We all know that Big Data is a hot topic. In fact, according to Reuters, Big Data will grow by 45% annually to reach a $25b industry by 2015. And, it's a hot topic in healthcare too; although the focus of many articles on the topic tend to center around the use of data to improve care or disease management, and rightly so. By using data to track underlying trends or causes, the goals are often to improve care, quality and/or reduce costs. What may be missing from the equation though are behavioral trends, and that’s where healthcare marketers data comes into play.
We may not have insights into why someone chooses to take their medicine (or not) for example, but we do have insights into behaviors that are occurring in our hospital’s digital footprint. Healthcare marketers have large amounts of data at their disposal from Web site and digital marketing analytics that can shed light on consumer behavior. If a CRM solution is in use, the amount of data available compounds.
What can marketers contribute to the use of big data to improve outcomes? First, think in terms of what is trackable; here are a few examples:
- Newsletter subscribers – if topical and/or demographic subscription information is available that’s even better.
- Web site conversions, with conversions being visitors who completed an online transaction and who are identifiable - requested an appointment, signed up for a class or event, refilled a prescription online, etc.
- Web site exits – where, when, why, etc.
What can healthcare marketers do with this information? We can target message. The days of marketing to the masses are quickly changing; the more we can message to the individual the better the chances that person or group of like people will take action. Conversely – if we know who took action, we may be able to mine our data to determine who didn’t and why. That’s behavior analysis and may lead you to new insights on how to attract, engage and retain our target audiences. We can determine what’s working effectively, for whom, and take steps to continue improving it. Conversely – we may see what’s not working and use that information to plan changes. There’s a "ying" and a "yang" to using data and sometimes it boils down to asking the right question.
Big data is all about capturing data, hypothesizing about it and analyzing it in relation to that hypothesis. Healthcare marketers have at least three types of data at their disposal, many have at least four. They include:
- Web site analytics
- Social Media analytics
- Other eMarketing analytics (i.e. – email campaigns, mobile analytics, etc.)
- Customer relationship management data.
Each type of analytic has various solutions available to enable measurement, ranging from free to paid systems and services. Regardless of the solution used, keys to successful analytics endeavors include clean data and visualization tools.
What is clean data? Data that has incomplete or inaccurate records removed from a dataset. An example is the Web site awareness measure of site visitors. Depending upon what the organization specifically wants to measure, that statistic is often cleansed to remove visits from within the organization, to capture only unique new visitors or to capture only repeat visitors. As data analysts, we need to think in terms of clean data when conducting analysis and ensure the inputs used are appropriate measures for the question being measured (i.e., the hypothesis) and that those measures correlate to desired outcomes.
Visualization tools involve the use of charts and graphs to depict or predict trends. They are typically available within analytics systems or you can create them yourself using available desktop graphing tools (i.e., Excel, PowerPoint, etc.) Use them! Often it’s far easier to understand information and get our message across when depicted graphically as opposed to sharing tables of numbers. When using charts and graphs to relay a message or insight, pay close attention to the scale used. A relevant statistical change can get lost if the scale used is too large.
And finally, we need to consider “what if?” and “how would we …?” questions as we analyze our data. Those are the hypothesis questions that the analysis of data, big or small, helps to answer.