specials and end-of-season clearance sales are familiar examples. Consumers also encounter dynamic pricing when they buy plane tickets online or order Uber rides.

There is little evidence that dynamic pricing raises prices overall. In some industries, including airlines and restaurants, it has actually enabled companies to offer lower prices.

Airlines have long used dynamic pricing to offer lower fares to travelers who are flexible on their travel dates or willing to fly at off-peak hours. A large restaurant chain that used an algorithm to adjust food delivery prices found that it helped smooth out orders throughout the day, according to a 2023 Harvard University study . The algorithms slashed prices during slow hours to attract customers who would otherwise dine elsewhere, bringing the average price down across the day.

Nicolas Economou/NurPhoto/Getty Images

Lower prices weren’t the only benefits of dynamic prices. In the grocery industry, research from the University of San Diego in 2023 found that dynamically pricing perishables cuts down food waste by 21%.

But as AI has expanded companies’ ability to personalize prices across more industries, the practice has gradually been described in more nefarious terms such as “AI surveillance pricing” and “price discrimination.”

Economic populists such as Sens. Elizabeth Warren and Bernie Sanders blame “surveillance pricing” for driving up prices, claiming AI will make things worse. Lindsey Owens, the CEO of progressive economic think tank Groundwork Collaborative, gained traction for her warnings about the opaque ways corporations use customers’ personal data to rip them off.

Owens pointed to a patent Walmart received in 2023 for technology that can remotely deploy dynamic price changes to electronic shelf tags: “If a customer has tuna fish, they may be offered a different price for mayonnaise,” the patent says.

Facing public uproar, Walmart last month defended its practices and told customers: “We don’t set different prices based on who you are or the time of day.”

Walmart has little incentive to abandon its business strategy of stable, consistent prices, said Ioannis Stamatopoulos, a professor of operations at the University of Texas at Austin who has studied dynamic pricing.

The company relies on a model known as everyday low prices — rather than sales and coupons — to draw shoppers. Rapid price fluctuations would alienate Walmart’s customers, Stamatopoulos said.

“They would rather keep you happy and coming back instead of risking your continued business for a few opportunistic cents,” he said.

Spencer Platt/Getty Images

“监控式定价”听起来非常糟糕。消费者认为，公司会追踪他们的搜索历史、位置和收入，从而对食品杂货和其他商品收取更高的价格。

当然，立法者们正在努力打击这种所谓的噩梦。

今年夏天，科罗拉多州议会通过了一项法案，禁止公司利用客户的个人数据来设定个性化价格。人们担心，根据数据分析得出的消费者支付意愿，两个人购买牛奶的价格可能会有所不同。

但科罗拉多州州长贾里德·波利斯认为这项法案存在一个大问题。

“如果我签署这项法案，许多科罗拉多州居民将无法享受商品折扣，”波利斯在六月份的一封信中写道。“这将惩罚那些价格较低的商品，而不仅仅是价格较高的商品。”

美国民众对高物价、人工智能以及大型企业收集用户数据的愤怒情绪，正集中体现在“监控定价”问题上。越来越多的城市和州正在通过立法来遏制企业利用消费者个人数据定价的行为，尤其是在大型超市和快餐连锁店。一些禁令甚至针对商店的电子货架标签。工会支持这些针对电子标签的限制，因为他们担心这项技术会使员工每周更换数千个纸质标签的工作量减少。

但这项技术本身更为中立，波利斯的否决揭示了与日益高涨的反对声浪之间的矛盾：企业用来抬高价格的同一算法，也能降低价格、减少食物浪费，并为消费者带来其他益处。经济学家、律师和营销专家表示，州和地方政府的广泛禁令扼杀了降低价格和其他带来益处的技术的可能性。

“一些监控定价法不仅不太可能保护消费者，甚至可能损害消费者的利益，”芝加哥大学布斯商学院市场营销学教授让-皮埃尔·杜贝说。

证据还表明，“监控定价”和电子货架标签都不会提高整体价格。

去年一项追踪杂货店超过 1.8 亿条价格记录的研究发现，“安装这些数字标签后，定价行为基本上没有变化”。

围绕“监控价格”的争论常常混淆了两种不同的做法：动态定价（根据供求关系和其他市场状况调整价格）和个性化定价。

动态定价已经存在了几十年。欢乐时光、午餐特价和季末清仓甩卖都是常见的例子。消费者在网上购买机票或叫优步时也会遇到动态定价。

几乎没有证据表明动态定价会推高整体价格。在某些行业，例如航空和餐饮业，动态定价实际上使公司能够提供更低的价格。

航空公司长期以来一直采用动态定价，为出行日期灵活或愿意在非高峰时段出行的旅客提供更低的票价。哈佛大学2023年的一项研究表明，一家大型连锁餐厅使用算法调整外卖价格，发现这有助于平抑全天的订单波动。该算法会在客流低谷时段大幅降低价格，吸引那些原本会选择去其他地方用餐的顾客，从而降低全天的平均价格。

Nicolas Economou/NurPhoto/Getty Images

动态定价带来的好处不仅限于降低价格。圣地亚哥大学2023年的一项研究发现，在食品杂货行业，对易腐食品进行动态定价可以减少21%的食物浪费。

但随着人工智能扩大了公司在更多行业中个性化定价的能力，这种做法逐渐被描述为更邪恶的术语，例如“人工智能监控定价”和“价格歧视”。

伊丽莎白·沃伦和伯尼·桑德斯等经济民粹主义者将物价上涨归咎于“监控定价”，并声称人工智能会让情况变得更糟。进步经济智库“基层合作”（Groundwork Collaborative）的首席执行官林赛·欧文斯因其对企业如何以不透明的方式利用客户个人数据牟利的警告而备受关注。

欧文斯指出，沃尔玛在 2023 年获得了一项专利，该专利涉及一种可以远程向电子货架标签部署动态价格更改的技术：“如果顾客购买了金枪鱼，他们可能会收到不同价格的蛋黄酱，”该专利写道。

面对公众的强烈抗议，沃尔玛上个月为其做法辩护，并告诉顾客：“我们不会根据你是谁或一天中的时间设定不同的价格。”

德克萨斯大学奥斯汀分校运营学教授伊奥尼斯·斯塔马托普洛斯（Ioannis Stamatopoulos）研究过动态定价，他表示，沃尔玛几乎没有动力放弃其稳定、一致的价格的商业策略。

该公司依靠“天天低价”的模式——而非促销和优惠券——来吸引顾客。斯塔马托普洛斯表示，价格的快速波动会疏远沃尔玛的顾客。

他说：“他们宁愿让你满意并让你再次光顾，也不愿为了几个投机取巧的钱而冒失去你继续光顾的风险。”