The Psychology of Business Intelligence Implementations: Tableau and the Culture of Yes

Erik Erikson had a rock star name before his time. Born today, I imagine he’d front a band that is both lyrically profound and acoustically disturbing. Instead, Erik picked his new name and decided to go into psychology. Now we know that celebrities AND psychologists alike prefer aliases.

We know of Erik Erikson for his stages of development. He considered various ages different battles of how we respond to our environment. Naturally, they build, and of course, you can fall off a cliff into all the sad stuff. That’s where those profound lyrics and disturbing acoustics come in.

Business intelligence knows this battle well. Historically, BI has decided to mistrust all but a very few. You can guess how it matures from that point.

That is, until Tableau.

Then you saw analysts doing devious things like connecting to Excel data sources.

Or running Tableau server on machines stealthily hidden in the corners of cubicles.

Or, dare I say it, quickly exploring data during a meeting to come up with an answer to something.

It’s not that we couldn’t do certain things before. It’s that mostly it required 900 miles of red tape, 17 signatures, and a massive, massive disclaimer that ‘YOU SHALL NOT ___’.

Tableau starts from a position of trust and autonomy. It believes analysts should be able to take initiative and explore the data. Moreover, it rewards an industrious nature – can you make a dashboard? Can you solve this problem?

It supports people who proudly wear the badge of analyst, even for those who do so part time. It came from the idea that maybe, just maybe, analysts weren’t wholly nefarious with the numbers and that they, too, could put out accurate renditions of business information without a tightly controlled source, stringent requirements, and predetermination of the solution.

Greater insights can help pull people together. It can unify the vision of what data is and what we could be doing. Instead of operating from a rules-based culture, we do things because – gasp – that’s just who we are. We achieve greater things and we’re all happy.

OR…

We could choose the other side of the fence. We can say no, we don’t trust and go down from there. We can even start with trust, build some beautiful things, but then introduce role confusion or start isolating people from the data.

Does this sound a whole lot like psychology?

It starts with the data…

First, we’ll dig into the original history. In the days of yore (when Tableau was first made and logos were far more blocky), there was only a desktop client. You could connect to a few sources and, you won’t believe this, it was frightening easy to do so. You could then connect live or extract.

Then, you started making charts.

For anyone familiar with legacy BI, you’re noticing there was minimal data prep mentioned here. Tableau, built on the assumption I’d have access, didn’t require a whole lot around metadata management. Sure, I could get a view made or, if I was lucky, I’d have some read-only access to a database and real-live fact tables.

Or, I could even be sneaky and connect to Excel.

Trust, Autonomy, and Initiative

Tableau did NOT start off with the idea that someone else needed to limit and make all my data sources. Even when server first came out, it acted more like a repository FOR data sources than a massively tight control mechanism. That’s not to say security is lax, but it does start from a position of trust with ‘NO’ being the second answer – not the first. It supported my autonomy and rewarded my initiative, letting me connect to various data sources and publish.

Even as Tableau continues to deploy more rigorous controls like certification, it still provides mechanisms for analysts to, say, make a calculation without begging a department to do so (this, kids, is the types of pains cubes inflicted – you want that counted, averaged, or anything else? ONLY the cube can do it). With Tableau, I can be industrious, because I’m trusted. This isn’t about control versus anarchy, but about balance between the two.

When we trust our analysts to understand how to translate between numbers and business vernacular, we get faster, and even better, insights. We escape the reporting trap because we don’t have to fight to get the data. Usually…

Of course, you can run Tableau like a traditional BI implementation. You can check all the ‘no’ boxes and lock it down. And, yes, sometimes, there’s reasons to do so. But, mostly, it goes back to those Erik Erikson ideas of trust vs mistrust, autonomy vs shame and doubt, and so on. So, before you DO lock it down, ask why.

We are empowered to say yes with Tableau. We can choose a balance, rather than a wall, for deploying Tableau. And that is music to my ears.

(This song is both disturbing and profound. Enjoy!)