
Lots of folks are talking about the news out of Cornell that researchers have created a program to filter out fake reviews at online sites.
It's an interesting program, no doubt.
And it is, as the researchers describe it themselves, a good first round filter. It focuses on specific language used.
Here's my human filter that works for me. And it ought to not be that difficult for a second round filter to, well, filter out.
I look at the relationships between the reviewers.
Sure, that's easier for me, since I often know the relationships. If I know that the inn owner has a brother in Alaska named Fred, for example, and Fred from Anchorage posts a review, well, yes that's pretty obvious. Especially if it was for the same time period when Fred was visiting.
But the relationships I'm talking about go beyond that. It's the relationships between the reviewers themselves. Say there's an apparently solid review from a reviewer with, oh say, 25 reviews under their belt.
Must be a solid reviewer, right? No flags there, right.
Not so fast.
Say that reviewer has posted reviews of 6 bed and breakfasts. All within a thousand miles of home.
Fairly normal, right?
Say each of those 6 bed and breakfasts has received reviews from each of the other reviewers of the first bed and breakfast. Fairly easy to see that each of the reviewers, despite their geographic differences and seeming body of solid reviews, have a clear relationship with each other.
One's from Indiana, one's from West Virginia, one's from Virginia, one's from DC. You get the picture.
What do they have in common? They're all B&B owners. They belong to the same forum. They exchange free visits with each other - and follow up each visit with a glowing review that looks terribly terribly solid.
And that's just one ruse that's sadly quite common.
Here's another I see often. In today's economic climate, more and more small inns are being bought by investors who are buying up not one but three or four small inns. Several partners involved. Each partner then writes up a glowing review of each of the small inns they now own. Suddenly their inns move from bottom of the review list to top of the list, almost overnight.
But the reviews look solid, right? They're not one hit wonders, for sure.
And if you didn't know that the inn in, say Pacific Grove, was owned by the same set of investors as the inns in, say Elk, well, who would know the difference? Certainly not the unsuspecting guests who think they've just read some unbiased guest reviews. They haven't.
A program that looked at those sorts of relationships - that become surprisingly easy to spot once you're looking for them - would filter out a huge volume of fake reviews.
Why would a review site not want to do that? I can only guess that the number of fake reviews that would be discovered would expose the unreliability of the site itself.
Only a guess. But basically when I'm looking at online reviews without my own filter, that's all I have to go on anyway. A guess as to whether they're real or not.
And that, my dear friends, is why, as is so often the case, I agree with Arthur Frommer on the future of online review sites.
Unless, of course, they clean up their acts and responsibly remove the treasure trove of obviously fake reviews on their sites.
As for innkeepers who post fake reviews, well, regular readers know what I think of folks who stoop to fraud to attract guests. Those aren't innkeepers, those are...well, I'll let you fill in the blanks on that one. Suffice to say, the description I'd use would not be a nice one.
And suffice to say that the number of guests I would refer to an inn that posts fake reviews is a very consistent number. Zero.








