Recently I attended the RILA Leadership Council meeting in Orlando, which featured top-ranking executives from key retailers like Walgreens, Staples, and Target. We had many great conversations, but one topic that really engaged us all was how to avoid consumer-generated media PR crises.
In the digital age, where everyone is now a videographer, there’s no telling when, where, or how the next major PR crisis will hit. It could be a viral video of a customer service issue, an advertisement that was interpreted wrongly, or an unpopular strategic move by your company. As a retailer, you’ll need every available tool in your arsenal to ensure you don’t end up on the evening news for the wrong reasons.
A tried-and-true method that can be leveraged is identifying high-risk stores where a major issue is likely to strike next. (This is not to say that a crisis can’t start at a low-risk store, but it’s certainly a step in the right direction.) There’s a wealth of information on the Internet — including Facebook, Google, Twitter, and more — that can tell you exactly what’s happening at your stores. The question then becomes, how can you leverage that info?
There may be some software solutions that can help. But many of these have a major limitation: the inability to accurately capture intent. Intent is the underlying meaning behind a comment, and getting it right is crucial to getting accurate insights. For example, consider these two fictional comments about a pair of shoes:
- “I bought those cheap shoes.”
- “I bought those shoes cheap.”
Read them again. Notice that despite having the exact same words, the order in which the words appear changes the meaning. #1 has a negative intent, while #2 is more positive.
Enter sentiment analysis, powered by machine learning and a key capability of Profitect’s prescriptive analytics solution. Sentiment analysis combs through textual data like reviews and comments, isolating buzzwords and key themes that retailers value (think customer satisfaction, pricing, quality, etc.). The data containing these key themes is analyzed for intent, which is then used as a type of score for rating stores (and products) on various KPIs.
With the help of sentiment analysis, you can then identify stores that are at highest risk of a PR disaster and target retraining efforts. By leveraging this type of technology, perhaps PR disasters (such as the recent racial profiling issue at Starbucks) could be avoided. This approach is certainly easier and more resource-efficient than shutting down your entire chain for training. When you combine all the information you need — the unstructured data, the scores and intents assigned to it by sentiment analysis, and the corresponding KPIs (i.e. sales reports, return rates, merchandise quality, damages, etc.) — you have everything you need to help your organization weather the next crisis-communications storm.
At Profitect we’ve seen our customers take advantage of sentiment analysis in order to drive operational efficiency and improve the customer experience. To learn more about how it can help you too, contact us at email@example.com.