The effect of Groupon promotions on Yelp ratings | Hacker News Reader

The effect of Groupon promotions on Yelp ratings

mybiasedcoin.blogspot.com

93 pointseggspurt27 comments

Interesting data and conclusions in there. And it's not really surprising given what I've learned about Groupon customers from their behavior at brick and mortar businesses owned by some friends here in NYC. They classify Groupon users as demanding+cheap, meaning that they aren't willing to pay full price for much of anything (and unlikely to return unless given a similar deal) and far more demanding (special requests, asking for free extras, etc) than their customers who pay full price. Not one of them will ever do another Groupon promotion.

From the opposite perspective, I have found that owners of business treat me as a second class customer when I use a Groupon voucher with them. There is either an automatic assumption that I'm cheap and a less worthy customer, or they have overbooked their services and provide a poor experience.

Maybe the problem is your friends became resentful because they were incapable of handling the promotion, so they blamed the customers to protect their own egos?

Not all of my Groupon experiences have been bad. I did a scuba diving experience, which lead to me doing additional scuba qualifications and experiences with the same group, at full price.

Most of my friends who use Groupon have automatically defended Groupon users and blamed the businesses without knowing any of the story. Most of my friends who use Groupon regularly are, honestly, kind of cheap (you know the ones... who always cause an issue at the end of eating out with a group of people when it's time to split the check and leave a tip).

One business I know of had no predisposed opinion of Groupon users (they aren't tech savvy and never heard of it before doing it at the suggestion of a friend) and regularly does discounts for seniors and students as well as deep "first visit" discounts for Yelp users where they ran a promotion. They even did a similar promotion for NYC businesses with a similar deep discount to Groupon. They said that the Groupon users were, without fail, the most demanding of any group of promotion-using customers and some even made it obnoxiously clear that they would never use my friend's services without a similar discount in the future. Of all the Groupon users, not one became a repeat customer. In every other promotion they have run, they have gotten numerous repeat customers, including the similarly deep discount program for NYC businesses. So, they never plan to use Groupon again and have warned their fellow local businesses away from it.

Probably an important difference would be that while Yelp users are almost by definition looking for new experiences, Groupon users are likely to be either people looking for cheap new experiences, or people looking for any cheap experiences. I mean, be honest, if price wasn't regularly a big deal to you, would you spend much time on Groupon, where the options are drastically more limited?

That is to say, the difference between running a promotional discount in a regular magazine, and running a promotional discount in a coupon magazine.

I use Groupon for experience discovery. I could spend time combing Yelp for new restaurants (and sometimes I do), but it's nice to hear about a new one through Groupon.

Perhaps, but it also means that the pool of potential new experiences is being limited to businesses that just did a Groupon. I'm sure results vary widely, but in the aggregate that probably has implications for what can be expected of both the seeker and the provider of said new experience.

That is a very good point. In the one case (Yelp) they were advertising a good 'first visit' deal to folks who were already looking for their services. In the other case (Groupon) they were advertising to a group of people looking for cheap deals, not even people who were looking for their services. It is odd that the NYC business list (can't recall the name at the moment) is also a 'deals list' type of service and a similar discount to Groupon was offered but it had much better results. Checking them out myself at the time, the NYC business list seemed to be much more centered around getting a customer in once and keeping them as a customer and seemed to cater to a different demographic than Groupon.

I think it depends on the business. A friend of mine is a Rolfer and she is very happy with the results she got.

I definitely believe this. I had a bad experience at a restaurant at which I used a Groupon simply because the food took a long time, the servers were constantly rushing around and couldn't attend to us, etc. Eavesdropping a bit, it was clear most of the customers there were using the Groupon and the place wasn't used to that kind of volume.

this. I bought a Groupon once for a small restaurant and among about 12 customers there was just a single one who ordered conventionally. The restaurant was at capacity and was really small but they sold around 8000 Groupons and when I visited a couple months have passed and it was still completely full with Grouponers. Even though it was cheap as hell (18 EUR all you can eat tapas for 2 people) I won't go back there because the food made me sick.

Usually happens during the first and the last week of the promotion. I learned my lesson, and if you time your visit anywhere in-between, you can generally get into the place on a low-volume day.

One of the comments on the blog perfectly sums up what would have been my hypothesis, so I will simply repeat it here:

    Part of the drop may be due to Groupon inducing 
    people to visit a business selling something 
    they otherwise wouldn't be interested in. 
    E.g., I might check out a Greek restaurant 
    even though I'm indifferent to Greek food because 
    of a good deal and curiosity. Given the same 
    experience as someone who likes the cuisine, 
    I might then give a lower rating.
    - Anonymous

http://mybiasedcoin.blogspot.com/2012/03/groupon-effect-on-y...

Are people reading the complete article? I take the author not to be claiming that this happens because Groupon sucks. In fact, I think only the third hypothesized reason of five for the decline is actually a problem for Groupon. The fact that Groupon users may be more critical in Yelp reviews is only an indirect problem (ie. people don't want their Yelp reviews to go down), but isn't evidence of a fundamental problem with the model.

But the article suggests a number of very plausible reasons that have nothing to do with bad experiences. Further, as they say, until the compare these effects with any influx of customers, there's absolutely no evidence that Groupon isn't actually performing better with this metric than, say, running an ad in the paper, or getting featured on Oprah.

Such a plot really needs error bars. It's possible that the difference is not statistically significant.

Error bars on what? The ratings themselves can be measured with perfect accuracy, and the results were averaged over 5000 merchants, so any error in the mean would be too small to see.

It's true that the average clearly moved, but the author was trying to make a generalized inference from the average, that there was a structural break in the "true average" insofar as the measured average is just a sample/noisy measurement of the "true average". Therefore, error bars/confidence bars would indicate how likely the structural break in measured outcomes is to reflect a structural break in the underlying "true average."

The volatility/variance comes from the fact that 1. not everyone rates their experience, 2. there's variability in the experience, 3. different people have different expectations or likelihood to rate high/low, etc.

I hope that clarified what I believe was the parent's meaning.

That break at day 0 was because they reset the moving average calculation on that date.

With error bars/lines on that plot, this discussion wouldn't have been needed. There are nice ways of doing error contours - showing the 1% quantile lines on top and on the bottom.

You're right, statistical significance is a valid concern but error bars is not the way to go. Check the paper for a rigorous analysis. We are not drawing our conclusions from a single plot.

I didn't think that you were (and having glanced at the paper now, there seems to be a lot more meat to it), but I think others will see only the plot. In my opinion, such a plot is misleading given the large variance I would expect in the rating.

Whenever a Groupon deal appears at a place that I already frequent, my heart sinks a little. I know that I can look forward to crowded seating, poorer service, and harried staff for the next three months or so.

I've thought about how to fund a reverse-Groupon: a service where I can pay places to NOT take part.

I want to fund a "GroupOff", the opposite of GroupOn.

You pay more. Sometimes a lot more. Yet you get something extraordinary in return. What would that be? It's up to the vendor.

It'd probably have to work something like Kickstarter if there were different premium perks, or just a one-shot thing. It'd give mid-range restaurants a chance (and budget) to pull out all the stops for a special VIP night, for instance.

Nothing says to your retailer "I hate you and I want you to die" more than using a GroupOn coupon.

>You pay more. Sometimes a lot more Well, strange concept. Since every experience at a restaurant is carrying a risk of not enjoyment, I guess many people would not be willing to increase their risk by paying more. On the other hand, usually becoming a regular should lead to better service.

An anecdote from my life. I never used to eat uni (sea-urchin) in my local Sushiya here in Tsukiji. However, at some stage the Sushi-chef recommended it to me free of charge since it was season and he wanted to offer me a special treat. Since I liked the taste I regularly add uni to my order. Just by giving me this one uni for free he increased his revenue from me by 10 USD.

[Edit] So the lesson's learned here should be "Know thy customer"

This pretty much coincides with my experience; when I'm with a group of friends and we are using a Groupon, we get about 1 star less worth of service. (I've never used one on my own; I'd rather pick the restaurant on my own rather than being told where to go buy whoever happened to feel like buying an ad that day.)

That's why I don't get the Groupon hype. It seems like a way for businesses to pay money to upset customers. Probably a bad idea.

I wonder if the cost-benefit balances out on this one. Obviously it varies by retailer/restaurant.

I'm not just talking about the incremental sales<->lower ratings tradeoff, either. According to the paper's abstract, running a groupon ad earns you a significant amount of new reviews - new ratings which tend to be 10% lower on average. But is it more valuable to have a lot of "very good" reviews, or a small number of "excellent" reviews?

Groupon Spams the internet with cut rate, deceptive advertising, that is killing the legitimate scuba business in Florida. http://www.scubaboard.com/forums/new-divers-those-considerin...

This is one of my favorite blogs, it's worth reading the archives for the material about CS and statistics.

Reply on news.ycombinator.com