Metamusings on Life, Business, and Data October 2013

Many of us in the data management space have been involved in an Agile project, where often there is a clash of cultures between a more traditional “design then build” approach and the “responding to change rather than following a plan” approach of the Agile Manifesto.  One of the themes and tenets of Agile is to create models and software that are “just good enough.” Rather than working to perfection, it’s important to build just enough to keep going and get the project completed, with the ability to rework it later as need.

If you’re like me, you might have a problem with that.

I’ll admit it, that’s been a piece of the Agile approach that has bothered me.  “Really?” I think, “You’re planning to make something sub-par on purpose?  And you’re actually admitting that publicly?”  OK, after thinking this through beyond the first impression, I do understand it in theory.  Not everything needs to be perfectly designed.  A prototype is a prototype, and an internal presentation might only need to get the facts across—not win an award for PowerPoint design.  I get it—don’t spend time on things that don’t need it.  But deep down, I admit, I had this inner drive that my “good enough” will be better than others’ “good enough—sneaking extra hours to make that model or presentation “just right.”

But a recent experience brought home to me how insidious this tendency can be.  As some of you know, I live in the mountains of Colorado.  About seven years ago, I bought some land near my house that came with a mule shed.  Yes—a mule shed—an old-fashioned shed built with logs and metal sheeting that once housed a real, live mule.  (Don’t go there, yes, I live in the boondocks.)  With the old logs and lumber, it was quaint, and I had grand designs of fixing it up into maybe an office or guest house.  Seriously—I bought lumber, skylights—the whole deal.  After starting this project, I realized the momentousness of the task.  Nothing in that shed is plumb or level—it’s been there for about 75 years.  So I gave up on the project, and the lumber had sat in the yard ever since.

Several weeks ago, sick of that lumber sitting in the yard looking messy, in a fit of anger, I just nailed up the boards—put them up without using a level—just “winging it.”  No design, no sanding, no planning—nothing—just holding up the board and nailing it in.  And…it turned out looking really cool.  The rough-in nature fit perfectly with the “mule-shed” theme (a term I might just coin in home design circles).  Best of all, now I have a storage shed—one that I could have been using for the last seven years.

Now I wouldn’t recommend this approach if you’re a professional contractor.  It’s important to know when a project is a mule shed, and when it’s a home you’re planning on living in for another 50 years.  And it’s similar in data management.  In some cases, data quality needs to be near 100%, and projects are mission-critical needing strict attention to detail.  But many others are “mule shed” projects that need to be done “just enough” to move forward with an initiative.  We in data management, data modeling in particular, have often been criticized for treating every data model design as a work of art—requiring months and even years to complete.  While that’s sometimes necessary, just as often we need to be more tactical in our approach, and build incremental models that show rapid, iterative progress.  Else we risk going seven years without a usable shed, or a valuable data management initiative that can make things “better”, if not “perfect.”  Don’t mistake a mule shed for a Taj Mahal.  In many cases, “good enough” can be good enough.

Share this post

Donna Burbank

Donna Burbank

Donna Burbank, is a recognized industry expert in information management with over 20 years of experience in data management and enterprise architecture. She currently is the Managing Director of Global Data Strategy Ltd, an international data management consulting company. Her background is multi-faceted across consulting, product development, product management, brand strategy, marketing, and business leadership in organizations such as CA Technologies, Embarcadero Technologies, and PLATINUM Technologies. She has worked with dozens of Fortune 500 companies worldwide in the Americas, Europe, Asia, and Africa and speaks regularly at industry conferences. She has co-authored two books: Data Modeling for the Business and Data Modeling Made Simple with CA ERwin Data Modeler r8. She can be reached at and you can follow her on Twitter @donnaburbank.

scroll to top