See That Review? There’s (Almost) A Way To Know If It’s Fake

Something that started as both a great marketing tool and a great customer service tool has, over time, become a nuisance. Online reviews, a cool idea and a great way to crowd-source decisions, feel less and less useful.

Early this year I told you about a bad experience I had eating brunch at Penelope, one of those place-of-the-moment restaurants in New York City (I’ve since returned to Penelope, by the way; nothing’s changed).

In that piece, I pointed out that trying to ferret out “good” places by looking at on-line reviews was made difficult because statistically almost everyone ends up with a “star rating” in the 3.5 to 4 (out of five) range. But what’s making them equally bad is that reviews are now being bought.

The Yelps of the world, regardless of what you think of their business practices, have a big problem on their hands. Oh sure, their traffic keeps growing and sites like Yelp keep adding services, but the sad truth is that unless they can deliver useful content, eventually people will stop using sites like Yelp, TripAdvisor, and other review repositories.

Now, some researchers at Cornell University have come up with an algorithm to ferret out fake reviews.

That link is to the actual research paper on the subject, and it’s pretty dense stuff. And, because the nature of the subject is that the fake-ness of reviews just isn’t that easy to discern in real-time, you need to ask yourself whether there’s really any solution to this problem. But I’m struck by a couple of things.

First, the idea that too many fake good reviews are a problem is a stark counter-point to the problems that arise when there are too many bad reviews. OK, so maybe there’s no contrast at all; the problem is the same: reviews need to be REAL, or they’re useless (or will be used in some sort of statistics-bending kind of way that makes them useless).

Second, the question of whether it’s appropriate to solicit reviews and under what circumstances gets in the way of this conversation. Sure, we’d all agree that paying people to write good reviews about you or bad reviews about your competition is “dirty”, but what about when you ask customers for their opinions? That starts out sounding like an attempt to provide customer service by eliciting and (hopefully) responding to feedback, but  even this cleaner form of review solicitation can be—and is—used for less-clean purposes.

Third is this: The nature of Search Engine Optimization is such that the better SEO Consultants get the harder it will be to find the “right” content. I’m up to my eyeballs in that contest, so I’ll use a cliche to describe the situation: “Don’t blame the player, blame the game”.

I make that third point because in searching for the phrase “Cornell’s Fake Review Detector Is A+++ Would Use Again ” I came up with reference after reference to this article in The New York Times. The article is less than three weeks old, and it’s been so heavily cross-referenced all over the Internet that it’s likely to be the #1 result for people looking to research this topic … forever.

Refer back to the fifth paragraph in this piece for the link to the actual research paper. See what I mean? An article bemoaning a problem has become part of the problem.

In this case it wasn’t intentional; it was just sloppy. All the people who write about the Times article referred to it and linked it in a way that buried the actual point of the story. That’s right; their didn’t-try-to-do-search-engine-optimization references helped the Times, but hurt the actual subject.

I didn’t intend for this piece to be an advertisement for our SEO Services, but … there you go. On the Internet, nobody can hear you scream … unless your SEO Consultant wants them to. And that’s business change.