Fewer Dashboards, More Analysis

It pains me to say it but Business Intelligence has gotten itself a bad rep. And we're all partly to blame.

It’s pretty clear that to most people in the data science world, BI means building endless reports and dashboards - and that’s all it’s good for.

Take a look around your own workplace. I don’t care if it’s a bootstrapped start-up or a global corporate giant. Odds are that the virus-like spread of the BI dashboard has reached epidemic levels. BI teams are stretched thin building more dashboards than the Tesla factory. And the requests just keep growing.

The business teams are demanding more and more and more. “I need a new sales report”, “I need another traffic report”, “I need another hundred metrics on this dashboard or else it’s useless and I can’t run my business and it’s all your fault and...and….and….”

Someone please wake me when the nightmare is over.

WE. NEED. TO STOP. FEEDING. THE. DASHBOARD. ADDICTION.

It’s not that I’m against providing valuable insight to management based on the data we have at our disposal. Quite the opposite.

I’ve spent most of my working life trying to deliver better insights through data-driven projects. I might get jaded at times when it feels like I’m banging my head against the metaphorical (and sometimes physical) brick wall. But I’m not suggesting we throw the baby out with the bathwater.

Gut feel (or what should more correctly be termed Subject Matter Expertise and Experience) alone doesn’t work either. We need a joined up approach that brings the two worlds together.

I want everyone to slow down on the “drive to dashboards”. Endless reports filled with countless metrics are not the best way to get full value out of either your data OR your data team.

Benn Stancil, the Chief Analyst at Mode, wrote an excellent article called Don’t Choose Dashboards Over Analysis.  Benn says:

“Though it sounds counter-intuitive, more dashboards often make people less informed and less aligned...a company with hundreds of dashboards can’t focus on anything.”

And he’s dead right.

I see some dashboards so jam-packed with metrics it’s difficult to even work out what they contain, never mind what they are telling us. The "More Data/More Dashboards" game-plan really only provides a comfort blanket for indecisive managers. They really want an inanimate object to blame if they make a decision that doesn’t quite pan out.

It’s a symptom of poor company culture that a scapegoat and sacrificial lamb have to be prepared well in advance of any business choices happening. But that is often the choice junior and middle management have put in front of them. They carry all of the risk but little of the responsibility to really move the dial. So it’s easier to sit and play spreadsheet games with the moving numbers on the BI dashboard instead.

How does incessant dashboard building impact the data team?

My philosophy is that data analysts should be analysing data. Not blindly building more and more reports. It’s our job to analyse and, most importantly, interpret what is happening in the business. Spotting trends and investigating the causes and possible drivers is where the real value is gleaned from having a strong data analyst team.

If all you have them doing is building mega-dashboards which mash up and squeeze in hundreds of different figures, you’re wasting their time. Every second is vital and the opportunity cost of that time sink is the inability to get focused thinking time to properly analyse the real data at their disposal.

What gets measured, gets managed.

But does it really? Basecamp CEO Jason Fried spoke about his views on this oft quoted business mantra when he appeared on Tim Ferriss’s podcast (episode 329). Jason suggested measuring less. He tests by removing elements of what he does measure then sees if it makes any difference to their business performance.

It clearly hasn’t had a negative impact on Basecamp’s success over the years so maybe Jason is really on to something.

I advise business people to really focus on no more than five key metrics if they really do have to measure their KPIs. I mentioned a senior exec in another article who was only interested in one key metric. At her level that was all of the management info she needed to see how her division was working. I balked at the time when I heard the story but understand it better now.

Less is more. It’s OK to not look at everything all of the time lest we get snowblind, as Benn Stancil suggests in the quote above.

I’ve seen this from both sides.

I’ve been a data analyst for many years. And I've worked very closely with business stakeholders to try and understand exactly what they think they will get from their latest round of MI report requests.

Conversely I’ve ran businesses where it’s easy to get caught up obsessing over small details. There is always the anxiety that you’re missing something important if you aren’t proactively measuring EVERYTHING, every second of the day.

Real-time reporting feels like complete control but it comes at a price. You gain a close-up view but lose the big picture.

There has to be a trade-off.

As an analytics manager, it’s your job to talk the business side down from the idea that they NEED to see everything all at once and all on the one page. They don’t. They need to focus on what really moves the dial for their area of the business.

As an analyst, you need the space and time to dig into the data and ask it better questions. You're wasting everyone’s time if you're on a constant treadmill to deliver more reports that will rarely get looked at. And no-one needs that realisation.

If you're in the business owner’s chair you need to take a deep breath and take the advice I give when working as your data consultant. Focus. Don’t spread your attention too thin. And use the data to lead you to a place where it’s telling you a story, not just showing you more and more numbers on a screen.

Getting to that stage needs the ever useful help of the data detectives in your analyst team. But you’ll never know if you keep them tied up on those dashboards and comfort blankets all week instead. Free them. Utilise them. Or you will inevitably lose the best of them.

The choice is yours.

If you enjoyed this article you should sign up for my weekly newsletter at SimpleAnalytical.com. You’ll get advance notice of new articles, along with a veritable treasure trove of added data-related bonuses. Cheatsheets. Ebooks. Code samples. And more.