In 2006, the average redemption rate for the Sunday Newspaper Free Standing Inserts was reported to be less than 1% at 0.065%, and declining 30% from the previous year.
Based on these numbers, PBT's SmartShop looks destined to completely reinvent the loyalty industry
Here's the article:
By Jesse Quatse (Research Director, Pay By Touch)
Published by The Wise Marketer in September 2007.
Jesse Quatse of Pay By Touch updates us (thewisemarketer.com) on the 'statistical inference' technology behind SmartShop, and how grocers are benefiting from the new kiosk marketing system...
In August 2005, a Wise Marketer feature article summarized the technology of One-to-One Marketing and introduced an entirely new approach referred to as "Statistical Inference." Now, two years later, the new technology has completed productization, passed a 6-month beta-site testing, and is 6 months into initial commercial use at several well known grocery superstores.
The results are remarkable. The retailer's redemption rates are much higher than anticipated and the CPG (Consumer Packaged Goods) industry is taking notice. This article briefly reports on the impressive results and casts light on the technological future of everybody's retail superstore.
Originally named "LoyaltySuite," the product is now known as "SmartShop," and is provided through a service by the advanced retail technology company, Pay By Touch. The purpose of the product is to target personalized promotions and discount offers to the individual customers of bricks and mortar retail stores, as contrasted with the targeting of personalized recommendations by eCommerce retailers such as Amazon.com, Netflix, eBay, and others.
The so called "recommender systems" of eCommerce do not work well on the greater part of the bricks and mortar retail market because of fundamental differences in the purchased items themselves. As explained in the previous article, "Several of the differences between e-commerce and bricks-and-mortar retailing can be exploited technologically to the significant benefit of the retailer or CPG."
The question is, What are we waiting for? The overwhelming ratio of market sizes (2006: by some estimates, $634B for bricks and mortar groceries alone vs. $130B for all of eCommerce together) suggests that the demand is already there, waiting for the right technology to come along. The initial technological approaches to bricks and mortar retail, principally CRM and "rule-based" targeting, have not met with wide scale success. They can impose a burdensome task on the user who must individually judge which offer should go to what kind of a person. People are notoriously different from each other and not so easy to lump into the hundreds of precisely defined categories that are necessary in order to target offers with a high degree of accuracy.
The innovation of the SmartShop technology is in the automation of the process. It is based upon the purchasing habits and patterns of the individual customer rather than upon the human judgment of the retailer. The predictive task is shifted from the human user to the product mathematics.
Launching the Technology
As of the date of this article, four major grocery superstores are in various stages of pilot trials for SmartShop, five more chains are starting, and full chain roll-outs to several thousand stores sites are anticipated through the next year.
Shoppers report that the offers are all interesting and relevant to their own tastes and choices. Therefore they continue to use the service in increasing numbers. The CPG offers are presented on the same print-out sheet as the store's offers, but targeted differently. Often the CPG prefers a more predictable offer distribution than do the retailers, even when the trade-off for predictability is weaker individual relevancy.
The trade-off works because the strong relevancy of the retailer's set of offers attracts the shopper to the CPG-based offers as well. Thus the CPG can target in a more conventional and controlled manor while the store-based offers target purely automatically through mathematically emulated relevance to the individual. For the technically minded reader, the math and technology were recently published by Dr. Quatse and Najmi in the proceedings of WORLDCOMP'07, the annual international conference of computer societies.
Long Term Benefits
The long term benefit of personalized marketing is customer loyalty; that is to say, the lifetime revenue value of the customer. The term, "loyalty," has been defined by a plethora of books and articles, some of which object to the term because "customer loyalty" is not the same as "family loyalty" or "baseball fan loyalty." Ignoring the other uses of the word, for many retailers "customer loyalty" simply means keeping the customers they have and increasing the spending of each. Since the grocery retailer cannot induce the household to eat more, the only way to extract higher revenues from existing customers is to increase the number of visits to the retailer's store at the expense of visits to competing stores. Personalized marketing can entice the customer to stop at store A over store B for the sake of even one interesting discount offer, all else being equal. Presumably, while at store A, the customer then purchases the rest of the basket that would have been purchased at store B. The long term benefit is in attracting a larger portion of the household budget through more frequent visits.
The personalized offers impart two loyalty reinforcements: one obvious and the other subtle. The obvious reinforcement is the cash reward to high value shoppers in a visible, personal, way: through discounts on items of interest to them. More subtly, the personal relationship itself has its own powerful long term effect on the shopper, even without the cash reward. The dawning significance of Personalized Marketing is recognized by the Eisenberg brothers in their recent book, Waiting for Your Cat to Bark?. Loosely translating the title, the brothers are trying to avoid waiting for the benefits of a marketing program that is doomed from the start. It can't work because it isn't relevant to the customers. They observe that "When someone acknowledges us as individuals and personalizes our experience based upon our unique characteristics, we feel understood and valued." They conclude that, "Personalization casts a powerful spell. Marketers understand this." The long term benefits to the retailer and the CPG are clear. Loyalty is enhanced by personalized offers.
Short Term Indicators
Yet, if the product effectiveness is to be evaluated, some form of immediate feedback is essential. Not many can wait to see if the "cat will bark." There are two immediate indicators of the long term effectiveness: redemption rate and lift. The redemption rate is a direct measure of how relevant the offers are to the customer. Redemption is defined as the percentage of print-outs for which any discount has been used by the shopper. (A specific definition of "redemption rate" is needed for personalized electronic marketing because the situation is unlike any other.) The automatic nature of the process requires a customer identification by loyalty card or finger touch recognition, followed by a print-out of the personalized recommendations for that identified customer. That process is called "activation" and all else is handled automatically at checkout. This definition is hard to compare to paper based discounts but the numbers are still enlightening. In 2006, the average redemption rate for the Sunday Newspaper Free Standing Inserts was reported to be 0.065%, and declining 30% from the previous year.
The SmartShop average redemption rates for all superstores is currently above 40% every week and rising.
One surprising conclusion is that the redemption average is so high even though the discounts are not particularly high (averaging around 70¢ - 80¢) and the items offered are not exceptionally popular (by intent). The most important conclusion is that the consumer does indeed appreciate personalization at the extreme level where each image is truly relevant to the individual recipient. The key word is "relevance."
The lift is a measure of how much retail revenue boost is obtained through Personalized Marketing.
Lift can be measured by sales figures, for example by directly comparing sales of a product with and without an offer redemption. However, measuring lift in that way has its problems. The markdown cost of increasing the revenue is obscured and the effect of a redemption on the total basket value is not clear. Lift can be more revealing when measured in terms of increased number of visits. The long term benefit of increasing visits is evident regardless of other purchasing factors. Chris Cubba, Analytics Manager at Pay By Touch, is monitoring the combined lift at three superstores of a more recent commercial client.
He reports that shoppers who activated SmartShop offers visited the store 12.3% more often than those who did not. One interesting conclusion is that personalization in and of itself increases visits. Customers who choose to stop at the SmartShop kiosk appear to visit the store 12.3% more often then those who don't. They won't stop at the kiosk unless they expect the print-out to be personally relevant. So lift too is dependent upon the key word, "relevance."
Lift and redemption are not the whole story. Neither require personalization. Lift on everything is easy to obtain by simply dropping the prices of everything by 20% or more. It's called Wal-Mart. Redemption can be maximized just as easily by always offering the top selling items at discounted prices. That too is called Wal-Mart.
Personalized Marketing requires individually preselected discounts on only the relatively few items that are personally important to the individual buyer. As Jim Stengel, Procter & Gamble's global marketing officer, described customers in Fortune magazine, "They want to be understood, they want to be respected, they want to be listened to." Personalized Marketing is part of that understanding and respect, the part that can distinguish one customer from another.
That distinction cannot be made by simply recommending whatever each individual purchases most often. Most of the frequently purchased items of any one individual are similar to those of any other. They correspond to the "hits" in movies, music, groceries, or in any retail industry, because the most popular products make up the bulk of the sales. (Grocery examples are bananas, milk, and chicken.) The so called "80/20 rule," or "Pareto Distribution," pops up everywhere. That is to say, about 20% of the items usually make up about 80% of the sales. The 20% form the "head" of the demand curve. Personalization cannot be based upon the head of the curve alone because hits are hits for the reason that they are in fact purchased by most customers. So hits tend to lump customers together in a common preference. By contrast, the items which are not purchased in common must show the individual differences between the customers. Our individuality resides in the tail of the demand curve, not the head.
Relevance of the Long Tail
In his significant book The Long Tail, Chris Anderson points out that the rules of marketing have been radically and permanently changed. For example, he points out that Wal-Mart can sell no more than 60,000 different CD tracks profitably, but Rhapsody can sell 900,000 CD tracks, all of which are profitable. His point is that when given the opportunity, we each make our own different choices in the long tail. Personalization requires that the technology "flatten" the demand curve so that uncommon purchases are given greater weight than the common ones. That is done in SmartShop by ranking offers according to how much more the individual customer purchases the item, compared to the average purchasing. To receive an offer for bananas, for example, the customer must consume many more bananas than the average customer of the store. In those terms, the long tail is the basis of Personalized, One-to-One, Marketing in general, not just SmartShop in particular.
The Personalized Future
The question is not "Why Personalized Marketing," but "When." In the flood of information engulfing us all today, Personalized Marketing technology acts as a filter. The technological goal is to remove the irrelevant information from that flow and to admit only what is relevant to each of us individually. The filter is imperfect, but improving quickly. Without them, we are shouting our message at the noisy crowd. We are waiting for our cat to bark. (Ed. Note: Woof!)
- Jesse Quatse. "The technology of one-to-one marketing." The Wise Marketer, August 2005.
- Jesse Quatse and Amir Najmi. "Empirical Bayesian Targeting." Proceedings of the 2007 International Congress on Machine Learning; Models, Technologies & Applications, WORLDCOMP'07 June 2007, pp 153-159.
- Bryan & Jeffrey Eisenberg. Waiting for Your Cat to Bark?. Nelson Publishers, 2006.
- Geoff Colvin. "Selling P&G." Fortune Sept 17, 2007, pp 163-169.
- Chris Anderson, The Long Tail, Hyperion, NY, 2006, p22.