What happens if AI personalizes every price you see?

Dear Friends,

We know the adjectives that probably immediately popped into your mind when you read that headline: Orwellian. Dystopian. Scary. Unfair.

But the adjectives don’t all need to be negative. We can’t rule out the idea that there are ways for consumers to benefit from that trend, or at least have some agency over the course it takes.

In this edition of the newsletter, my co-author Arnab Sinha and I stretch some of the current trends in AI-driven pricing to their practical limits and explore whether – and how – a world of intensely personalized day-to-day prices could bring more benefits than drawbacks to consumers.

Where do things stand right now?

What would happen if artificial intelligence set almost every price you see in your day-to-day life, and what if the AI relied on your personal data to set those prices dynamically? Answering those questions is important and increasingly urgent, because we’re closing in on that future faster than you might imagine.

Two stories this month raised our eyebrows especially high, because of their broader implications for business and for society.

In mid-June on its quarterly earnings call, Delta Air Lines revealed that it hopes to have an AI-powered revenue management system trained and ready to set prices across 20% of its domestic network by the end of 2025, up from around 3% currently. Delta is not alone in partnering with the Israeli tech company Fetcherr to put such a system in place. This article from last summer refers to legacy airline revenue management systems as “ancient” and cites examples from several other carriers that are pursuing that path.

Earlier this week, the Wall Street Journal ran a story about a Norwegian retail chain that changes shelf prices on some products up to 100 times per day. The technology that makes these changes visible to consumers is the electronic shelf label (ESL), which we discussed in an edition of our newsletter last year.

It seems every day brings fresh stories about how artificial intelligence is penetrating deeper into our personal lives. To answer what would happen if AI relied on personal data to set most prices dynamically, we need to rephrase the questions. The more incisive questions, in our view, are these:

• What are the benefits of personalized dynamic pricing?

• How much price discrimination are people willing to tolerate, especially when companies use customers’ personal data to set prices?

• What can companies do to increase the tolerance of customers and mitigate the risk of a backlash?

The benefits of personalized dynamic pricing

The fact that many people can travel affordably by air is a byproduct of dynamic pricing. When airlines began to charge more for some classes and much less for others – and vary the prices over time – they opened up much greater access to air travel and helped lead a surge in consumption. The number of passenger miles flown in the United States rose from 191 billion in 1980 to an initial peak of 608 billion in 2007, before the Great Recession.

Repeated exposure to the reasons for the price differences – which have historically been advance booking, loyalty status, booking class, and departure time – mean that passengers can often work out on their own, without emotion, why the person seated next to them may have paid 50% more or 50% less for their ticket to the same destination.

Getting to that point requires consistent education over time to help customers understand the nature of price differences when established norms offer little or no clear guidance on what a fair price is. Ride-share services are making the same transition. Their growth has also led to lower overall prices and better options for point-to-point travel, despite the occasional protests when “surge pricing” occurs.

How much personal price differentiation will customers tolerate?

The answer to this question revolves around the question of fairness. It’s popular right now to sprinkle the word “fairness” into stories about personalized dynamic pricing, but the phrase isn’t usually helpful without a common understanding of what a fair price is, how much pricing fairness matters, and how companies can influence it

That’s why I spent over a year at the BCG Henderson Institute (BHI) researching the topic of price fairness, including a survey of over 10,000 consumers around the world. The results of that research motivated two chapters in the Game Changer book.

Is it fair when the person next to you in line paid half of what you did or paid nothing at all, because of who they are? Think of an amusement park or museum, where ticket prices are not trivial. The person standing next to you in line may have paid half price – or nothing at all – because they are over or under a certain age, serve in the military, attend a certain school, or belong to a particular loyalty program. Our research at BHI showed that people around the world tend to accept such forms of price discrimination, but their tolerance often depends on their own age, nationality, income, and in some cases even their political affiliation.

Our working definition of a fair price is one that shares value equally across buyers and equitably between buyer and sellers. This focus on value sharing means that only true peers will be charged the same price. A seller can charge a higher price to a customer who derives much higher value, without perceived unfairness, as long as the seller communicates the clear rationale.

That last sentence is the kicker. It explains why many dynamic pricing initiatives face heavy public resistance from customers. The obligation to justify value as price change – higher or lower – is an ongoing process of education and information, not a one-off message or something to bury in fine print.

The fairest price – not the highest price – optimizes profits

Success with personalized dynamic pricing will depend on whether customers find something they can see, feel, or experience as a benefit. If companies change prices without a corresponding change in value, they risk a backlash if customers perceive those changes to be unfair.

But we also learned in our research at BHI that fair prices not only keep customers happy, but also shareholders. Companies that focus on offering fair prices over time perform better than companies that focus on maximizing prices by extracting full value. That leads to recommendations for companies to take advantage of AI-driven personalized pricing in ways that benefit their customers and their bottom line.

• Vary prices in a wider range: A wider price range that keeps average prices roughly similar can result in a significant increase in volume or consumption. That’s why retailers are initially talking about using ESLs or other means to offer lower prices, not higher ones. The retail chain in Norway usually lowers selected prices over the course of the day. But the balance and discipline is critical to maintain overall margins and avoid a price war. This flexibility is widest for offerings with little or no marginal cost, as I explained back in 2019 under the term progressive pricing.

• Communicate clearly and constantly about value: Price variation – whether in its simplest form or its extreme personalized form – requires a communication strategy that makes the justifications positive, transparent, and explicit.

• Preserve trust regarding data usage: Companies need to become trusted custodians of customers’ personal data. There are lines between abusive and constructive uses of personal data.

We also recognize the potential for unintended consequences if personalized pricing becomes a de facto standard in many sectors. In retail, for example, the key value items that help define a retailer’s price positioning may lose their ability to serve as price anchors. Similarly, “magic” price points or price thresholds lose their meaning.

As always, please continue to share your thoughts and questions with us. If you haven’t ordered your copy of Game Changer yet, you can do that HERE. Thanks for your interest and support.