People Are Talking About You: The Hidden Value of User-Generated Content
I’m going to make a couple of assumptions at the outset of this column that I happen to know are partially unfounded. The first is that, as a local business owner or brand marketer, you are already closely monitoring your consumer reviews. I know for a fact, working as I do with many brand marketers on a regular basis, that fewer of them are paying proper attention to reviews than you’d expect. Even with the right tools to help make the process more efficient, brand marketers are overwhelmed by the constant flood of user-generated consumer content, and many would rather hide their heads in the sand than make a commitment to engage at the local level.
As a gauge of consumer sentiment, reviews may not be fully unbiased. In fact, a 2014 study of restaurant reviews suggested that consumers will rate a restaurant higher if the prices are high and lower if the weather is bad. However, 97% of consumers read online reviews, and 87% won’t visit a business with less than a three-star rating. What’s more, review signals such as volume of reviews and average star rating are among the eight most influential local ranking factors.
Some brands and businesses, despite knowing that reviews matter, feel they are powerless to influence the trend of reviews one way or another. It’s true that consumer opinion is ultimately outside the brand’s control – though it is mostly reflective of realities the brand would do well to acknowledge, including product and service issues that can be improved. We’ve found, too, that brands actively engaged in responding to reviews can convert individual reviewers from detractors to fans and can increase the overall sentiment of reviews over time.
In one real example, a restaurant reviewer changed a one-star review to a four-star review after the brand responded, noting, “This is an update from last visit. I decided to give it another chance and come back, as corporate was very nice and professional with the situation we had last time we visited this location.” Another reviewer wrote, “Since customer service responded within 24 hours of this review and showed me where I was wrong initially the rating goes from one star to five stars.”
Assuming you are already responding to your reviews – although, in fact, many brands and businesses are still not doing this – we come to the next critical value point in review management. Simply put, insights gleaned from reviews can help you do business better. Though reviews may contain bias of various kinds, they are still the best source you can find of detailed feedback from real customers.
If you’re a small business, you should read all of your reviews and categorize them by topic. How many customers complain because you close too early on weekdays? Would staying open until six instead of five increase your daily revenue by making it easier to visit your store after work? What are people saying about the servers at your restaurant? Do some of them need retraining due to a pattern of issues from different reviewers?
Such insights are ripe for the picking in consumer reviews, though for multilocation brands in particular, the challenge of wading through hundreds or thousands of reviews to find trending topics can be daunting. Tools like review tagging and topic extraction can help. With review tagging, users tag reviews as they read, noting common themes like “wait time” or “rude service.” Using a platform like Brandify, users can sort reviews by these tagged themes in order to find historical, regional, and location-specific trends.
One of our retail brand clients has been tagging reviews as a regular practice for more than two years. This discipline has led to significant improvements in customer service training, where employees are guided through real-world examples of positive and negative customer experiences as case studies for proper employee behavior. The brand also uses reviews as the consumer side of a feedback loop to improve its messaging. Whenever a customer’s complaint can be traced to lack of clarity regarding a policy or service, the team uses that feedback to create more thorough and relevant messaging for the brand website and in-store materials.
Topic extraction gleans insight from reviews in an automated fashion. For example, Brandify uses IBM Watson technology to unearth trending topics in reviews and other social content. As a result, brands can analyze frequently mentioned topics within negative reviews, positive reviews, reviews by geography, historical reviews, and so on.
The most effective method of mining reviews for operational insight involves natural language processing, sentiment analysis, and machine learning. Ideally, the tool you use can be trained through machine learning to understand the phrases consumers typically use when talking about businesses like yours. A simple word like “dish” may refer to a menu item in a restaurant review, whereas in a review of a retail store it would mean a kitchenware item. Hundreds of such nuances crop up in text related to different verticals. A platform like IBM Watson lets you train its textual analysis engine by feeding in representative datasets as well as input from users (“more like this,” “mark as irrelevant”). Over time, the tool does a better and better job of isolating patterns of consumer sentiment.
Regardless of the tools or methods you use, you can’t afford to ignore consumer reviews. Surveys and net promoter scores have their place, but nothing can match the power of reviews for providing detailed evidence of consumer opinions that directly impact your bottom line.
