Dynamic Pricing: “How Much” Is Not a Simple Question

The algorithms that price airline seats are being applied to cameras and cereal

Loading the podcast player...

Steven Cherry: Hi, this is Steven Cherry for IEEE Spectrum’s “Techwise Conversations.” Listen to this.

“Hi um, this other store has these for twenty cents less.”


“Match it.”

“Match it.”

“Match it.”

“Match it.”

“Twenty cents less? Wha-?”

That’s the start of a television commercial from Walmart, offering to match any price, anywhere, on any product they have in the store. On the flip side, Amazon has an app called Price Check that lets you photograph a barcode with your phone and look up the Amazon price for it. Of course, you can then order it with Amazon’s famous one-click buying, right from the aisles of the store you’re cuckolding.

Stores have always fought with one another, but in the past, the battles were over variety, quality, value, and, of course, location, location, location. Nowadays, every retailer seemingly has everything, and it’s all about price, price, price. Consumers have an unparalleled ability to shop around for the best prices, whether it’s knowing the exact markup that an auto dealer gets on every model in the showroom, or scraping off that last 20 cents on a box of cereal at the supermarket.

Or do they?

A recent study by a shopping site showed that retail websites often don’t show the best price for ordinary goods like digital cameras.

My guest today, to talk about the changing world of retail pricing, is Eric Best. He’s the CEO of Mercent, a Seattle software firm that helps companies better price their goods. He joins us by phone.

Eric, welcome to the podcast.

Eric Best: Thank you very much, Steven.

Steven Cherry: Eric, we’re used to airline tickets changing all the time in price, and if we buy on eBay, of course, the price of something can change from one auction to the next. But now we’re starting to see variable pricing for all sorts of goods at all sorts of stores. Is the price of toilet paper going to change as often as airline seats, and how pervasive is this?

Eric Best: Yeah, at Mercent we think it is. We think that consumer product prices are changing more frequently, and those prices are being expressed more broadly. And it’s interesting you start with tickets and travel. Certainly, dynamic pricing has been around for a long time in those industries. I think there are two big, disruptive trends that we’re observing in the marketplace that provide evidence that this dynamic pricing model is moving into consumer goods in a big way. The first is that we can see that consumer behavior is changing, and first and foremost that means that shoppers are becoming more comfortable with the idea that a price is going to change frequently, and increasingly, very frequently, meaning intrahourly on certain sites, like eBay, Amazon, and Google, and we expect this to continue.

That consumer behavior, or comfort with the dynamic price, is driven by the prevalence of smartphones, so consumers are armed with real-time pricing data when they’re making purchasing decisions in brick-and-mortar stores. It’s driven by the frequency with which retailers are willing to submit or promote products with deep discounts via e-mail to existing customers or prospective customers. And it’s certainly driven by recent emerging retail business models from the likes of Groupon, with Groupon goods, and ShopLocal, Amazon, and eBay, of course, but also the private-label sites Gilt Groupe and Rue La La. So consumer behavior is changing. We’re seeing retailers react to that with the application of technology as well.

Steven Cherry: So your business starts with the idea that we now have price aggregators online. I mean, I guess we . . . for a while we’ve had the Orbitz and Travelocity sites, but now we have the ones that you mention just now, and PriceGrabber and Nextag and Google Shopping and so forth, and so your software helps retailers with their prices, getting them to show up properly? How does it work?

Eric Best: Yeah, so the shopping engines that you mentioned, Nextag and Shopzilla and so on, have been around for quite some time. You know, a decade or so. So the idea of aggregating product availability and regular updates on product pricing is certainly not new. But what is new is the frequency with which these decisions are being made by retailers.

So just talking about where Mercent plays in the retail ecosystem, we integrate our software platform directly with the e-commerce storefront and the fulfillment centers, the warehouse that the retailer may operate. And we take all the real-time information out of that catalog, including product pricing, product inventory availability, but also things like cost of goods sold, minimum advertised price restrictions that may be set by a manufacturer or supplier, and, of course, gross margin data for each product. And we look at that data, even down to the size and color variations that may exist. So when we’re talking about a particular product, like a North Face Denali Jacket, you know, a retailer has the ability to monitor and set a price for the women’s red extra-large variant on that product, separate from, say, the black or the small or, you know, other derivatives of the same parent product.

And so then from there, once we have the information from the retailer’s line of business software systems, we layer in our own real-time Web analytics. That tracks where shoppers are coming from, from across the Web, and what they’re buying, and most recently we’ve added to that data mix real-time monitoring of product availability and pricing so that our clients can use the Mercent platform to keep tabs on what’s happening in the market. And the rate at which we’re able to collect that competitive intelligence, match competitor products against our clients’ own catalog, and then ultimately determine a new price point for the SKU, is currently at about 2 million products per hour. Which means that over the course of our entire portfolio, we can reprice every product SKU at least once a day, and certainly more often than that for more-aggressive sellers.

Steven Cherry: So the color of a jacket was an interesting example. I guess in the past retailers would make too many extra-large red jackets, and eventually they would show up at some discount warehouse or something. But now a retailer can decide to sort of lower the price on that because they can see already that there are too many of them. Is that the sort of thing that would change a price quickly?

Eric Best: Yeah, that’s exactly right. But there are other signals that—not just related to end of life, or product life cycle. There are many other signals in terms of the data set that we can collect and expose from across the Web. Most notably, if one of our client’s competitors runs out of inventory on a particular product, we’re in a position where we can actually raise that price if we know that the client has unique access to that inventory. Certainly we’re in a position where we can lower the price to react to a sales promotion, or some other branded sales event that a competitor might be holding at a particular time. And I should also add that price is not the only signal. In order to really effectively manage price in the context of the retailer’s overall business, we need to understand the shipping cost, the speed of the shipment to the doorstep, even the competitor’s reputation online determines how aggressively the software sets a new price point within that hour.

Steven Cherry: So what about the user experience? You said that users are getting a little bit more used to it, but I think that if prices are continually fluctuating, and I don’t know, if I see the camera that I just bought for [US] $100 listed at the same place, or even somewhere else, for $85 a few minutes later, I may not want to shop at that place again.

Eric Best: Yeah, I think that’s a risk, and I think it’s certainly a short-term risk, given consumers’ legacy expectations about how prices are set and how consistent they may be from one moment to the next. But I think if you look at travel and tickets, you know, some of these categories that we’ve already discussed briefly, you can see a precedent for how the frequency of price changes has accelerated over time, and really, the consumer market has come around to the idea that those prices are going to be set dynamically in real time. We also see a number of interesting business models emerging to address some of this uncertainty, if you will. In particular, there’s another firm here in Seattle called Decide.com, and they are a consumer-facing shopping engine, and they actually have recently announced that they will guarantee that the price point that they will recommend that a consumer buy is the lowest price in a certain time frame. I think it’s two weeks today, and they will actually compensate a consumer for the difference if the price drops further.

Steven Cherry: Online shopping has definitely hurt brick-and-mortar stores. Is dynamic pricing going to have yet a further impact on them?

Eric Best: There’s no doubt that when a price is set dynamically based on real-time market signals that you’re going to see margin compression somewhere. I think what this does, ultimately, is it serves to, one, make pricing more perfect. You know, in the free-market, free-enterprise sense, consumers have more-accurate data on where they can find the market-optimized price for a particular consumer product. And I think to the extent that brick-and-mortar businesses are saddled with higher operating expense, whether that’s real estate, or whether it’s inventory and cost to get that inventory into the hands of the consumer, they may be at a competitive disadvantage. But even in brick-and-mortar, and, again, I think it’s important to remind your listeners, your audience, that 92 percent of all retail transactions still occur off-line—although more than half of those off-line purchases are now in some way influenced by online research in advance, particularly when you’re dealing with considered purchases for higher-ticket items like consumer electronics and so on. But the point is that we see retailers, brick-and-mortar retailers, reacting to the more-aggressive and lower-margin marketplace by changing their business models as well. So we see smaller-footprint stores, we see a focus on value-added services, you know, classic example there would be Best Buy adding Geek Squad services alongside the consumer electronics products that they’re selling. And I would also call out entertainment value as something that is hard to reproduce if you don’t have a physical in-person connection with the consumer, and we’ve seen that with Disney stores, where more and more of the in-store real estate is being converted to a stage. And so the experience when you walk into a Disney store in the mall is as much about engaging with Disney characters, and then buying items which are not necessarily stocked in the store, but which are fulfilled later from a central warehouse, as part of that Disney shopping experience.

Steven Cherry: You mentioned before prices going up as well as down, and, you know, advertisers have known for a long time, based on, say, a home address, whether to try to sell somebody a Lexus or a Volvo or a Hyundai. What about different prices for the same product? On the Web a lot more is known about you, including your home address. Why not try to charge the Lexus owner more than the Hyundai owner for a camera or something?

Eric Best: Yeah, you know, it’s not a new concept. I think behavioral targeting, certainly from an advertising optimization standpoint, is very commonplace today. You may have already experienced display ads from across the Web following you after you’ve placed an item in a shopping cart and then left an e-commerce site without buying. And that concept of being able to track you as an anonymous shopper, generally speaking these targeting companies are not collecting personally identifiable information about you. They just know that you visited Best Buy, you added a television to the shopping cart, but then you did not purchase that television. And when they see you show up later on other sites across the Web, they’re going to display an ad for that same product and potentially even a promotional offer to try to convince you to come back and finish your purchase.

In the same way, there’s nothing really that prevents retailers from modifying the price point that they show in the cart, or on the website, or in an ad across the Web, except trust, right? And at some point I think there is a distinction—it may be subtle or it may be profound—between setting the price based on real-time market signals versus setting a price that targets a specific shopper. I think we’re headed in that direction. I think what we’re going to have to resolve in order for that to be commonplace are some, you know, perceived or real serious concerns with consumer privacy, and also the possibility that retailers are going to be viewed as targeting specific, you know, consumer demographic profiles, you know, higher net worth, or even targeting by race or color or creed, and I’m not sure that that is socially acceptable at this time.

Steven Cherry: That may not even be legally acceptable.

Eric Best: And, yeah, certainly you run into all sorts of interesting legal and policy issues when you get to that point. In terms of setting prices for specific constituencies.

Steven Cherry: Very good. Well, it’s pretty amazing how the Internet and big data is changing just everything. I think we’ve probably had about 10 shows this year about that, one way or another, and I didn’t really expect my local mall to be one of them, but it turns out it is. So thanks for letting us know about it, and thanks for taking the time today.

Eric Best: Yeah, thanks again, Steven.

Steven Cherry: We’ve been speaking with Eric Best, CEO of the software company Mercent, about how dynamic pricing is coming to a retailer near you.

For IEEE Spectrum’s “Techwise Conversations,” I’m Steven Cherry.

This interview was recorded 19 September 2012.

Segment producer: Barbara Finkelstein; audio engineer: Francesco Ferorelli

Read more “Techwise Conversations” or follow us on Twitter.

NOTE: Transcripts are created for the convenience of our readers and listeners and may not perfectly match their associated interviews and narratives. The authoritative record of IEEE Spectrum’s audio programming is the audio version.