Art by fellow Texan, Hugh MacLeod.

Re-read this classic post about the Consumer/Marketer Control Framework, and this one: How Advertising Will Evolve Using Word of Mouth.

Many people have been emailing this article to us asking for our thoughts, as Bazaarvoice has built a reputation for over three years of hosting and managing authentic user generated content for major online retailers and manufacturers.

Here are my thoughts on this situation…

99.9% Authenticity

In this incident, neither Belkin nor Amazon were our client. We can’t speak to their policies, technology and practices to prevent this from happening. Here at Bazaarvoice, we take authenticity very seriously and employ a wide variety of ways to check for fraud and discourage cheating. Account verification, purchase verification (optional), IP address, time/date stamp data, text analysis (human + technology) and a reach of reviews across many retailers give us a variety of methods to catch this stuff.

The danger is that some people (mostly industry people…as consumers may not internalize this story) may conclude from the Belkin incident that product reviews are not authentic. End of story. As serious as this is, that's an unrealistic and “chicken little” perspective on the situation. Instead, here is the data that we’re finding:

For over three years, we’ve managed user generated content for nearly 300 major retailers and manufacturers. We’ve moderated (meaning read by humans) millions of reviews. Through the human moderation, technology, and data analysis testing for fraud we’ve found that there is a fraction (less than .1%) of reviews that appear fraudulent.

You may ask, “What kind of fraud do you see from this small fraction?” We mostly see someone post the same review multiple times for the same product or different products. Sometimes this is an attempt to win a contest. Those reviews are easy to catch, and they don’t get published. Typically the account and/or IP are blacklisted from the site, and the retailer or manufacturer pursues further investigation.

Some have asked about about our moderation. At Bazaarvoice we DO NOT moderate for negative reviews. And we won't take on a client who asks us to do that. All valid product reviews are posted. Those that are rejected — also a small fraction — are usually because the reviewer is writing about customer service, not the product. And these are often sent to our clients' customer service departments. I believe we have the most rigorous process in the business. This is what we do, and we've got the best technology and moderators in the business. In fact, we have a 40+ page presentation JUST going over on our moderation process. We don't rest on our laurels, and shouldn't with situations like this. We continue to add process and technology to fight the <1%. With every 7-week release of technology, with a growing analyst team, and with collaboration from some of the largest retailers in the business we continue to get better.

Regardless of everything we do, there is always a way to fake, hack, steal, and manipulate systems on the Internet. Ecommerce coupon codes are manipulated. I've been in the online industry for a long time. I know that software keys are distributed. People download copyrighted music and movies. Contests are gamed. Viruses are spread. Computers are hacked. And, yes, a few people even post fraudulent reviews. But those few “dingys” who post fraudulent reviews are drowned out by a sea of tens of millions of authentic reviews. And with each public instance of outrage, such as this Belkin case, we are reminded of the need to raise our game to eradicate potential issues for clients.

Downside (and Upside) To This Situation

The downside of this situation is it can scare companies into inaction. It will be terrible for us — as consumers and marketers — if executives pull back from social media and user generated content initiatives because of this press. The momentum on corporate transparency and customer participation could be slowed or reversed. Executives may return to their "control the message" doctrine for PR and marketing. But we know consumers are seeking authenticity. Backing away from UGC and market conversations is the wrong direction for companies.

How could there be good from a situation like this? Because situations like this can also scare companies into action. I’m going to postulate that Belkin managers and executives never proactively set forth policies of creating or managing user generated content for their employees or agencies to follow. As a result, I’m guessing Mark Bayard, incentivized to improve Amazon sales performance, combined initiative with poor judgment to write his own reviews and pay for more. I’ll postulate that -– given the backlash against the brand — that Belkin executives learned a valuable lesson. I would imagine there are proactive measures being handed down right now to try to prevent this from happening again. I hope Amazon also takes proactive measures.

Belkin and its predecessors who have fumbled user generated content initiatives teach the industry a strong lesson. With each story like this, more marketers are spurred into taking proactive steps to prevent fraudulent UGC.  They see the kind of press and social media coverage fraudulent action has on their brand. Any executive thinking about condoning activity like this will think twice. Employees and executives don’t want to lose their jobs. Executives will put forth proactive policies to employees and agencies with regards to social media. They should also join WOMMA and agree to its ethics guidelines.

Do you think Walmart, Whole Foods, and now Belkin learned a lot from their situations of fake user generated content? Did they take action? I’ll bet they did. And I’ll bet each of these public mistakes teaches the rest of us to do the same.

A Wake Up Call for Brands

This is an unfortunate incident. But as I said, there positives from this. It’s a wake-up call for brands to take their user generated content policies and strategy seriously. The demand for UGC from consumers is increasing. 8 out of 10 consumers read reviews this holiday before purchasing (Nielsen, 1/09). And more and more UGC is coming up as search results. The answer is not to turn away from UGC; it’s to bring forth new solutions to improve the authenticity. This is an opportunity to evolve, to improve and make customer-to-customer even more effective. And that is to say, as a company and as an industry, we will continue to work to maintain its authenticity.

According to a study by MIT and University of Michigan they found a high correlation between offline word of mouth and online reviews for movies:

Another interesting finding is the higher correlation of offline word of mouth to online word of mouth compared to offline word of mouth to critics reviews.

This doesn’t surprise me. Usually when I talk to friends about a movie I hear similar feedback. Now a woman may give me slightly different ‘perspective’ on a romantic comedy than her husband ?, but the general recommendation is nowhere near as diverse as critics’ reviews. See the reviews on Invincible, from B+ to D+.

Ultimately, the question for authenticity is relevant for online reviews. There is the possibility of shills, fraud, and ‘astroturfing’. The truth is, with over 40 clients and hundreds of thousands reviews, we don’t see much fraud (and we’re getting pretty good at catching them). Of course, nothing is fool proof. With the growth of consumer-generated content, this will continue to be a focus for us to stay ahead. It’s also why our solution includes several fields to add context to the review. In usability studies we find that provides shoppers with review relevance as well as *reviewer* relevance.

The Edelman group found that ‘trust in someone like me’ has tripled over the last two years. The key phrase here is ‘someone like me’. Shoppers identify with the reviewer based on the content of the review, user attributes, and product attribute ratings.I would also assert the community of customers and shoppers are the best judges of authenticity and relevance. With helpfulness votes (which is sortable) and abuse reporting, shoppers themselves can ‘smell’ fraud. We find many of our clients' cutomers are compelled to jump in with their own review when a review that is ‘off’ is posted. So between our content analysts (we read every review and review several factors for fraud) and customers, we can maintain a high standard of authenticity.