Recently, I had the opportunity to provide some Business Intelligence (BI) training to someone with very limited prior exposure to IT and database systems. I was using whatever pointers I could provide to help this person ramp-up for a new job as an analyst. In the process, I gained a renewed appreciation for the power of metaphors for bridging knowledge gaps. As I will discuss, metaphors can be very useful things to collect.
In our conversation, the relevant topics were relational data models and querying, but instead of just introducing rows, columns, tables, keys, etc., my goal was to also show him how to “think in data sets.” This didn’t prove to be an easy task. I started by asking questions that would help identify prior knowledge and serve as building blocks for the discussion. There was a solid Excel foundation, so we opened a blank spreadsheet and sketched a few rows and columns to build a data set that would look familiar to him. Turning the discussion to relational concepts, he seemed to have a grasp, but when it came to a question like “what query asks how many dollars were spent in age group X?,” something critical seemed to be missing. It didn't take long to find the missing ingredient, which had to do with the dimensional role attributes play when creating aggregates out of other numerical fields. After several attempts, I realized the concept was foreign to him, and my explanations just weren't doing it.
Metaphors are perfect for these kinds of situations; pretty soon, I found myself talking about a population of people (rows) within a closed city (table), which had a set of designated sensors (fields), including one sensor that kept track of everyone’s gender, another that tracked everyone’s age and so on. With this metaphor, it became much easier for him to think of the query as asking the age sensor to group the people who fell between two boundaries (where clause), then sum up the total dollars spent by the group by getting each of their spend amounts (aggregate function).
Over time, I've managed to collect a list of favorite metaphors, and I'm always on the lookout for new ones. I've used them when teaching classes on BI software platforms, but just as often, they come in handy when working with a project team.
We tend to think of metaphors as lending creative and artistic meaning, taking for granted their practical utility. Consider recent evidence in cognitive science suggesting that our brains actually acquire much of our knowledge via metaphors, often without us even realizing it. Or, consider that complicated problems in science and engineering domains haven’t only been explained using metaphors (just do a search on the Higgs boson particle for plenty of recent examples), but also that metaphors have been used to develop new ideas and discoveries (both Einstein and Darwin cited the role of metaphors in developing their scientific ideas). Metaphors do this by turning a complex and unfamiliar process into one that can be approached more intuitively using existing knowledge, or what some would call a simplification. Good teachers are metaphor masters, building up their own collection of perfected metaphors, which they have many opportunities to refine with trial and error.
To develop professionally is to learn how to work more effectively with those who don’t share one’s specialized knowledge. This is true not only because, fundamentally, technical decisions impacting business processes or requiring investment should be discussed within cross-discipline groups, but also because the fields comprising BI, analytics, data science and many other related fields are quickly expanding. This expansion is also creating communication challenges as the new people try to quickly catch up and gain fluency in data-speak. It will help to be on the hunt for a few good metaphors so they will be there for you when they're needed. Then comes the challenging part: learning to use them just the right way—and only when needed—so people don't feel you're dumbing-down the conversation unnecessarily. Happy hunting!