Support Skift’s Independent JournalismMake a Contribution Now
Professor Camilla Vasquez of University of South Florida, will speak about about loudsourcing, the inside world of online reviews at the Skift Global Forum on October 14 and 15 in Brooklyn, New York. See the complete list of amazing speakers and topics at the Skift Global Forum.
Vasquez has culled through hundreds of thousands of reviews on TripAdvisor and Yelp in the last seven years to study how customers express themselves on these platforms. Skift geeked out with her about words like “definitely,” “unfortunately,” and “literally.” And we also had a fascinating talk about how technology shapes language, context and emphasis.
Below is an edited version of our conversation.
Skift: Why do you think online commenting has become such a big part of contemporary culture?
Vasquez: Some people argue that it’s related to this more democratic way of being in the world that’s afforded by being on the Internet. We’re networked with billions of people around the world, we can share our opinions, we are no longer dependent on top-down modes of information. So we, as consumers, don’t need to rely only on companies to give us information about themselves.
We’re not dependent only on the mass media to tell us what to think. We can also inform one another. Communications scholar Henry Jenkins talks about participatory culture. We all can use the internet to participate. Of course, other people critique and challenge that and say, “Well, you know, not everybody has equal access really and there are other things that we need to think about and not everyone’s participating equally, too.” If you’re on Facebook or Instagram, you might have five thousand friends, but only two hundred of them are active posters every day.
It’s the same thing with online reviews. Not everybody who stays in a hotel is going to post a review. It’s maybe ten percent of the total population. But it’s certainly a vocal and powerful and influential ten percent. And it’s also ten percent more than could do that, voice there opinions publicly, fifteen years ago, before we had this. If you were happy or unhappy with something, you could write a letter to the business and it wouldn’t really go much further than that. Now you can make all of that public and let lots and lots of people know.
Skift: What reasons compel people to immediately post about their hotel or restaurant experience?
Camilla: You’re not going to remember a restaurant experience maybe a year down the road. It’s one of those things you have it, it’s fleeting, then if you are really happy or unhappy, you may post something about it. Usually that’s what happens. Although I do have to say that I think in the earliest TripAdvisor reviews that I studied, some of the negative ones, there were people who said, “You know, I’m sorry it’s taken me so long to post.” I do have a few people who say that and who maybe would post a complaint six months to a year later.
One of the things that is interesting, now that I’m thinking about the site-specific stuff more, on both TripAvisor and Yelp, one of the very, very common features of the evaluation is some orientation to futurity. Like, “I would definitely come back.” Or, “I can’t wait to come back.” Or, “I want to eat here again.” Or, “I would never stay here again.” That seems to be the overall form of assessing it. Maybe, “The bathroom was nice and the room was small but the pool wasn’t that great. Overall, would I stay here again? Yeah, I guess I would.” That kind of final verdict.
Skift: Are you saying that the platforms are created in a way so that the consumers can voice their opinions that is closely related to a customer satisfaction survey?
Vasquez: Absolutely. In some sense they’re even broader than a customer satisfaction survey because usually when you’re given one of those, you have eight or nine different aspects that you’re being asked to orient to and you have to rate it on a certain given parameter.
But here, with these reviews, they’re open-ended, so people can orient to any different part of the experience that they feel is important to them. They’re not as constrained as they would be with a customer satisfaction survey.
Skift: With review volumes escalating on social media these days, brands are trying to sort through the conversations via sentiment analysis. How does this factor into your approach?
Vasquez: My approach is a complimentary approach. Sentiment analysis is a big data approach and it can give you a broad brush stroke of the big picture stuff. It’s certainly very useful. Especially if businesses are trying to target some particular thing.
What my approach does, is it shows how specific language features are used in context. I was just thinking as I’m planning my presentation for October, I’m thinking about one of the examples. The word “nice.” If you were to run a sentiment analysis software on something and you were to see the word “nice,” the way that those programs work is that they say, “Okay, here’s a positive word, here’s a negative word.” I mean, that’s an over-simplification, but that’s kind of what they do.
I found several uses of “nice” where nice doesn’t mean nice. Let me give you an example. “The walls are paper thin and I could here the couple next door arguing all night. Nice.” You could almost hear the tone. It’s an ironic kind of “nice.” It’s the opposite of what it actually means. Sentiment analysis isn’t going to pick that up because it can’t do irony. Only humans can do irony right now.
There are very context-specific meanings that are overlooked by those big data kind of approaches. That’s one of the things that I’m interested in. Or, what was another thing? “The pool was nice and black.” Well, obviously it’s not nice if the water in the pool is black. It’s being used, again, in a different kind of way.
Skift: How do people typically express their experiences online?
Vasquez: I looked at certain kinds of adverbs that are called stance adverbs [instead of adjectives]. I did compare those to general corpora of spoken language and I found certain ones that are super-frequent in online reviews.
“Definitely” is one that appears a ton. When you think about that, when we say, “Definitely,” you want to sound really convincing or persuasive. That’s speaking to that dimension of these texts. One of the functions of them is if I’m going to bother writing a review, it’s because I want to convince you to stay here or not to stay here or that it’s really a great place or not a great place.
“I would definitely stay at this property again” or, “We’ll definitely be coming back here.” But even in other ways. “I would definitely recommend staying here.” It connected a lot with recommendations, even with food. “The pork was definitely delicious.” “We were definitely satisfied.”
Skift: As opposed to just saying, “We were satisfied?”
Vasquez: Right. Because it makes it stronger. It sounds more convincing. “Definitely” was one. Another one that was way more frequent in reviews than in speech was “unfortunately.” That may be because I have a lot of negative reviews in my sample, too. Obviously, that function is often used at the beginning of a sentence. One of my findings was that even in super-super-negative reviews, reviewers will try to say at least one positive thing.
“The place was big, but unfortunately that’s the best thing I can say about it.” Or “They offered us an extra night free. Unfortunately, there’s no way I’ll stay there again.” People call it a subversion of expectations. You start off saying something positive, then when you want to reverse the sentiment, you preface it with “unfortunately.”
A third one, then I’ll stop with the adverbs. This is some language geek stuff. This is when you start looking at the micro-micro-micro-level.
“Literally” was another one that turned up a lot. Literally, I think, maybe has more to do with youth culture and informal language. That’s also a target of when people comment on what you shouldn’t do with language. You shouldn’t say “literally” because usually what you’re saying isn’t actually a literal thing.
But young people do this a lot. It’s another way, I think, of being emphatic or really stressing something. “It literally felt like we were staying in a jail cell.” “It’s literally in the middle of nowhere.” “It literally took ten seconds for the elevator to come.” That kind of stuff.
Skift: How do you feel that smartphones are changing the way that we communicate?
Vasquez: Obviously, we can do it any time, anywhere. The nature of texting is similar to communicating in 140 character on Twitter. We can write really short, compressed messages. Which is really convenient but then can also open up room for ambiguity and misinterpretation because you’re relying on fewer cues.
I don’t know if you’ve had that experience of agreeing to meet somewhere, but then you thought that the person meant something else and whatever, because you did that all by text? Somebody like me, who’s not a digital native, who had these experiences where it’s like, “Oh, that’s what you meant by over there, whatever.”
Skift: How about in terms of reviews sent from smartphones?
Vasquez: I know that TripAdvisor and Yelp, for example, have smartphone interfaces. What I’ve been hearing from folks that I’ve been talking to, this is on the business end, people are talking about their experiences. It was consumers, even though they’re speaking as PR people or restaurant managers or whatever.
One of the questions I’ve been asking folks is whether they think that either Yelp or TripAdvisor is more influential. This is in the realm of the restaurant world. I do have some restaurants that are connected to hotels, but some are just stand alone restaurants. Most people are saying that they actually don’t like Yelp as much but they say that it’s more influential and recently I talked to somebody who said, “Yeah, I think it’s just because of the ease of the interface on the smartphone.”
Yelp is just so easy to use and TripAdvisor is still a little bit clunky in that respect. That was the first time I’d heard of it. But I do know from just going out with friends and stuff that people who are trying to figure out where to eat, they’ll turn to Yelp first, not TripAdvisor.
Skift: In relation to your TripAdvisor work since 2007, do you feel now that people are using more of these emoticons, emojicons and also including pictures to round out their review and experience?
Vasquez: Absolutely to the second one. Not to the emoticons and emojis so much. We don’t use those as much as we think we do. That’s another one of those popular misconceptions about internet language. We’re just full of smileys all over the place. We’re really not. We use them selectively and carefully unless we’re talking about some particular segment of the population.
Overall, and I think in this type of communication as well, people are not using that kind of stuff. But the images, definitely. I’ve seen an increase in the number of photos, obviously, posted on TripAdvisor. Just in the last six or seven years since I’ve been studying it. I think it comes more with the mobile technology for one thing, as it becomes easier to upload images. That’s a big factor.