About puzzles and mysteries

When we are dealing with analytics and insights, we mostly are confronted with one of two distinct categories of problems: puzzles and mysteries.

What's the difference?

Puzzles have solutions; they have answers. Whether crosswords or Sudoku – a puzzle is designed to have a sole right answer, and the challenge with it is to overcome the lack of information it bears.

A mystery, on the other hand, cannot be solved – it can only be framed. Mysteries are often characterized by an abundance, even an exorbitance of information.

Some of the questions we are faced with can be answered by digging out missing data from our Analytics tools and reports. "How have we been able to grow our Like base on Facebook over the past six months?", "How has this month's revenue been compared to last month's?" and "Where did people on our website drop out of the sales funnel?" are clearly puzzle questions to which answers can be found from our available data.

But we get more complicated questions asked as well: "How can we increase our average revenue per user?", "How can we reduce churn?", "How can we shorten the lead times?", or "How can we increase engagement on our website?".

It is enticing to look at our past data, and to extrapolate trends over time from there. As seen with Kodak, or Nokia, a stable historic trend in units sold prevented both companies from asking “Are we selling the right products to the right people?”

Making extrapolations from historic data, or embarking on detailed planning for exact interventions involving several disconnected players lets us look at mysteries as if they were puzzles when, in fact, they are not.

Framing a mystery

We can frame a mystery, using questions: "Which additional subscriptions would customers consider buying that would complement (not replace) our offering?", "Is there a way to make it easier for customers to make a purchase decision by reducing their perceived risk?" or "Can we scale our offering so that it appears more attractive to pay a higher up-front price for it?" are questions that may have no definite answers at the very moment they are asked. The point is: once they do have answers, it's much more likely to discover a few potential solutions.

In essence: when we frame a mystery along the lines of constraints, variants and shifts associated with it, we tend to get a clearer understanding of its boundaries. Pretty much all business decisions are volatile these days, so they require both thinking in configurations and the reversal of hypotheses once their validity is refuted.

When we are trying to solve business problems by treating them as mysteries, it helps to ask questions about "What can be changed?" and "What has to remain the same?". Depending on the answers we find to each of these two questions, we can model the contingencies accordingly.

If we’re told a website has a well established and defined audience, we can ask: “Do we reach enough of this audience?” and “What previously unanticipated content needs does this particular audience have?”

This changes our perspective from the solution (of a puzzle) towards a more strategic approach. Even more importantly: it creates a sense of possibility.

In any case, we will have to find ways to constructively question the notion of any hypothesis as being irrefutable. There is a very fine line between 'what we know' and 'what we take for granted', and using the distinction between 'puzzles' and 'mysteries' can help us to understand these differences better.