Recently I ran into a Medium post by someone I worked with a long time ago. A very tenacious self-promoter and now an analytics guru, R did a good job of putting her LinkedIn network to work and that's how she showed up on my feed. She edifies readers on how App makers should consider the interests of the target demographics before building an app, some thoughts crossed my mind. R has basically taken three or four simple graphs produced from aggregated App download data and explained what each meant. Are we so mentally lazy these days that we need a bar-graph explainer? R also high-lighted the data points of note in a five paragraph article and concluding remarks to make it worth the five minutes it takes to read this thing. For some reason this is insightful to people and so she is busy speaking about her topic in conferences and doing webinars.
It is no surprise that clients chase after insights from data all day long and yet cannot take real action on it. In the real world, where the likes of R can't fit the graph to tell the story they need to tell, data can be made to tell any story. It all depends on who is telling it and what drives them to align with a certain narrative. I recall a simple topic and sentiment analysis work my team once did for a client based on their customer survey data. This is a fairly controlled situation where the set of questions was fixed though the answers could one of few options or a limited amount of free text. Technically a very benign project and the more serious geeks could not be bothered to deal with it - too yawn inducing for them. The client had a pretty decent sense of what their customers thought of them - the strengths and weaknesses of the brand so they were eager to learn more from us.
Once the data was put through it's paces, we did have some pretty solid looking charts and graphs. Someone had to come up with the supporting story-line for the read-out. We put L in charge of this job given his knowledge of the domain and also because he told a damn convincing story. Many past successes made him the clear choice. Once he was done, we looked it over, and everything computed ever so perfectly. I recall us having the collective feeling of how amazing it was to have truth teased out from the minds of the survey-takers unbeknownst to them. L was pretty happy and was ready for show-time.
The show ended up being a real disaster. Every time the client questioned L and team why this and not that was how a certain graph should be interpreted, we had no incontrovertible fact to support our answer. Given that the foundation was in question here, none of the insights or recommendations could remain credible. The client's perspective was just as valid as L's and they had five times more experience in the business and function than he did. So that was the fate of a our much anticipated client read-out after six weeks of work. They paid, said thanks for all the effort and never took any of the recommended actions. It was a pretty humbling moment for us all. R, I presume is yet to arrive at that point in her career.
Comments