Wednesday News: Yelp ratings are affecting restaurants; Bruce Willis Apple Suit Fake; Sex Euphemisms
How online reviews are crucial to a restaurant’s takings – There have been no scientific studies determining whether negative reviews impact book sales except for the Stanford Graduate School study issued in 2011.
Looking at 240 fiction book titles reviewed by the New York Times, investigators found that positive reviews, not surprisingly, always increased sales by anywhere from 32 to 52%. For books by established authors, negative reviews, also not surprisingly, led to a 15% decrease in sales. For books by relatively unknown authors, however, negative publicity had the opposite effect, increasing sales by a significant 45%. Follow-up studies affirmed the reason: Even bad reviews drew attention to works that otherwise would have gone unnoted. Moreover, the “negative” impression bad reviews created seemed to diminish over time.
Economists at UC Berkley have published a study in the Economic Journal that suggested Yelp ratings are having a significant impact on restaurants.In previous articles about paid for reviews, there is some intimation that restaurants and other places are buying Yelp reviews and this might be why:
They found that a restaurant with a rating improved by just half a star – on a scale of 1 to 5 – was much more likely to be full at peak dining times. Indeed, an extra half-star rating caused a restaurant’s 7pm bookings to sell out on from 30% to 49% of the evenings it was open for business. The Observer
Bruce Willis Isn’t Suing Apple Over iTunes Music Ownership Rights – There was a story that circulated yesterday that Bruce Willis was considering a suit against Apple to determine the true ownership of iTunes purchased music. It was a great companion piece to my article on Sunday (and the ReDigi case I’ll review next Sunday). Alas, this Bruce Willis story is a hoax. TechCrunch
17 Euphemisms for Sex From the 1800s – Mental Floss has collected (as the title says) 17 euphemisms for sex. I particularly liked Basket making and clicket. You? Mental Floss
Tesco Buys E-bookseller Mobcast For $7.2 million As It Squares Up To Amazon And B&N In The UK – B&N’s Nook is scheduled to launch in the UK next month. Amazon and Kobo is already there but the digital market in the UK is still in its early stages and so retail giant Tesco (a discounter like WalMart or Target) has purchased Mobcast, a digital bookseller. Based on previous moves, it looks like Tesco wants to position itself as a competitor to Apple and Amazon: The deal comes on the heels of Tesco buying movie and TV streaming service blinkbox in 2011 and Internet radio service WE7 in June 2012. TechCrunch
It’s a sad sad day when Adelaide Now can’t be trusted as a reliable news source. Oh, wait… Sorry for the bum steer. :)
On the restaurants, (as my statistics professor would write on the board) “correlation is not causation.” Possibly the restaurants have both better ratings and more customers because the food & service are better?
@romsfuulynn: The researchers admit that there is a problem withe cause/effect but said that
As with anything, more studies need to be done before any concrete conclusions can be drawn and there isn’t any evidence that the same affect is being extrapolated into other product markets.
I took a look at the Yelp article. They’re not analyzing whether the ratings are accurate (as to quality) but rather whether a star variation that signals a quality difference is effective in driving more/less customers. In the study they separate restaurants with other measures of quality (Zagat, Michelin stars, etc.) from those without and find that Yelp reviews have less effect for restaurants with the additional information.
There are some issues with the sample. San Francisco has a very high number of Yelp members and participants, but the results aren’t necessarily generalizable to cities and town with different characteristics. They don’t tell us what online reservation system they used. I’m assuming it’s OpenTable, which is (a) not used by all restaurants (which they recognize); and (b) restaurants differ in how many tables they allocate to OpenTable, which the study doesn’t take into account. And they’re aggregating all restaurants into one group, so everything from the corner take-away gyro house to Gary Danko is treated as equivalent.
They do look hard for evidence of restaurants gaming the system and don’t find any, which is reassuring. And the take-away message is powerful: rounding decisions have a measurable impact on availability. If we saw two restaurants with a 3.24 and a 3.26 rating each, we’d rightly assume that they are more or less the same according to Yelpers. But that’s not what we see. We see 3 and 3.5 stars, which we then treat as a non-trivial quality difference.
What a huge incentive to get all your friends to write reviews! Two 4* reviews would bump 20 reviews averaging 3.74*s to 4*s
Most new restaurants don’t have many reviews, so five reviews by family and friends, solicited with a review me! Review me! rather than pay could make a big difference.
Obviously, with rewards like these to be had, reviews will be gamed.
That’s exactly what the researchers thought (they are economists, after all). They performed a number of different tests on the data to see if they could find evidence of such gaming, and nothing reached statistical significance. However, the tests were performed only on those restaurants that were also in the online reservation database. It’s entirely possible that the ones not in the reservation database behave differently, but we’d want to have a theory for why that would be so.
A local sushi place I really liked recently turned me off by putting requests for yelp reviews on all their tables. Any customer who reviews the restaurant on yelp (and/or a couple of other sites) and brings the review in to show the restaurant staff gets a free sushi roll. They did not insist on positive reviews but it’s hard to believe anyone would bring a negative review in to show the manager.
@Sunita
Producing a lot of fake reviews is time consuming and probably visible. But if you are at 3.24 (3*) and you write one 4* review, you bump your rating to 3.26 (3.5*) and it will probably be undetectable.
I’d bet a restaurant that isn’t a part of an on-line reservation system might:
1) be smaller, thus get fewer reviews allowing a fake to have more weight in the score
2) have customers who aren’t as computer literate and less likely to do an on-line review, also allowing fakes to have more weight.
On the other hand, if a rating sets expectations too high, then there might be a backlash. I was at a hotel in Prague that had been on Trip Advisor as a top rated hotel for the city the year before. It was a great 3.5 star hotel, but when people started expecting the best in the city, it didn’t measure up and they started getting ratings about all the ways it didn’t meet expectations. Of course, in gaming your own ratings, dissing the competition helps.
@Sao: Interestingly, neither one of those conditions is likely to hold for their sample, which is San Francisco restaurants (although I agree it might well be true in other locations). OpenTable is a costly choice for restaurants because of the cut it takes, and restaurants at every part of the spectrum choose to abstain from it. I’m sure more at the small or inexpensive end do, but it’s not exclusively low-end restaurants. And given the population of SF, there are plenty of computer-savvy people who are constantly seeking out new, hole-in-the-wall and family restaurants. We have a favorite dumpling restaurant that doesn’t take reservations and looks like a regular Chinese takeaway from the outside but has a clientele of young IT workers and families. I found it because of Yelp. Because young people eat out so much, fakes may be more likely to be discovered and the ratings adjusted to reflect that.
The TripAdvisor ratings are a source of endless entertainment to me. I’ve become fairly good at spotting the sock puppet reviews for the hotels in places I frequent.
@Janine:
That’s interesting that they do that because under the terms of a Yelp Business Account, you are NOT allowed to do that.