Nick Valiotti illustrates a mismatch in his post about data strategy: the business asks why customers leave in month three, and the data team replies that the pipeline is stable. He argues that placing data under IT gives it IT's incentives: uptime, cost control, security and system health. Each deserves care. The question is what else the company expects from the data function.
A function inherits the incentives of where it sits and is rewarded. Its placement and mandate establish what counts as doing a good job. Those incentives shape behavior, so a well-meant structural choice can quietly harm a company without anyone intending it.
What the structure notices
Valiotti describes the rationale: data feels technical, IT feels technical, so putting them together seems like alignment. But where a team sits decides what it is measured on, who reviews its plan, what earns praise in a quarterly review, and which failures get noticed.
This is what I mean when I say structure breeds behavior. People do what the structure around them notices. The questions a manager asks in a one-on-one and the priorities defended in a budget review give the mandate practical meaning. A team rewarded for a stable warehouse is being asked to make that warehouse dependable. Doing that well deserves recognition.
The harm is quiet because it looks like competence
Consider a team whose objectives cover system health and cost control, while its executives also expect it to help explain customer loss. The team meets its objectives, and its managers have good reasons to recognize that work. The unanswered business question has no place in the assessment. From within that mandate, the team has succeeded.
This is how the way a discipline is conceived can harm a company unknowingly: every local signal looks like competence. The company keeps rewarding work that leaves an important question unanswered. The favorable review gives it no reason to reconsider what it has asked the function to do.
Change what the function answers for
The reporting line and the mandate need to be considered together. Valiotti describes data strategy as a set of hard choices, including which metrics drive the business and what decision changes when they move. Those choices give leaders a basis for deciding what the data function should contribute and how that contribution will be assessed.
For the customer-loss question, a performance review should examine whether the analysis helped the responsible business leader understand the problem and decide how to respond. Reliability should remain part of the review. The person setting priorities must also allow time for that analysis and recognize the work it requires.
Writing down a broader mandate is a start. It needs to be reflected in the work managers approve and the results they recognize. Telling a team to inform business decisions while assessing it only on system health leaves the same incentives in place.
Look at the last few things your data team was thanked for. If every one of them is about systems, compare that recognition with the contribution the company says it expects from data. That gap belongs in the next discussion of the team's priorities and performance.
Sources
- Data strategy is “IT’s problem.”, LinkedIn (Nick Valiotti)