There's a good bit of hand-wringing going on this week over the future of Groupon and Groupon-style daily deals. Groupon itself saw traffic plunge by 50 percent, with speculation that part of the drop is attributable to its surging competitors. New daily-deals sites are being launched almost as rapidly as others are going out of business. Facebook and Yelp are reportedly scaling back or discontinuing their forays into the Deals space. (Full disclosure: the Phoenix, like most newspapers these days, offers discount deals similar to Groupon and Living Social.)
Here's one bit of insight gleaned from research on the business side of the Phoenix's deals: in the case of Yelp's deals, the company's biggest enemy may have been itself.
For better or for worse -- and many would say for worse -- Yelp has become the primary arbiter of user-generated restaurant ratings. The Yelp community is fairly narrow, but it gets amplified at the macro level by inclusion in Google search results, and at the micro level by restauranteurs who obsess over their Yelp rankings. (The company itself has been accused in the past of manipulating those ratings for profit.)
As the Phoenix researched business models to improve our own deals product, we created a rough predictive algorithm to see whether we could guess -- based on a restaurant's average Yelp rating -- whether the deal would be successful. Our thought wasn't that Yelp itself drives people to buy deals -- but instead that Yelp reviews reflect, in the broad aggregate, a measure of a restaurant's popularity. We were attempting to build evidence for something that should seem pretty obvious, and that based on our research seems to hold true: Deals for good restaurants will succeed, while deals for dud restaurants will bomb. (Obvously, that's not by any means the entire story of Deals products: the size of a company's email list, for instance is a huge factor. We were merely trying to figure out, all other things being equal, whether the quality of a restaurant would trump, say, the steepness of the discount offered.)
So why would that be a problem for Yelp?
This part is speculation. But here's the argument: there's an inherent conflict on Yelp between two business models. On the one hand, they want to sell businesses enhanced profiles that will help raise their overall ratings. But the businesses with the highest ratings would be the hardest to land for a Deals product, because they typically wouldn't need to do a Deals promo. (Essentially, the best-reviewed restaurants may be becoming an expensive commodity -- the operators of Deals networks are beginning to get a sense of which businesses sell and which don't -- forming a pyramid of sorts with Yelp
the fact that Yelp has the reviews right there in front of the potential buyers
may have been what held them back. I built a prediction model based on a
weighted scale of Yelp reviews and found that it was possible to make a fairly
accurate prediction about the potential success of an offer in more cases than
not. Yelp reviewers tend to be pretty harsh, so Yelp may have been backed into a
corner where only the best-reviewed establishments were worth promoting – and
they are the hardest ones to land because they typically wouldn’t need to do a
Deals promo. My biggest takeaway from the article is just a reinforcement of
what I’ve already discovered through basic analysis – it all comes down to the
quality of the offer, which thanks to Yelp can be quantified.