Why Doesn’t My Data Listen to Me?

Apr 19, 2012 by Phil Simon in Data Migration, Data Quality

One of my favorite Porcupine Tree songs is “I Drive The Hearse.” Here are some of the song’s haunting lyrics:

And silence is another way

Of saying what I wanna say

And lying is another way

Of hoping it will go away

I am often reminded of these words as I look back on many of the IT and data migration projects on which I have worked over the last ten years. More than a few of these informed my first book, Why New Systems Fail.

On one particularly contentious project, I was on a team of consultants replacing a hospital’s disparate systems with one unifying ERP application (not that everyone at the hospital actually wanted the new system). For a bunch of reasons, things weren’t going well and it appeared that the project would take at least six months longer than initially anticipated–costing more than $1M (USD) than budgeted. Things were getting testy.

Note to project managers (PMs): Screaming at your consultants isn’t a terribly effective management technique.

Your Data Won’t Listen to You. Ever.

During late-night meetings discussing problems, assigning blame, and plotting theoretical next steps, never once would the client’s PM (call her Barbara here) talk about data quality. Not a single time. Ever. More disturbingly, she would quash consultants’ suggestions that her latest master plan be altered because of pesky data issues.

In her demented view of the world, data was supposed to just “fall in line”, as if her decree would do the following:

Correct years of sloppy data management
Magically convert and cleanse tens of thousands of questionable or errant records
Ultimately get the project back on track

As readers of this blog know, data just doesn’t work that way, not that Barbara wanted to hear about little things like reality. All of her threats and brooding didn’t correct a single data-oriented problem. Not one. Barbara’s screaming and yelling ultimately had no effect on the data. It only served to annoy the very people who had to fix or cleanse her organization’s compromised data.

Simon Says

I have written before about the voice of data and, trust me, data only stays silent for so long. At some point, bad or questionable data will make its voice heard. And its voice is much, much louder than the histrionics of the most irritated client or consultant.

The question is not if – it’s when – your organization’s data will speak up. And not saying anything counts, especially when the data should say something. To paraphrase from Simon and Garfunkel, the sounds of data silence can be deafening.

Feedback

What say you?