Don’t Let Business Intelligence Make You Dumb

I love Business Intelligence. I love hard numbers and charts. I love the flexibility it provides and the problems it helps solve. I am a Business Intelligence junkie. However, for all my enthusiasm about Pareto charts, scatter plots and what-if analyses there is one saying that reins it in heavily.

“Just because nobody complains doesn’t mean all parachutes are perfect” -Benny Hill

This reminds me that Business Intelligence requires, well, intelligence. Earlier in my career, before I learned not to say the first thing that came to my mind (many would argue I still haven’t learned), I was interviewing for a job. After 6 grueling hours of technical interviews it came time for the HR interview. The first question posed to me was “If I were to give you an elephant what would you do with it?” Reflexively, I blurted out “I would ask it better questions.” As glib as it is, that is the craft of Business Intelligence. It’s not in architecting a scalable schema, creating a flexible logical layer, or even in creating pretty charts, all those are required and help, but the true craft in Business Intelligence is in “better questions.”

There is no shortage of companies that got seduced by the easy answers business intelligence could give them and were lead down harmful paths. Most of these fell into some common pitfalls. Here are a few tips on how to avoid them:

Know your audience/keep strategy in strategy and tactics in tactics

Executives are great and handy, but they will never be as “close to the road” as the people that report to them. Likewise, mid level managers are great and handy but they will never be as “close to the strategy” as the executives. Too many dashboards and reports go unused by the majority of the intended audience because they were built to specifications from the wrong level.

The Future of Big Data

With some guidance, you can craft a data platform that is right for your organization’s needs and gets the most return from your data capital.

Get the Guide

While both strategic and tactical involvement is crucial, it is often easy to make the mistake of building too “top heavy” or “bottom heavy” dashboards/reports. Often times these mistakes are made because there’s simply no one else to give tactical or strategic direction. A good remedy is to set the strategic vs. tactical boundaries early. Ensure that every intended audience group has an engaged member that can help with crafting reports that are meaningful to the group they represent.

People react when measured. Think things through.

Many companies have experienced blowback from forgetting to think through their reporting implications. For example; a company deciding call wait times are too high and measures their employee on physical number of calls can lead to less customer satisfaction as support reps will be incentivized to have the customer call back for additional issues. Or a company that tracks to “reported wins” instead of reported revenue will see an increase of reported wins but not increased revenue. Or a company that gives credit to the first person putting in a lead will see an influx of terrible leads being entered as people spam the system trying to be the “first one” in.

Newton once said, “For every action there is an equal and opposite reaction.” The same holds true for Business Intelligence. Every measure you impose will be studied, worked around, stretched and gamed. Ergo, before imposing any measures be sure to think through how whom you’re measuring will realistically react and take action accordingly.

Strive for uniformity in reporting

It does little good to collect measures if everyone is going to interpret them differently. Foster a strong unified approach and set of uniform definitions. Just like data architects have data dictionaries, business intelligence users should have a data glossary with definitions of the measures, the formulas behind them and what is used and where.

Beware the excel jockey, they can derail progress by manipulating data in excel to their own interpretation. Certainly there is a need for deeper analysis from time to time. However, it should be carried out within the agreed upon rules set in the data glossary. There are few things less useful than to spend the first half of a meeting about data arguing about how the data was gathered and analyzed. Even if most worries can be put to rest the data will always be suspect to some people.

Eliminate “Elephant Questions”/”Business Trivia”

In the interview when the HR manager asked me what I’d do with an elephant I knew why she was asking, but there wasn’t very much value in it. Whatever my answer was, it didn’t have much to do with whether or not I could do the job. Just because data is interesting doesn’t necessarily make it useful.

I remember sitting with a sales vp once and going through the set of reports he and his subordinates used. He had to scroll past no less than 9 other reports before he got to a report that meant anything to him. Intelligence is only a benefit if it can be put to use, otherwise it’s just trivia. So too the goal with business intelligence should be for anyone to see a report, know what it is, what it means, and be able to take action on the data. Otherwise it’s not business intelligence, it’s business trivia.