Among the group was a friend I'll call Jane, who runs business research and insights at a major media company. She is intimately and credibly versed in the strengths and weakness of the various partners that support her in her role of providing insights to every single property under her roof. Discussing Nielsen's major correction last week of bad performance data on major media properties led to a chat on what she regards as an ugly truth: online data is bad all the time. In short, this correction was deemed necessary after a glitch in the system revealed an error of the type that Nielsen assures will no longer happen after its much anticipated overhaul to a new system this summer. But, the conversations I had over the weekend shed light on the idea that such instances are not so isolated. Though, they are not always so grand in stature -- involving properties such as Fox, the New York Times, and USA Today to adjust their numbers -- corrections are issued all the time by parties and players across the galaxy that is data and analytics. In a word: terrifying.

The High Stakes of Progress

Last week, I discussed analytics as a field still warranting a smart and righteous guide. Though many see this age of metrics as "golden," it holds pockets of things we still must understand and standardize. It was interesting to consider for a moment the more primal alert raised by Jane: that the industry may not care about accuracy in the same way that its predecessors in magazine and television always have. That a certain nonchalance about errors prevails; that corrections are frequent, a constant agitation to companies that care about getting it right.

Most of us can recall reading corrections in our own daily newspaper that were so deeply alarming, yet nestled as barely a mention on the corrections page, so as to diminish impact. We ask, "What?! How can this be? Are people seeing this?"

So, during tough economic times when business courses must be charted evermore shrewdly, and data companies undergo their own system advance and methodological overhaul -- can we really afford these cycles of crude and constant correction? My friend would say "no." I'm just starting to really think about it, provoked by this chat over the weekend. I would really like to hear from those who are passionate about the field -- what's the real scope of this issue?

The Pain and the Promise

Yes, you might agree with Jane, who believes that online dominates the cross-media sphere on available data and on its dexterity. I know I do. But, do we agree it's a lot of bad data that people only apply sparingly and not always strategically? Jane in fact credits her traditional-media roots for her devotion to quality and accuracy -- that's where she gets good stuff she can trust and on whose back she can plan or direct her properties to follow suit. This foundational strength frees her up to focus on application and analysis. She now faces this deep internal struggle as she integrates her own practices to encompass digital -- given her almost total discomfort with digital data integrity. Does this seem like a lot of drama -- or does it ring true?

While we might waste time trying to decide whether any given outcry over issued corrections is a "shoot the messenger" issue or not, every time one of these corrections is delivered, it seems important to encourage a reminder about the purported sweep of this issue. We must remind ourselves that one person's rounding error might be another's death knell. That we'd do well to take a step back and align on a primal commitment to accuracy. In a world that is large and complicated, such focus on accuracy is a big ethos to drive. But, we cannot really afford flippancy when it comes to corrections. That is at least one true thing.