November 29 2007

[Comment] Review authenticity – Is there really a problem?

Online reviews, friend or foe? Bazaarvoice’s Chief Marketing Officer, Sam Decker, put us straight.

Our recent piece on a kitemark for review authenticity created more than a little controversy. Sam Decker, Chief Marketing Officer at Bazaarvoice, sorts the hype from the facts.

When a retailer is considering embracing user generated content, there are two questions often raised. The first one, ‘What about negative reviews?’, is the easiest to address. Of the millions of reviews submitted across 150 Bazaarvoice clients in 15 different industries over the past two years, 80% were positive. The second question is ‘What about fraudulent reviews?’. In answer to this we want to say ‘What about them? What makes you think this is a big problem?’. Mark Twain wrote: “The reports of my death are greatly exaggerated.” Hype and edge-case thinking ensures that every so often we run into companies with this concern when they are considering adding the customer voice to their sites. So, first it’s important to put this issue into rational context before we over-engineer solutions to an exaggerated problem.

With the rise of user generated content and reviews, occasionally a company gets caught making the mistake of posting user generated content in their favour. This, of course, gets a lot of press and attention. Is the attention these stories get a barometer for the frequency of fraudulent reviews? On the odd occasion this does happen, how does it compare to the volume of authentic participation? Or is it that in this age of consumer participation an action like this may increase the outrage and coverage, and therefore awareness? And logically, isn’t this awareness itself a deterrent for every other manufacturer, hotel or author not to make the same mistake? The risk of posting a fraudulent blog, review, forum comment or any other type of user generated content far outweighs the reputation cost. An executive who enjoys their job would not endorse or pay the salary of someone to engage in this practice. Who would take that job anyway? Frankly, it’s unlikely that large manufacturers and brands have the competency, focus and resources to keep up a ‘fraudulent reviews’ operation.

That’s the first and easy answer to the fraudulent reviews question. It’s logical, and most people are savvy enough to trust their judgment if they suspect they’re reading fraudulent user generated content.

When the issue does arise, it should of course be taken seriously – and there are several ways for online retailers in particular to deal with such a situation.

Whatever tactic you take to avoid fraudulent reviews, you must make sure that it doesn’t deter your best customers from participating. You shouldn’t limit the number and type customers who can participate, because they are part of the solution. After working with 150 online retailers over the past two years, we’ve found that a balanced approach encourages participation while deterring and identifying fraudulent activity.

Here are the best practices:

1) First off, ask customers to submit reviews via a submission form. It would be far more tempting for a fraudulent reviewer to submit a number of ratings-only posts to boost the average rating. By requiring the customer to write a review, of at least 150 words, you make it more time consuming for fraudulent writers and raise the risk for them (the following reasons will explain why). Besides, you wouldn’t want customers to only rate products. It’s the review text that really helps others make a purchase decision.

2) Integration with a retailer’s ‘my account’ system is a best practice. When a customer writes a review, you should ask them to log in or register with your ‘my account’ system. If they’ve purchased online they will have an account, and if they purchased offline, they can create an account. This is also valuable from a marketing perspective as you can identify your influential customers. From an authenticity perspective, the retailer can enforce their own level of authentication and verification in their ‘my account’ registration system, rather than being forced into a standard or third party verification.

3) One thing to remember is that nothing replaces human moderation. When you actually read each review before it is posted, you can ‘sniff out’ suspected fraudsters. After reading millions of reviews, this is why we’ve come to the first conclusion in this article: that fraudulent activity is exaggerated. However, when something doesn’t read right – it’s the first review that’s highly positive and is too polished – we can raise a red flag and investigate the IP address, date of registration, purchase history and more. Fortunately, there aren’t a lot of these.

4) Customers are the best police in a community of people posting user generated content. In the same way moderators can catch suspicious reviews, thousands of customers who read the reviews can click on the ‘Report an issue’ button which can induce an investigation.

5) The community always wins. If for example a pet supplies company sells a dog treat that has received 700 reviews and fortunately, almost all of them are positive, but a competitor then posts a negative review, then two things would happen. If the customer didn’t suspect it as a fraudulent review and report the issue, they would be compelled to participate themselves, to ‘tell their truth’ about the product. Other customers would do the same. In the end, the authentic reviews from customers far outweigh the fraction of postings (or potential for them) from fraudsters.

Finally, if none of the logic or deterrents convinces you, purchase verification for accepting reviews is always an option. Through email or ‘my account’ verification we can identify if the person writing the review is a customer. However out of our 100 plus multi-channel retailers none have chosen this approach.

There are two reasons for this. First, you prevent participation of offline – those buying in-store - customers. We’ve found that with the right promotion, over a third of a multi-channel retailer’s review volume can come from store customers. These customers create reviews that help drive conversion and help discredit any possible fraudulent review. Secondly, if some reviews have a purchase verification ‘badge’ and others don’t, it makes the store customers’ reviews (without verification badge) look less credible, which in turn diminishes the credibility of the overall rating for that product. Verification (and verification badging) may be a good option for online-only brands, but not for multi-channel retailers.

So, while there will always be unscrupulous people who abuse the trust of consumers by posting fake user generated content, we can rest assured that they are few and far between. We can outsmart them with technology and business processes, while consumers possess perhaps the cleverest thing of all: instinct.

by Marcus Austin (Web Editor)