This month the Analytic Hospitality Executive blog is all about social media. Being the data lovers and analytics geeks that we are, we are particularly interested in social media not as a marketing tool, but rather as a new and exciting source of data that can be leveraged across the enterprise to augment and enhance your existing analytics efforts.
I have been giving a lot of thought to this topic, and it has become a particular passion of mine. In fact, later this month you’ll hear from an academic colleague of mine about some research we’ve been doing into how revenue managers could take advantage of social media, and the challenges and opportunities associated with developing pricing strategies in a social world. We’ll also talk to operators who have started incorporating social media analytics into their business strategy development. However, before we get started with that, I thought I’d take a step back and provide some definitions that will put this month’s posts in context. Despite the attention that social media has received in the popular press, at conferences, and by consultants, it can still be a bit of a mystery -- particularly from an analytics perspective.
Starting from the very beginning, for the purposes of our research, we define social media as a group of Internet-based applications that allow the creation and exchange of user-generated content[i]. These applications are designed to facilitate conversation, allowing the consumer to participate in the development and dissemination of content. Examples include review sites (TripAdvisor and Yelp), news sites (Digg), social sharing (Flickr, YouTube), social bookmarking (Delicious, Faves), and purchase/review sites (Amazon, Travelocity). There is a myriad of these types of applications, which means that conversations about your brand are taking place in many places in many languages across the globe. While this can be intimidating, it also represents a huge opportunity for hospitality companies not only to engage with customers, but also to turn all of this social activity into meaningful, actionable information.
We are all familiar with the volumes of unstructured text data generated by these sites, but there are also many other types of data in social sites. Basic quantitative data like number of reviews, number of fans, number of friends or followers, or aggregate consumer ratings is available. Social sites also contain images (How many of you have been hearing about Pinterest lately?), video, and audio. Demographic information can be mined either through site-mandated entry (e.g. age range, gender, purpose of travel, location) or through the comments themselves (e.g. my wife ordered the fish; we don’t have a place like this in Chicago). Finally, the connections among users within the networks and the impact of their social media activity can provide an interesting source of information.
I generally separate social media analytics into three basic categories: descriptive statistics, social network analysis, and text analysis. Descriptive statistics provide a snapshot of historical and current performance. They answer questions like: How many fans do I have? How many reviews have been posted over the last six months? What is my average rating on each of the Online Travel Agents (OTA)? Social network analysis (SNA) is an advanced analytic technique that uses the connections among users, and the impact of their social activity, to determine the degree of influence each participant has within these social networks, and who they are influencing. Once you have this information, targeted marketing efforts can be directed at the most influential users such that they spread your message for you. For example, you could use SNA to identify the influencers in a cruise community and invite them to preview a new ship or a new itinerary with the hopes that they will rave about it within their community.
Text analysis, which is used to quantify unstructured text data, includes: content categorization, text mining, and sentiment analysis. Content categorization automates the process of categorizing text documents according to their content and tagging them to optimize search. This time-saving technique keeps organizations from having to read through every document and manually tag them. Content categorization could help hotels and casinos to automatically categorize open-ended responses on guest surveys so they could be routed to the appropriate department, whether it is housekeeping, front desk, or maintenance. Text mining, similar to data mining, uncovers related concepts in large volumes of conversations, and surfaces key topics that can be used to understand or predict guest behavior.
For example, text mining could tell you that the most frequent topic of conversation among your customers relates to the loyalty program, and when guests mention the loyalty program, they also talk about earning and redeeming points. The final technique is sentiment analysis, which uses natural language processing to determine how guests feel about attributes of your brand, product, or service. Sentiment analysis will tell you that guests are relatively positive about your beds, but they have negative feelings about the service at the restaurant. Best of all, social data is public data, so you can grab your competitor’s data and perform the exact same analysis on it that you do on your own.
Once you have quantified unstructured text and collected descriptive statistics, traditional advanced analysis like forecasting, predictive modeling, and optimization can be used on this data to gain additional insight. For example, you can track trends in sentiment analysis to predict where service problems may occur, to target training efforts, or to help understand where to invest in refurbishments. Volume of chatter after a promotion can be used to predict success of future promotional activity. Results of text mining can be incorporated into behavioral analysis to enhance acquisition, retention, and attrition modeling. Influencer scores can be incorporated into the customer value calculation to place a dollar value on a particular guest’s network.
Where do you start?
With so many options, it can be overwhelming to determine where and how to get started. To help with this, some colleagues and I developed a framework[ii] (below) that places the direction of social media communication (inbound -- or consumer generated and outbound -- or firm generated) against the scope of the decisions must make (short term -- tactical and long term -- strategic). Inbound information flow is the data generated in social networks by the consumers, and outbound represents the firm’s communications through social media channels. Most hospitality managers, particularly analytic hospitality managers, are used to working with data and with channels, so social media simply adds another data source and another channel. My co-author, Dr. Noone, will place this framework in the context of revenue management later this month (which is where the quadrant examples come from), but I propose that it can be used by any department that is struggling to find a way to take advantage of social media.
What you will realize as you begin to apply this framework, and what I think is key to successful social media analytics, is that the right approach is to start by identifying a business problem you need to solve, and then to determine how social media analysis could be used to improve the solution. Instead of letting social media drive you, think of incorporating it as another data source and channel to augment your existing tool basket. This approach makes it much easier to envision how social media can be leveraged across the organization (and will probably give you more ammunition to justify your investment in Social Media!).
[i] Kaplan, A.M and Haenlein, M. (2010) Users of the world, unite! The challenges and opportunities of social media Business Horizons 53 (1) 59-63.
[ii] Adapted from Noone, B.M, McGuire, K.A. and Rohlfs, K.V (2011) Social media meets hotel revenue management: Opportunities, issues and unanswered questions, Journal of Revenue and Pricing Management 10, 293-505.
How and when hotels and casinos should participate in social media conversations was a topic addressed in the conclusions paper, Getting In on the Conversation: The Power of Social Media in Hospitality and Gaming, from a Webcast sponsored by SAS and the Center for Hospitality Research at Cornell University’s School of Hotel Administration.