Skip to main content

Quality Uncertainty On-Line: Yelp Reviews—Everyone’s a Critic

Yelp is one of the most intriguing online review sites I’ve experienced during my time in the service industry. What sets it apart from many online reviews available to the general public (ie, reviews within a business’s self-managed website) is that business owners, once registered, are unable to deactivate their registry or “start over.” This is also what sets Yelp apart from other reputation systems. With eBay, for example, a seller could register, generate “trustworthiness” within the market via their profile and interactions with other buyers, and then destroy this good reputation by engaging in shoddy business (ie, selling a “lemon” or another product that is significantly less nice than what was initially promised or advertised).

The restaurant business, in these ways, is exactly like eBay. There is nothing stopping business owners from creating their own glowing reviews via a fake profile or soliciting others to do so. Additionally, the market for lemons is an example of the effect had on businesses by poor Yelp reviews: let us hypothetically say that a couple going out to dinner in New York City for their anniversary have the good sense to do a little research on their dining options. When they search the query, “restaurants new york city,” they may come across several Yelp pages. If this couple deduces from this information that the chance of receiving bad service is too high, then their reservation price (defined as the maximum price they will be willing to pay) will be extremely low (meaning they wouldn’t even be willing to pay, for example, $1 for a meal from a badly reviewed restaurant). In this case, the restaurant’s market could fail completely, all due to a poor review.

Another analogy that correlates directly to the service industry is the labor market in the context of the market for lemons. We will assume that in these examples, the “workers” are the restaurants and the “companies” are the prospective customers (it is easier to think of it in terms of: who do people decide to “hire” for the night to provide food for them?).

According to the textbook, the market for lemons has four basic assumptions:

“(i) There are different qualities of workers — some are very productive while others are less productive, and this affects the value they will produce for the company that hires them.

In this assumption, one can liken the quality of workers to the quality of the restaurants. Because there are varying qualities within restaurants, peoples’ reservation prices that they will pay for a meal at a specific business also vary (ie, how much would you pay for a hamburger at McDonalds vs. a crème brulée in Paris’s fanciest, most expensive café?), just as some employers would be willing to pay a more highly skilled worker a larger salary than a less-skilled worker.

“(ii) It is natural to consider a setting where there are different kinds of jobs at different levels of wages, and where companies would be willing to hire any given prospective employee if they could accurately determine which jobs and wage levels were appropriate for them.

This assumption’s analogy to the service industry is very similar to the previous one.

“(iii) There is asymmetric information: a person generally has a better sense for how productive they are than a prospective employer does.

Throughout the process of getting hired, the only person aware of their intrinsic capability for a certain employment is the interviewee. It is the job of the interviewer to deduce this information and make an informed decision based off these deductions. However, an interviewer has no way of knowing for sure other than their assumptions made based off their collected data.

“(iv) If we take a strong but plausible version of (iii), where employers can’t reliably determine the quality of the people they are hiring, then employers can’t hire only productive workers and wages can’t directly depend on the quality of the person being hired. Rather, a uniform wage will be offered, and only applicants who believe this wage acceptably values their skills will take the job.

There is inconsistency in the analogy with this point, as it is not the restaurants (the potential employees) who are choosing their customers, but the customers who are choosing them. Still, the similarity may be drawn with customers who go out with the knowledge that they only want to spend, for example, $100 maximum on dinner for the evening. Thus, they may “offer this wage” and only restaurants able to accommodate this precondition would be in the running for their selection process.

Another component of Yelp that is worth considering is the prominency with which the reviews are displayed on a search engine. Review websites lead to vast amounts of links to other webpages, making these pages seemingly more well-connected than others, which, as we’ve studied, leads to them being featured higher on the list of search results. The accumulation of reviews results in an increased likeliness that a business’s Yelp page will be one of the first impressions researchers see, like our New York City couple. In the case of a mediocre or poor restaurant, the cycle occurring as a result of reputation systems is vicious. When customers’ opinions are processed via an algorithm, one can anticipate many potential customers will shape their dinner plans to benefit themselves. As Professor Kleinberg quotes in our textbook, “Designing reputation systems that are robust in the presence of these kinds of difficulties is an ongoing research question.”

I found the source insightful, as it shed light onto the “evidence” Yelp displays to the public on the business’s Yelp page when it has found evidence of a business attempting to manipulate their review history (ie, a craigslist ad soliciting good reviews). I found this not only amusing but good marketing on Yelp’s part, as it shows their in-depth attempts to demonstrate to their clientele the lengths they will go to in order to provide the most thorough investigation possible on behalf of the consumer, and how devastating this exposure can be to businesses trying to “illegally” pad their online reputation.


Leave a Reply

Blogging Calendar

November 2017
« Oct   Dec »