These days, brands can live and die by their online reputation, and research shows that hoteliers ignore online reviews at their own peril. According to a Cornell Hospitality Research study, electronic word of mouth helps drive revenue per available room (RevPar): A one-point increase in a hotel’s average user rating on a five-point scale makes customers 13.5% more likely to book that hotel, and a 1% increase in the hotel’s online reputation score can boost RevPAR by almost 1%.
Sentiment analysis offers an automated way to synthesize the very large volume of information available from online reviews, and thereby provides information that can help shape the customer experience, as well as insights on pricing and revenue management. Because sentiment analysis offers a look at direct customer feedback, it’s extremely valuable for marketing leaders, but it isn’t always easy to execute.
Here are six guiding principles to get sentiment analysis right:
- Put data in the right context. Algorithms often misinterpret the tone of online reviews. “This was a great experience” could be said sarcastically, for example, so consider the words around the comment to identify the right context and tone. With large volumes of data coming from a large volume of customers, you have to use big data techniques and natural language processors to get the right picture of what your customers are really saying, minus ancillary comments or sarcasm.
- Get taxonomy right. This is where industry and domain knowledge come into play. There are a lot of terms that mean different things, depending on the company you’re talking about. For example, the Wyndham Wyzard refers to the Wyndham Hotels loyalty rewards program. Create a dictionary of taxonomy specific to the industry or company you’re studying.
- Assess your current approach: reactive vs. proactive. Most companies, out of necessity, have a reactive approach to responding to online chatter, setting up a social media command center or communication teams to respond to social media feedback. Of course, you do have to respond to social chatter on the latest customer service incident or react if your information is hacked, but there’s a big difference between just reacting versus gleaning insights and trends to inform your business from the large volume of available online feedback. Instead of merely responding to the topic du jour, use sentiment analysis to be proactive and more accurately take the pulse of your customer.
- Break organization silos. Sentiment analysis is a cross-functional issue. It can’t be done in isolation, so make sure that key stakeholders such as marketing, digital, loyalty, and IT are collaborating to implement it and harness it for maximum value. The marketing team can use insights from sentiment analysis to personalize the customer experience. Customer service can leverage the insights to more effectively engage, respond to and address customer concerns. The pricing and revenue management teams can tap into the insights to ensure that they’re making data-driven decisions. The data itself may fall under the ownership purview of many of these divisions, and IT could be a key enabler. Consider who can and should benefit from sentiment analysis, and who can best enable the work before plotting a way forward.
- Monetize insights on an ongoing basis. Create a strategic road map, but focus on a tactical execution approach to achieve incremental ROI. Once you get the information, what do you do with it? For instance, if customers say that the checkout process was particularly bad at a certain property, you should get that information to the right stakeholders at the right time and take action on it, and make sure that there’s a methodology for doing it. Monitor this on an ongoing basis to keep track of the latest feedback.
- Engage with a partner. Working with a partner can help you accelerate time to market by leveraging the partner’s platforms and sharing domain or company knowledge.
While online reviews and ratings are important, sentiment analysis also requires thinking about a holistic data picture. Insights can be gathered from third party websites such as TripAdvisor, hoteliers’ own websites, and internal customer feedback and call center data. Taking data from hotel websites, for instance, lets you identify who the customer is because the reviewer is required to log in with his ID or loyalty number. Then you can start creating more personalized marketing tactics, putting you even further down the path of understanding what your customer thinks of you, and tailoring the customer experience accordingly.