As a linguist, how did you get into comment/review analysis?
My background as a linguist is in spoken discourse or spoken language. I made the transition to online language in a non-deliberate fashion. I started reading reviews and as I read them, I got more interested in the language features that they included. From there I searched the scholarly databases where I didn’t find a lot of work from the linguistic perspective being done on online reviews. Obviously Economics, Marketing and Business folks have been researching them – but not a lot of research from a linguistic perspective. So I stumbled into it and thought “Wow this stuff is really interesting!” My blog will keep you updated on my latest interests.
What can brands learn from comments?
Quite a lot. It depends on what they’re interested in finding out. Businesses have more customers than the ones who actually provide reviews. Reviewers represent a small but specialized sample. I’ve read estimates that they’re about 10% of a business’ total client base. It’s a self-selected population. It’s not going to tell you everything about everybody, but it’s 10% of the people who are motivated enough to want to write about their experiences online and share what’s important for them. And, as we know, they wield a lot of influence because people pay attention to not only the aggregate ratings but also to the comments from reviewers.
In my book, The Discourse of Online Consumer Reviews, I looked at reviews from five different websites which were TripAdvisor, Yelp, Amazon, Epicurious and Netflix. Recipe and film reviews are different than reviews of products and services …so it really depends on what kind of business we’re talking about, and what they’re interested in finding out. But they can definitely mine reviews for whatever kinds of information they want. Many businesses are mining using big data approaches like sentiment analysis however my approach is a more complimentary, smaller data approach which provides a more nuanced view, one that can pinpoint or target specific features. So it would be up to the business to figure out what they are interested in looking for.
Has a brand changed its course because of a user’s comment?
(Laughs) Yes, there are a couple of things I can tell you about that. I’ve been looking at restaurant reviews on TripAdvisor and Yelp, comparing the two and comparing businesses’ responses. There have been a few case studies of restaurants in my area that have opened in the last few years, got negative reviews initially on opening and had a really active presence on social media (not only in responding to reviews but also on Twitter and other sites as well) and have actually kind of turned sentiment around. They have gone from being a business that wasn’t getting good reviews to now being really successful, thriving businesses. It seems to me that the initial investment of time, really paying attention to customers, following up and showing concern and maybe even addressing some of the problems that are being mentioned can actually turn popular opinion around. We’ve seen some cases like that.
I’ll give you another example, this is sort of a funny one. My research on legitimate reviews has led me to this side area of research about parody reviews. I don’t know if you’ve looked on Amazon, where certain products are the target of jokes that people write funny reviews of? Some of the better known ones are the Bic Pen for Her, Banana Slicer and the Three Wolf Moon T-shirt. Those products have something like 3000 reviews and and the most popular reviews have over 50 000 helpfulness votes, which are basically the same as “likes.” So the Bic Pen for Her became the target of thousands of these parody reviews, where some people would write funny reviews like: ‘Oh as a woman my tiny delicate female hands can’t open the package’ or ‘I’d use them for math but I can’t do math because I’m a girl’ and all this stuff. Sort of mocking marketers’ attempts to gender a gender-neutral product. Ellen DeGeneres even had a skit about it on her show. So it reached this massive audience, this social critique that started on Amazon reviews. I was searching for those funny Bic pen parody reviews yesterday to follow up on a paper that I’m giving, and the product is no longer on Amazon! I might be overlooking something, but I looked at a number of different search terms and I can’t find those reviews anymore. It is possible that when there’s enough of a response in online review forums, that these kinds of things can happen.
You said brands should not always answer negative comments. Why is that?
Figuring out how to reply to these reviews is a sensitive form of professional communication and I think it takes an expert communicator to figure out how to do this online where communication is somewhat decontextualized. Sometimes people end up really doing more harm than good. So before they fire off a response which may come across as sounding defensive I think it’s a good idea to have a strategy in place. Overall what’s the big picture? How many negative reviews to positive reviews are there? Are you only going to reply to negative reviews, or are you going to reply to all of them, or reply to a certain percentage? Are you going to reply privately to that customer, or are you going to make that public? And if you want to make it public, are you only speaking to that person or are you trying to address your comments to a larger reading audience of prospective consumers? There are a lot of factors at play and from the interviews I’ve done with local businesses on this project that I’m doing with local restaurants on TripAdvisor and Yelp, it seems that people make it up as they go along and they haven’t put a lot of thought into their larger strategy. I believe that’s an oversight, it needs to be given some thought. Some businesses farm it out and have another company address their online feedback. I’ve gathered some anecdotal evidence about really generic responses like ‘thanks for your feedback, we’re looking forward to having you back again and hope you’ll have a better experience’. For many people, responses like that don’t really convey that they care, or that they’re paying attention to the content of the message. Is that better than nothing? I don’t really know. I think that’s something that needs to be studied and I’m just starting to take my research in that direction.
Can brands benefit from user comments as users do?
Absolutely. There are studies that have been done with brands, like in the travel and tourism industry for example, which have found that several hotel chains regularly review the data from their online reviews. They check to see how they’re doing, if there’s anything that needs to be followed up on or improved, make changes based on the feedback that they get, or look to see if certain employees have been mentioned in a favourable light and reward them. Those kinds of practices are already taking place. Businesses are using reviews as a virtual focus group to gather information (not unlike a customer satisfaction survey) from at least from some segment of their customer base.
Do people “speak” differently when they leave comments?
Linguists and people who study language in social context like I do, follow the fundamental premise that we adjust our language according to who we’re communicating with, when, where and for what purposes. The way that you would tell a story to your grandmother would differ from the way you would tell that same story to your best friend who’s your age peer. You’d use different language, different expressions, you may use more or less slang, include more details or not. We call this audience design and yes absolutely, reviews are very much audience designed.
Although there are a number of factors at play with reviews. For some people it feels more like a stream of consciousness text that they compose, whereas others try to be very credible in how they portray their comments, maybe they’re more careful in composing grammatically correct sentences. They’re projecting a certain identity with varying degrees of awareness. Also, depending on the site there’s a different kind of sensibility that comes across. On Facebook reviews, for example, you’re authenticated through your profile so that connects you to your offline identity much more so than on other review sites where you can be a little more anonymous, where it’s not necessarily linked to your other online accounts. Some people may feel a little inhibited to be really honest if others know exactly who they are and how to find them. That’s another issue.
Of the sites that I’ve studied, Yelp asks readers to rate reviews not only according to how helpful or useful they are, but also asks readers to rate reviews on how funny and cool they are, which sets up a kind of ethos on the site. So there’s a lot more slang on Yelp than on TripAdvisor, for example. There are a lot more references to popular culture, people are more informal, and they try to be more clever and creative. You see different kinds of texts being composed there than you do on a site like TripAdvisor because of the way the site has been set up, and as a result, the audience’s expectations.
Cover photo credit: “scream and shout” by Mindaugas Danys, flickr.