I showed the manager that there were mentions that could be resolved in seconds. Kaiser is going to make the social team a 24-hour operation because of information that I found. That’s pretty cool
Kaiser Permanente is one of America’s leading healthcare providers. Thomas Hall is senior analyst at Kaiser Permanente, where he discovered a way to keep an eye on social chatter to identify trends and issues by geography, and empowered Kaiser Permanente to respond to comments at the speed of social.
Customers expect companies to respond via social communication. Kaiser Permanente wanted to monitor mentions on social media outlets including Twitter, Facebook, Pinterest, Instagram, Yelp, blogs, and microblogs to proactively reach out and address these complaints. Hall explains that 72% of people who complain on Twitter expect a response within an hour, and 60% of people said they had negative feelings towards a brand that did not respond quickly.
Hall explains, “At Kaiser Permanente, we love data. And we have access to a ton of social data. It’s all public. It’s all free. You can go and connect to any social API you want and get however much data you want from there. But it’s unstructured. It can be messy. There can be misspellings. It can be all over the place.” Hall needed to make sense of it all, and he needed to create a process so that his team could identify and respond to company mentions on social media quickly.
Kaiser Permanente uses Alteryx to monitor social media. Hall used the Twitter API connector tool to build a workflow that looks at topics, checks for frequency for topics and words, considers two- and three-group words, and considers how many followers a user has. Hall developed algorithms that run on a pre-defined schedule, and he gets an email alert if certain conditions are met concerning social mentions. Hall explains, “With Alteryx, you can build your own social monitoring tool, which is pretty awesome.”
Hall uses the R tool to run custom R scripts that process text algorithms. The workflow breaks sentences up into little pieces or tokens, called “tokenization” to allow analysts to understand what people are talking about. He also talks about complex algorithms and machine learning to improve the ability to decipher sarcasm in sentiment analysis.
The Twitter search tool also allows Hall to use spatial capability with geolocation search to see exact places where tweets are coming from. Kaiser Permanente has different regions, with medical offices in each region. So if there’s a spike in negative tweets about a facility problem, Hall can identify where the problem is happening and at which facility. He can notify management, so they can fix it.
Hall notes that he uses the RegEx tool to clean up filler words like and, the, or but so he can focus on the content that he wants to analyze. He also likes that he has the flexibility to use R to write any additional script and include it in his Alteryx workflow.
Alteryx enables Kaiser Permanente to undertake fast analysis of current social data and make the best decisions to help their customers. Once data is available and analyzed, the team can see real stories about their customers’ experiences and how they can help.
Hall shares an example about how Kaiser Permanente used Alteryx for social media monitoring. Recently, a user with more than half a million Twitter followers tweeted that he had lost his Kaiser account and couldn’t get it back. About an hour after the member tweeted, Kaiser Permanente’s social media unit was alerted and reached out to help. The customer responded and his problem was resolved the next day because Hall had the proper infrastructure in place to allow the social team to respond quickly. The member’s social followers saw Kaiser’s quick action and resolution to the problem, which turned a potential negative situation into a positive one.
Hall explains, “Because of the tools that we have, we can see data, we can analyze data, and we can get the information to the managers who need to make a decision. And ultimately, that’s the point of our work. That’s the point of analysis — to give managers who make decisions better data so that they can make better decisions.”
The social team is now empowered to have significant impact, and Hall knows that the best impact will come from fast response time in social media. He did some research and found that most tweets come in after business hours when the social team was offline. Based on the results of his analysis, the social unit changed their hours of operation to be able to respond faster to activity that occurs after regular business hours.
Hall plans to continue to build Kaiser Permanente’s analytical infrastructure to allow them to be proactive in responding to customers via social media. Kaiser looks forward to unleashing the full power of Alteryx. Hall says, “The Alteryx software has really helped us at KP get inspired from social data and we’re really looking forward to the future.”