August is get-away month, so the hospitality industry should start gearing up for what happens when all those travelers get back home….and start to record their impressions of the properties at which they stayed across the social media landscape.
Last week saw the integration of Lexalytics’ sentiment analytics engine into Revinate’s software-as-a-service social media reviews-tracking solution for the hotel industry. In the fairly recent past, Aptech Computer Systems signed on to use Clarabridge’s sentiment and text analytics software for its Execuvue Business Intelligence software for hotel operators. And many other names in the social analytics space, from Attensity to SAS, also see the hospitality sector as a key segment when it comes to mining customer sentiment. At this spring’s Sentiment Analytics symposium, Lexalytics CEO Jeff Catlin called “travel and tourism a natural spot to use it. There’s a lot of data feeding back all the time that helps them make money,” he said. “You can learn things when you’re scoring tones on certain attributes.”
By integrating into its mix Lexalytics’ sentiment analytics (see story here ) – technology that includes capabilities such as Facets for selecting the most important ideas or subjects within content and their pertinent attributes—Revinate now lets hotels measure guest sentiment of all key topics, with scores for each key category of operations (rooms, beds, blankets, etc.). They can view this information for a specific time period or how it trends over time, and take action from the software itself to do things like assign in-house tickets to personnel to address. They can also view sentiment from a competitive angle, to see how they rate against other properties. Prior to the integration, Revinate let users understand through keyword analysis words that were popping up a lot in social media reviews – “bathroom,” for instance, or “bed” – but it was up to the user to manually review the content to understand whether those were generally positive, negative, or neutral contexts.
Business increasingly depends on the perception online visitors glean about hotels through web site reviews, says Michelle Wohl, VP of Marketing and Client Services. “We’re seeing that when hotels are on the first or second page of TripAdvisor or Expedia it drives more business,” she says. Getting into that hotspot on TripAdvisor, for instance, takes into account its Popularity Index that includes how many reviews a hotel receives and user ratings. She cites findings from hospitality feedback solutions company Market Metrix that show that reputation, recommendations and online reviews now drive more bookings (51 percent) than either location (48 percent) or price (42 percent). The e-tailing group has found that 92 percent of Internet users read product reviews and 89 percent of people say that reviews influence their purchasing decision. And 35 percent of social media users changed their hotel after browsing a social platform, according to World Travel Market, which runs global events for the travel industry.
It’s not just the business of the Mr. and Mrs. Jones’ of the world and their two children that are at stake, either. Hotels are increasingly hiring social media managers and others with that savvy in their sales and marketing functions who hope to use positive sentiment and other social media metrics as an in with corporate travel departments. Users of Revinate, for instance, could attach to RFPs (requests for proposals) reports showing how they rate against the hotels with which they compete.
Revinate says its integration of the Lexalytics’ technology focused on transparency, so that users can drill down to entire reviews or snippets to see if they agree with its interpretation of the sentiment expressed, and change it if they choose. “We know sentiment is not foolproof so we provide a safeguard to reassess it if they feel differently,” she says, rather than providing black box analytics that require users to trust the interpretations.
The company says it covers all major sources of online hospitality reviews, and that customers are leveraging the capabilities to do things ranging from re-educating employees to fix a conspicuously reported issue, to figuring out where they should spend renovation money. One hotel company, she notes, discovered that it was getting negative commentary on the fact that it still had tube TVs in its room, so it knew to put the money to flat-screen upgrades. “Or maybe they want to change their parking fees, to raise them. They can see how much people are talking about that as issue, and what competitive hotels are charging, and do people not like that,” she says. The company also recently integrated its system with Klout, so now hotel operators can suss out sentiment among those with high Klout scores – and perhaps more quickly address the negative opinions brought up by those users.
Revinate's pricing algorithm is determined by a hotel opreator's average daily rate and number of rooms.
But Are Those Reviews Real?
Another question entirely is how real are some of the opinions that get expressed in online review forums. In the Chronicle Online, the daily news site of Cornell University, it was reported last week that Cornell researchers are working on computer software to identify deceptive reviews, starting with the hotel sector in Chicago. At the 49th annual meeting of the Association for Computational Linguistics in Portland, Ore. last month, professor of computer science Claire Cardie, associate professor of communication Jeff Hancock and graduate students Myle Ott and Yejin Choi reported on their approach of ferreting out opinion spam.
The work includes analyzing text for the presence of concrete words (bed or bathroom, for instance) and a greater number of nouns than verbs as an indicator of truthfulness. By combining keyword analysis with the ways certain words are combined in pairs, they were able to identify deceptive reviews with 89.8 percent accuracy, it says. The researchers are looking to extend this to other categories.