Targeted marketing: Groupon's secret sauce?

Felix Salmon summarizes the reaction to his piece on Groupon's business model, including my take posted here. He and many others believe that Groupon cannot succeed without "targeted" offers. Since I spend a lot of my work day thinking about targeted marketing, let me first talk about how targeting generates value, and then whether this makes sense for Groupon.

***

Let's start with Felix's earlier example of a $15 Groupon for $30 value at a local restaurant Giorgio's. The restaurant can choose to offer this deal to everyone: this reminds me of a certain small retail chain in Manhattan that advertises "up to 90% off, everything" 365 days of the year. That could only work if the retailer jacks up its list price and prey on unknowledgeable customers, or in an industry in which the profit margins are insane.

Alternatively, the restaurant can decide it has a budget for deals, and will have to limit the number of takers. In this case, most companies will attempt to direct the offers to certain customers. Based on what?

Some companies target customers who are most likely to use the offer. The more customers use the offer, the higher the revenues. At least, that is the hope. But as I pointed out in the previous post on Groupon, more customers does not always translate into more profits. If your most loyal customers are the ones taking the deals, then the offer strategy will backfire.

Here, the economics concept of adverse selection comes into play. Deals disproportionately attract two types of people: your loyal customers who love your products, and deal-seekers who does not want to pay full price.

More sophisticated companies target customers who are most likely to increase profits. There has to be enough incremental revenues to offset any lost revenues due to offers.

***

Let's imagine that a perfect targeting technology is available. What's the perfect target for one of Giorgio's Groupon offers? This would be someone who lives in the neighborhood who has a generally positive impression of Giorgio's but has never taken a meal there. Besides, this person should be someone who eats out frequently, and likes the type of Italian fare Giorgio's serves. Further, one hopes that this person is not entirely motivated by deals, and would come back in the future for full-priced meals; or be so overwhelmed by the meal as to raise one's willingness to pay.

The point is there may not be many such people. If the restaurant is new, then yes. If someone just moves into the neighborhood, then yes. But if neither the restaurant nor the resident is new to the area, then he or she either has tried the place, or doesn't bother. 

By definition, targeting reduces the number of possible dealtakers while increasing the overall profitability of the deals. If the target universe is too small, these deals are profitable but do not generate meaningful increases in revenues; if the target universe is too large, these deals are unprofitable despite producing revenues. The art of targeting is a balancing act between those extremes.

***

Going back to the perfect target for Giorgio's. The next question is how would one find such people? Data is more readily available on some items, such as how much one spends on food and eating out, or how long one has lived at an address. (NB: Available does not mean accurate.) It's harder to know which type of cuisine one prefers, or whether one is a dealseeker. These preferences are not necessarily stable or generalizable. For example, you can be very picky when it comes to French food but much less so when eating Indian.

And don't forget Giorgio's only knows about what diners do at their restaurant. To know anything else, they need to buy data, and they need to analyze the data.

One could say that Groupon can act as the aggregator of such data and provide the intelligence for designing these deals. The challenge is that if these deals become truly local and tailor-made for merchants, Groupon would need to be experts at many businesses and know a great deal about local business conditions.

That may be possible but it is a tall order.

 

Comments

[Erik]

This comment is unrelated to your Groupon posting; I just wanted to provide this link to a story in Saturday's FT on how statistics is changing soccer. I recently read (and thoroughly enjoyed your book), and I thought that the story told via the article was similar to some of the phenomenon you highlighted in the book. Keep up the good work.

http://www.ft.com/intl/cms/s/2/9471db52-97bb-11e0-9c37-00144feab49a.html