Data is useless

Data stories, on the other hand, are very important, as most businesses are now drowning in data, but still haven’t learned how to swim

drowning big data 2

When Alan Turing and his colleagues cracked the Nazi Enigma codes during World War II, the impact on Britain's war-planning was minimal. The reason little changed is because Britain's armed forces simply didn't know how to take advantage of the abundance of data they now had.

The military's operational systems were built for a world of information scarcity, but now a bunch of civilians were flooding the system with details of every U-boat position and every troop movement, and war planners didn’t know how to integrate all that information.

bundesarchiv bild 183 2007 0705 502 chiffriermaschine  enigma Bundesarchiv, Bild 183-2007-0705-502 / Walther / CC-BY-SA 3.0

Enigma Machine in use

Businesses today find themselves in a similar predicament. Data used to be hard to collect, expensive to store, and slow to analyze. All that has changed. Most businesses are now drowning in data, but still haven't learned how to swim.

Everyone is embracing data because, well, right now data is sexy. (And big data? Big data is really sexy.) But businesses are quickly discovering that they don't actually want data. They want information and knowledge they can use to make better decisions.

So how do we get there? There are three components.

First, businesses today will only realize the enormous value that data offers if they're willing to find new ways of working. Just as Britain reconfigured its military bureaucracy to efficiently get information to commanders in the field, businesses will have to adopt new processes. Just adding data to old processes and hoping it makes them better won't work.

If business leaders refuse to adapt their methods and try to jam new information into stale processes, the result is just going to be frustration -- frustration that new systems aren't yielding the promised returns and frustration from analysts that their work is being wasted.

Second, teams need to use the right tool for the job. They can't become beholden to particular tools and let the tools dictate what's possible. With the advent of MPP data warehouses, machine-learning algorithms, web-based visualization libraries, and the Hadoop ecosystem, the array of tools that data analysts and data scientists have to familiarize themselves with has grown enormously. But with that huge array of tools comes the risk of fetishizing the tool rather than focusing on the goal.

Data teams need to remember that sometimes a linear regression works just as well as a random forest or a neural network and that it doesn’t matter which you use, as long as you get the information you need.

And that brings us to the third component of getting from data to knowledge: data analysts' responsibilities. Analysts have to take their analyses that last mile, from interesting to useful, from analyses to data stories that answer the business question that was asked. Because too often analysts find data that is fascinating to them and hand it over to co-workers, forgetting that data is useless.

I say that as an analyst myself -- as someone who loves data. I love digging around in it and finding hidden patterns and telling others about them. And it's hard to admit that something you love is useless.

But over my decade-plus working with data, I’ve finally come to terms with the fact that I’m not normal -- that my unconditional love for data isn’t shared by most humans. And so handing over a table of numbers to a co-worker without much explanation because "I don’t want to prejudice their interpretation of the data" does a disservice to us both.

Analysts need to take responsibility for integrating their findings into a story that makes sense to their "customers." Because most humans think in stories, not in data. I realize that spending time on this last step can be annoying -- thinking about the most intuitive way to graph data, labeling your axes well, writing up your interpretation in plain English, and all the other things that go into telling good data stories take time.

Portrait of Alan Turing

Alan Turing in 1930

But it's time well spent. Because when you skip that last step, you often undercut all the good analysis you did. Good analysis presented poorly is just as useful as bad analysis presented well.

And Alan Turing figured that out. His team put aside the most interesting intellectual cryptography problems to focus on the tasks that would most help win the war. And when they couldn't get the resources they needed, Turing wrote to Winston Churchill with a story about what they could accomplish with the right resources.

Churchill made sure they got those resources, and Turing's team played a critical role in helping win the war.

Because as Turing realized, data is useless. Data stories aren't.

Copyright © 2016 IDG Communications, Inc.