Podcast: Download (Duration: 39:13 — 45.2MB)
AI dynamic pricing uses data about an individual shopper, such as their purchase history, browsing, and loyalty status, to set the price that shopper sees. Two people can look at the same product at the same moment and pay different amounts.
Airlines, Instacart, and Wendy’s have all drawn backlash for versions of it. Firms now sell the same data and tools to ecommerce stores, so a small seller can price each product by shopper.
I have actually thought about using AI dynamic pricing in my own store, and the news stories gave me second thoughts. In this episode, my co-host Toni Herrbach and I debate whether it is fair, share the times it has already happened to us, and decide whether sellers should use it.
Get My Free Mini Course On How To Start A Successful Ecommerce Store
If you are interested in starting an ecommerce business, I put together a comprehensive package of resources that will help you launch your own online store from complete scratch. Be sure to grab it before you leave!
Table of Contents
Key takeaways
- AI dynamic pricing is personal. Traditional dynamic pricing raises prices for everyone during peak demand, while AI dynamic pricing changes the price based on what it knows about you.
- It is already happening. Toni’s usual Delta round trip to Austin jumped from about $247 to close to $400, and none of her usual tricks brought it down.
- Small differences add up. In one Instacart example, the same box of Wheat Thins ranged from about $4 to $5, a roughly 20% spread most shoppers would never notice.
- The payoff can be real. Case studies show profit increases in the low double digits.
- Loyal customers carry the risk. The shoppers you know the most about are your most loyal ones, so personalized pricing can punish loyalty.
- Neither of us is likely to use it. My wedding hankie customers are already price insensitive, and Toni thinks it widens the gap between the haves and have nots.
What is AI dynamic pricing?
AI dynamic pricing is when a business uses AI to gather data about a customer and then prices its products based on what the AI predicts that customer is willing to pay. It goes further than the dynamic pricing airlines and Amazon have used for years.
The data can include your purchase history, your search behavior, your loyalty status, and even signals about why you are buying. A store does not need you to log in, because tracking tools can identify shoppers across the web.
| Pricing type | What sets the price | Example |
|---|---|---|
| Demand-based dynamic pricing | Overall demand, the same for everyone | Higher airfare at Thanksgiving |
| Surge pricing | Real-time supply and demand | Uber rides in the rain |
| Time-based pricing | Time of day | Restaurant happy hour |
| AI dynamic pricing | Data about the individual shopper | A loyal flyer seeing a higher fare on a familiar route |
AI dynamic pricing examples from Delta, Instacart, and Wendy’s
The best-known AI dynamic pricing examples come from Delta, Instacart, and Wendy’s, and each one triggered public backlash. Here is what happened with each.
Delta’s AI airfare pricing
Delta’s use of AI pricing drew the most anger, partly because of a story about a woman flying to a relative’s funeral. AI flagged her unusual buying behavior, and her fare went up because Delta knew she needed to go.
The president of Delta has said the system will save customers money in certain cases, such as off-peak periods. Toni and I are skeptical.
Toni thinks it already happened to her. She flies Delta to Austin every three months to record, on the same flights, and the round trip was always about $247 no matter the time of year.
On her last booking, the fare was close to $400, about $140 more than normal. She tried all her usual tricks and could not get it any lower, so she paid it.
Toni believes Delta knows she is chasing her next status level and will not book through someone else’s account. Loyalty programs know their most loyal customers best, and in Toni’s experience they treat them worse because they can get away with it.
Instacart’s personalized grocery prices
Instacart’s case study showed the same product priced differently for different shoppers. Wheat Thins ranged from about $4 to $5, a spread of roughly 20% depending on who you are and how you spend.
Groceries are the perfect place to hide it because each cart holds many items. A shopper who pays 20 cents more here and 72 cents more there only notices when a bill that is usually $72 comes to $84.
Wendy’s surge pricing backlash
Wendy’s drew backlash when it talked about surge pricing, with burgers costing more at lunch and less during slow hours. I asked Toni how that differs from happy hour, and she pointed out that most people cannot get to a restaurant at 4 p.m.
How AI dynamic pricing differs from surge pricing
AI dynamic pricing differs from surge pricing because it targets you personally, while surge pricing hits everyone at once. Everyone pays more at Thanksgiving, and everyone gets the same Uber surge in a rainstorm.
Toni compared it to COVID, when masks, hand sanitizer, and toilet paper went through the roof. Sellers’ costs really did go up, so the higher prices were shared across everybody.
With AI dynamic pricing, one business profits from one person’s habits, tendencies, or misfortunes. Toni also says surge pricing leaves you a choice, and in a New York City downpour she would rather walk to the subway than pay $100 for a ride.
Why AI dynamic pricing feels unfair to shoppers
AI dynamic pricing feels unfair because it takes data that shoppers share for convenience and uses it against them. Toni is fine being tracked for ads, and she draws the line at pricing.
She recently spent 25 minutes on the Oats Overnight website after her copywriter sent her one of their ads. She is happy to get their retargeting ads, and she would welcome promotions from a brand she loves like Yeti.
Toni’s biggest objection is that personalized pricing punishes loyalty. A store knows the most about the customers who visit and buy often, so those are the people who would end up paying more.
Finding out stings. Toni’s friend Sandy hired the same cleaning service Toni used, paid significantly more for a smaller house, and fired them.
Toni asked me to picture meeting a stranger in a Whole Foods parking lot with the exact same car. They paid $40,000 at the same dealership in the same month, and I paid $50,000.
Personalized pricing already happens offline
Personalized pricing already happens offline every day, and that is my main argument. The only real difference with AI dynamic pricing is that software is setting the price instead of a human.
- Event venues might charge $100 to rent a space for a business event and $1,000 for the same space for a wedding.
- Banks whisk customers with large balances past the line to a private consultant, who then upsells them.
- Car prices vary by dealer, and the same Toyota Toni bought last year was significantly cheaper across the country.
- Service businesses negotiate, and our new cleaning service charges our neighbors less because they have been grandfathered in for 10 years.
- Software companies like Klaviyo know how much your business makes, and enterprise plans at Klaviyo and Shopify Plus are negotiable.
- Funeral homes sell to people at their most emotional, and when my dad passed away, they tried to upsell me a $20,000 coffin.
Should ecommerce sellers use AI dynamic pricing?
Ecommerce sellers can make real money with AI dynamic pricing, and Toni and I agree it is most defensible for luxury products and wrong for necessities. Case studies show profit increases in the low double digits, and firms already sell the data and tools to stores.
I put the business case to Toni this way. If you could make 20% more from a portion of your customers and only 3% of people ever found out, would you do it?
Toni thinks you are a jerk if you do this with a necessity. Most ecommerce stores sell non-essential products, so she has less of a problem with dynamic pricing on luxuries.
I also argued that we are in a K-shaped economy, where the people with money keep getting richer and the stock market is at an all-time high. If you target wealthy, price-insensitive shoppers, you are effectively selling in a boom.
Our conclusion was simple. If you sell luxury handbags, AI dynamic pricing is fair game, and if you sell doorknobs, please do not do this to people.
How AI dynamic pricing can hurt your store’s credibility
AI dynamic pricing can hurt your credibility the moment a shopper sees a lower price in an incognito window or on a friend’s computer. Once that happens, they will question every price on your site.
Toni ran dynamic pricing in her old jewelry business, and it was based on aggregate data across all shoppers. She would not use individual pricing if she had a store today.
Her bigger worry is the economy. So many people are struggling that she thinks personalized pricing only helps the businesses that can afford the tools, and the gap between the haves and have nots grows even wider.
Why I won’t use AI dynamic pricing in my hankie store
I won’t use AI dynamic pricing in my store because it would not help my business much. Most of my customers buy handkerchiefs for weddings, and wedding shoppers are already price insensitive.
People also buy our hankies for funerals. A shopper who finds us by searching for a funeral keepsake hankie has an urgent, time-sensitive need, so in theory I could charge whatever I wanted.
I would never add a funeral surcharge. After seeing how the funeral industry treated my family, I know how it feels, and that kind of pricing is wrong.
AI dynamic pricing vs buy now, pay later: which is worse?
I think buy now, pay later is ethically worse than AI dynamic pricing, because it puts struggling people further into debt. Toni was all for buy now, pay later when we covered it in episode 653, and I was the one with ethical problems.
AI dynamic pricing mostly charges people who have money a bit more, often without them noticing. Toni thinks I have a warped sense of what counts as bad, and we would love to hear where you land.
How shoppers can avoid AI dynamic pricing
Shoppers can avoid the worst of AI dynamic pricing by tracking prices on what they buy often and comparing before they pay. Here is what Toni and I suggest.
- I buy exchangeable plane tickets. I track the price after booking and rebook when the fare drops.
- Compare prices across browsers and accounts. When Toni’s fare jumped, I wondered what price someone else’s Delta account would have shown, and a lower price in an incognito window is a sign you are being priced personally.
- Toni keeps a grocery price book. She knows milk was $2.89 a gallon and her apple juice six-pack went from $3.89 to $4.67, and she offers a downloadable price book on her site.
Most shoppers do not track prices this closely, which is exactly why personalized pricing works. Checking your staples is the simplest defense.
Frequently asked questions
Is AI dynamic pricing ethical?
AI dynamic pricing is most defensible on luxury products sold to price-insensitive shoppers and hardest to justify on necessities. Charging more to someone in a desperate situation, like a traveler flying to a funeral, crosses the line for most people.
How do I know if a website is charging me more than other shoppers?
Check the same product in an incognito window, on another device, or on someone else’s account. A consistently lower price elsewhere suggests the site is pricing you personally.
Does Instacart charge different prices to different customers?
In one Instacart example, the same box of Wheat Thins ranged from about $4 to $5 depending on the shopper, a spread of roughly 20%. Most shoppers never notice because the differences are spread across a full cart.
Do airlines raise prices based on your search history?
Airlines have long used dynamic pricing, and many travelers see higher fares when they return to a route they searched before. Delta’s newer AI pricing goes further by using what it knows about the individual flyer.
What is the difference between dynamic pricing and personalized pricing?
Dynamic pricing changes prices for everyone based on demand, like higher airfare at Thanksgiving. Personalized pricing, which AI makes possible at scale, changes the price for one shopper based on that shopper’s data.


