“There’s a creative moment when you think of a hypothesis, maybe it’s that interest rate data drives” currency rates, she says. “So we think about that first before we mine the data.  We don’t mine the data to come up with ideas.” Lady Braga Steps Out From Behind The Curtain, Bloomberg Businessweek, Mar 2, 2015.

I’ve faced a lot of PowerPoint charts that showed a reality different from what I was seeing.  I couldn’t figure out why these folks felt their data showed anything useful.  Yet here I was with data and trends painstakingly coaxed into showing themselves up against someone’s PowerPoint chart that they had just developed within the last few hours.  It was clear that they saw it as a competition of “can we get people to buy our idea over his idea!”  The notion of if it was real or not was not the concern.  It was all about if they could sell the chart as resonating best with what management wanted.  It was not about capturing what really was going on in the project nor the organization.

For similar see The Trend Is Your Friend

I watched as a VP of development for a Fortune 50 company sat and scribbled on some paper as we sat waiting for a meeting to start.  He was sketching a chart, adjusting the trending with a lot of erasures, until it showed a nice downward pattern over the next few months.  He then pulled out his laptop and punched numbers into an Excel spreadsheet, examining the trend line and how it looked, adjusting it until it looked like the one he had on paper.  He then cut and pasted the Excel chart into a Powerpoint slide.  Once the meeting started, the VP displayed the chart when asked by his boss what our plans were to complete development on time. His boss then asked, somewhat incredulously, “this is what your team agreed to?”  The development VP’s answer was “yes!”  Of course, we delivered months late as the numbers the VP had created on the spot had no basis in any reality.

See instead Honesty Is Just More Efficient

Some companies, including many drug and medical-device makers, already have more information than they know what to do with, says Pratap Khedkar, managing principle of consulting firm ZS Associates. “Everyone in Silicon Valley wants to start with data.  They think, if we get all the data, then magic will happen,” Khedkar says.  More important, he argues, is asking the right questions, which can often be answered with just a few sources.  In Search Of The Golden Prescription Pad, Bloomberg Businessweek, Feb 8, 2016.[Update]

Data is a wonderful thing.  It can give us insights into our team, project or organization that can all but assure that we deliver our projects on time and with good quality.  But getting good data and coaxing out of it the insights we need is often not easy.  For me the best approach was  always to first see the phenomena that is going on in the organization (e.g., always delivering projects late, not fixing defects when promised, etc.) and then finding data that accurately describes that phenomena.  Once we’ve done this we then had truly profound insight that would guide us and be predictive of how our project would unfold.

See why Getting Good Data Is Hard

I’ve had people get tired of seeing my charts and trends because they made management of the project almost automatic, it was obvious what to do based upon the charts. Very little discussion was needed to make a decision. This was in contrast to the PowerPoint charts-of-the-week approach that differed with every presentation.  Nothing in the charts ever guided our decision making (e.g. a stop light chart of issues) and if used again was just a “standard chart to be shown” and ignored in the next status review. Note that it was managers who often disliked my charts as it made a lot of what they were there for unneeded (e.g., making charts to guide projects it seems).

Compare with Successful Projects Are Boring

Analysis of data is very useful when it shows an accurate description of what is going on in the project.  An accurate description just about always has the characteristic of being persistent and predictive of the future (i.e., we can use it to make management decisions). There may be other useful insights when mining data but if we don’t first understand how our project is going, then those deeper insights (e.g., making things faster, cheaper, etc.) are of little use if the project is careening out of control.

“If you don’t have deep expertise in how energy is distributed or generated, if you don’t understand how a power plant runs, you’re not really going to be able to build an analytical model and do much with it”  Bloomberg Businessweek, May 6, 2012

Is your data mining driven by wanting to understand what is going on or is it a random search for anything that just sounds good?