Imagine you had a machine that could predict with perfect accuracy every grocery item your customer was going to buy next week. It would tell you about the bananas and the yogurt, the pulp-free orange juice and the organic granola bars. It would tell you about the favourite hot sauce and the detergent that was running low and the cocoa for the birthday cake. It would not tell you where these things would be purchased, mind you. It might be with you or it might be with your competitor. Or, most likely, a mix. But you could rest assured the customer would purchase each of these things at some point in the next seven days. If you had such a machine, would you know what to do with it?

This machine is a fantasy, of course. Needs change. Surprises happen. Spur of the moment decisions are made. But the truth is that with advances in big data and the power of predictive analytics, many retailers – especially high frequency retailers like drug and grocery stores – can predict what customers will buy with a surprising degree of accuracy.

But knowing what a customer is likely to do is not enough. The real secret is in knowing how to leverage that information. Faced with the question of how to use retail analytics for relationship marketing, the retailers we talk to tend to default to one of two basic approaches. The first is to send customers offers on the things they buy most. Imagine a list of every item the store sells, ranked for an individual customer from the item she is most likely to buy down to the product she is least likely to buy. This approach says we should give her discounts on the items at the top of the list. This will build loyalty, drive trips to the store, and help retain her as a customer for the long term.

The second approach is the opposite. We should give her offers on the things she doesn’t buy…the things at the bottom of the list. This is where the incremental sales come from. After all, the thinking goes, why would we give her a reward to do what she was going to do anyway?

Relative Relevance

We want to suggest a third approach. In a previous post, we examined how retailers can leverage the long tail to give customers personalized offers on items that may not have high sales volume overall, but are important to that individual customer. We shared the surprising statistic that 90% of products are purchased by 1% or less of customers, yet together these products make up 58% of total sales.

In this post we want to expand that idea by sharing a tactic for pinpointing exactly which products in the long tail make for the best personalized offers. It is simple but powerful concept called relative relevance. It works like this: Instead of asking the question, “What product is this customer most likely to buy?”, we ask, “What product is this customer most likely to buy relative to other customers?” The chance a customer will buy a product compared to the chance an average customer will buy a product gives a relevance index that is extremely helpful in identifying the best offers.

In the example below, we see a customer’s purchase prediction score for two items. Looking at the scores in isolation, you would say that the detergent on the left (38% chance of purchasing) is the most relevant offer. But if you asked Jennifer, an offer for the protein powder on the right would actually feel much more relevant. She is seven times as likely as the average customer to buy it – it is likely something she uses, cares about, and identifies with.

The second product sits in the long tail. But so do thousands of other products. The relative relevance approach allows you to quickly identify which long tail products will be meaningful to customers. When trying to choose offers that will motivate a customer to shop at your store, this is the sweet spot.

Why it Works

There are many reasons that focusing on highly relevant long tail products is an effective strategy. Here are three of the key ones:

1. The offer feels more personalized

It is no secret that personalization in retail is here to stay. According to an Accenture study, 91% of consumers are more likely to shop with brands that provide them relevant offers and recommendations. But there is a difference between an offer that is relevant and an offer that feels relevant. The detergent in the example above is a relevant offer for Jennifer. She buys it and uses it. But it is also a very common product. It is the kind of product that is regularly featured in the flyer or in-store display. Everybody buys it, and Jennifer knows it. It is the kind of product a retailer would recommend to Jennifer whether they knew anything about her or not. The protein powder, on the other hand, is unmistakably personalized. This builds an emotional connection with the customer in the way an offer for a popular mainstream brand simply cannot.

2. The offer has greater perceived value

Precisely because niche items have lower volumes, they are less likely to be prominently featured in mass promotions. Flyer pages and end caps are reserved for products that are going to drive traffic and build baskets. Specialty products don’t have that power. All this means that customers who do buy the products know that it is rarely, if ever, on sale. If soda is featured at a discount on the front page this week, and a customer doesn’t take advantage, she knows it will come around again in a few weeks. But a discount on a niche item she values is a rare opportunity, and therefore more likely to influence a purchase decision.

 3. Niche items tend to have higher margins

Finally, we must note that the niche items we are talking about tend to be more economically advantageous for retailers. They are often needed or strongly desired by their customers for health, diet, religious, or other reasons. Their prices have not been driven to the bottom by fierce competition. Which means the margins are higher. Clearly, offers on niche items create a win for both retailers and customers.