How effectively does Groupon leverage its size?

By Felix Salmon
June 4, 2011
David Sinsky has some extremely smart Groupon analysis over at the Yipit blog, using numbers from the company's S-1 to throw into question just how good Groupon is at making ever-increasing amounts of money once it has entered a market.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

David Sinsky has some extremely smart Groupon analysis over at the Yipit blog, using numbers from the company’s S-1 to throw into question just how good Groupon is at making ever-increasing amounts of money once it has entered a market.

That is, after all, the explicit rationale behind Groupon’s ever-increasing losses:

We spend a lot of money acquiring new subscribers because we can measure the return and believe in the long-term value of the marketplace we’re creating. In the past, we’ve made investments in growth that turned a healthy forecasted quarterly profit into a sizable loss. When we see opportunities to invest in long-term growth, expect that we will pursue them regardless of certain short-term consequences.

Groupon breaks out granular details for four cities in its S-1 — Chicago, its home; Boston, its second-oldest US market; Berlin; and London. The first is unique in many ways, while the last two were acquired when Groupon bought CityDeal, so Sinsky concentrates on Boston. And finds this:

Rev-per-Sub3.png Rev-per-Cust1.png

These charts show quarterly revenue per subscriber, on the left — that’s people getting Groupon’s emails — and per customer, on the right, which is people who have actually bought coupons. Both are going down, which is worrisome.

The reason to worry about these trends is that as many people including Joshua Gans have said, the daily-deal space has very low barriers to entry and is highly competitive. As such, if Groupon has a “moat” — a comparative advantage over its competitors which is very difficult to overcome — it’s its sheer size.

Size confers many advantages, not least targeting. As Kaiser Fung notes, the more new customers and fewer existing customers that a Groupon reaches, the more profitable it’s likely to be for the merchant concerned. The best Groupons target a whole new audience for the merchant in question: the S-1 gives the example of a cruise line which was doing very well with its murder-mystery cruises but much less well with its romantic-dinner cruises. By offering a Groupon valid only for the romantic-dinner cruises, the company managed to broaden its customer base without losing revenues from its existing customers.

Benjamin Edelman says that for a Groupon to be highly profitable, it should be targeted at people who are “particularly unfamiliar with a participating merchant’s services.” And it stands to reason that Groupon, with the largest subscriber base and the most sophisticated targeting technology, should be better able to target Groupons at relatively narrow classes of people than any of its subscribers.

At the same time, because the best Groupons are aimed at people who are likely to become regular customers of the merchant in question and who are well-disposed towards trying it out, sophisticated targeting should increase Groupon’s conversion rate substantially. Rather than just send the same deal out to everybody in a city, Groupon should be able to show you only those offers you’re most likely to want, and thereby increase its conversion rate.

Sinsky says that doesn’t seem to have happened in practice: “As the average purchases per customer continues to decline,” he writes, “so will overall conversion rates on personalized deals.” The trend isn’t good: Groupon’s Boston customers are less engaged, and less profitable — even as Groupon’s costs for acquiring new customers continue to rise.

That said, Sinsky doesn’t think that what we’re seeing here is a failure of Groupon’s targeting per se. Instead, we’re seeing the effects of competition. Which is where Yipit’s own data comes in:

A year ago, according to Yipit data, there were 9 daily deal services in Boston offering 15 active deals. Today, Yipit Boston shows 23 separate services offering daily deals including new successful entrants like TravelZoo and Yelp. The 23 services are responsible for creating 91 active daily deals. Worse for Groupon, there’s no sign of this ending with Google and Facebook on the horizon. Plus, successful entrants like TravelZoo are still only running two deals a week.

Realistically, what we’re seeing is improvements in Groupon’s targeting technology failing to compensate for the rise in competition (which also increases Groupon’s customer-acquisition costs.) Anecdotally, targeting technology still has a long way to go: for one thing, Groupon still isn’t remotely local enough in cities where people spend most of their money within a mile or so of where they live, and where they often get offers for merchants on the other side of town. But it’s not even clear how Groupon is going to get the kind of highly granular data about its subscribers’ income, and profession, and exact location, and other things which would help it target offers. That’s where the likes of Facebook and Foursquare have a clear advantage.

So while Groupon is still the biggest of the deal sites, the jury’s still out on the degree to which its size is going to prove an unassailable advantage over the long term. In principle, a massive subscriber base could be a huge competitive advantage. But in practice, this could be one area where Groupon finds it hard to execute.

14 comments

Comments are closed.