In an interview today National Public Radio reported on home construction data that the U.S. Commerce Department had been reporting erroneously for ten years. Not a sexy topic, but read on, since it addresses the whole subject of why we bother to collect data, and whether we need to spend more of our
In the interview, a prominent business economist – Mark Zandi of Moody’s Analytics – observes that because of “a processing error in the tabulation of data,” the Commerce Department recently revised ten years (yes, that’s YEARS!) of data on construction spending. Zandi goes on to argue that this is an example of just why the government needs to spend more money on data collection.
I’m not sure I agree. In order to justify collecting data, those data must be:
- Accurate. Need I say more?
- Comprehensible. More on this below – it’s important.
- Meaningful. The data need to help you make a better decision.
- Worth it. The value of that better decision – as measured by, say, higher revenues or lower expenses – must exceed the cost to collect the data.
If any of the four links in the above chain is broken, the data collection effort wasn’t worth it. And the Commerce Department wasn’t publishing accurate data. But even if they were, I have to ask: Was it data worth collecting? After all, Commerce had been publishing this data for ten years and nobody noticed the errors. Was anyone actually using the data to make really important decisions? Did anyone make any bad decisions that they wouldn’t have made but for the erroneous data? And if we had spent more money, would we have gotten better data?
Point #2 – comprehensibility – gets less attention, but it’s important. In fact, I wrote an entire book on the subject of making your numbers understandable, how to achieve that, and why it’s important. The worlds of business and government are littered with great ideas and powerful information that failed or were ignored because nobody understood what they were being shown.
Just out of curiosity, I went to the Commerce Department website to glance at the underlying data. Click here and here for a couple of the spreadsheets they publish, and judge for yourself. While they may be neat and consistent, I find them to be mind-numbing piles of rows and columns that only someone who works with them week in and week out can make sense of. Commerce can do better.
Yes, I know that good data is the lifeblood of good decision-making. And I know we’re in the new world of of immensely powerful data collection technologies like Big Data (click here for my recent post on that). But before we spend a whole lot of money collecting even more data, let’s make sure we’re doing a good job collecting and explaining the data we have now. And I hope you’re all having a great New Year!
“Painting with Numbers” is my effort to get people to focus on making numbers understandable. I welcome your feedback and your favorite examples. And see more about my writing and “Painting with Numbers” at www.painting-with-numbers.com.