In addition, there was kind of a subtle effect of this negative information such
that if I got a negative review on a book that I had written, a one star review,
that seemed, in some, sense to also slow down the rate of positive
reviews that were coming along subsequently.
So, again, a very, very interesting study showing that reviews are affecting sales.
Now, of course, this might motivate people who are selling books
to review their own books if they know that it's going to have a positive effect.
So what I'm showing here, and you can read the articles and the links, is a picture
of a very interesting lady, Harriet, who has reviewed over 25,000 books on Amazon.
So I'm not suggesting that she hasn't, but
it's just interesting to notice that one person can generate so many reviews.
Now, in addition to our friend, Harriet, there was another fellow who was doing
something perhaps a little more devious but in some ways quite clever.
So there's a fellow called Mr. Rutherford.
You can read the article I provided a link for.
What he decided to do was to offer himself as a review-writing service for authors.
So if you'd just written a book, he might be willing to write 20 reviews for you for
some sum of money, let's say $100 or $500, completely, of course, fake reviews.
And what was interesting about this was Mr. Rutherford
was making about $28,000 not a year, but a month by doing these fake reviews.
So what does this tell us?
It tells us that information's powerful and it also tells us that sellers might
want to manipulate the information that's offered.
So, you might be sitting there scratching your head as I was when I first read
that story.
Boy, somebody making almost $30,000 a month providing fake reviews.
So clearly, this is a bit of a problem, and it's come to the attention of
many other writers and kind of experts, and so, how might one address this problem
of fake information being out there affecting people's behavior?
Well on the next slide,
I'm showing a story, you can read the full story in the New York Times,
about a gentleman who's an expert on something called data mining.
So what he would do is he would go through all of the text of reviews and try and
figure out what reviews are real and what reviews are fake and so on.
Now, even though that's a pretty sophisticated technique, it turns out that
it's actually difficult for a computer to figure out what's really true and
what's not under certain kinds of circumstances.
So it would be difficult, for example, for a data mining
algorithm to figure out if I was saying something sarcastically, for example.
So those nuances of language might be quite difficult.
So to look at this question from a different perspective, Dina Mayzlin,
who's a professor now at USC, she was one of the original authors on the other study
I mentioned about looking at the effect of reviews on Barnes & Noble and
Amazon, decided to tackle this problem in a very, very interesting way.
So let me illustrate that for you.
What she did is she wanted to compare ratings on TripAdvisor
versus ratings on Expedia for the same property.
So let's choose a property so we can all be clear about what we're talking about.
So it turns out there's a Sheraton Hotel on the corner of 36th and
Chestnut Street here in Philadelphia.
So Dina and her co-authors were curious as to whether the reviews of that property
were the same on TripAdvisor as they were on Expedia.
So most of you have probably been to those websites,
and you'll notice again it's a five star review system, and
you get a histogram of reviews, so some number of one star reviews,
two star reviews, three star reviews, four star reviews, and five star reviews.
Now if both of those sites are providing the same objective
information those histograms should be identically matched, right?
So that was the test that she wanted to see.
Were the histograms different for the same property?
And what she found was on TripAdvisor there were slightly more one star reviews
for certain properties and also slightly more five star reviews.
And when she dug in a little bit more what she found was if you were operating that
hotel, the Sheraton Hotel at 36th and Chestnut Street and located quite close
to you was a competitor, a competitor who perhaps just owned their own little hotel.
Let's call it Chris' hotel, Chris Heffron, the videographer.
And Chris's hotel is competing with the Sheraton, and because he's a one man shop,
not really accountable to a big organization as I would be if I were
a manager at the Sheraton, he would be more likely to write a one star review and
try and knock his competitor, the Sheraton at 36th, and he would be more likely to
write a five star review and to praise his own property.
That's what Dina and her colleagues found.
So again, it's very, very interesting to think about whether review information is
valid, because certainly people are acting on the basis of review information.
So this is a good point for me to now summarize the key idea,
if you were to introduce reviews into a market,
what you'd want to do is make sure those reviews are objective and verifiable.
And so what Dina did in her study is she made an assumption that reviews
on Expedia are more objective than reviews on TripAdvisor.
So I'll let you think about that for
a moment as to what the reason might be if you've looked at both of those sites.
I use both myself when myself when I'm deciding where to stay.
TripAdvisor was less objective than Expedia for the following reason.
In order to be able to post a review on Expedia you have
to have stayed at that property within the last six months.
So Expedia will send you an email and they'll know
from your credit card receipt that you actually booked and you stayed there.
Where as at TripAdvisor, you are supposed to have stayed there, but
that need not be the case.
I can go and write a review for a property that I had not stayed at.
So again, I think this is a fascinating area of online offline competition, and
again, I'd like to challenge and
encourage those of you out there who are thinking about starting your own business.
If you can start a business that brings information into a market,
that can be something very, very powerful.
If you think about the story of how TripAdvisor has grown into a multi billion
dollar company, that's essentially just gathering and
processing and providing review information that's generated by all of us.
[MUSIC]