Separation of Church and State
I occasionally find myself justifying why I prefer a centralized structure for analytics/business intelligence/science teams within a company. There are actually numerous benefits including team morale and team retention rates that can be found in white papers such as the ones by Accenture in 2015-2016. However, there is another important factor for why your data professionals should be reporting into an independent team and two recent pieces of news have made me want to discuss this topic.
In Florida, a person(s) of authority appears to want facts around COVID-19 to tell one story but the scientist involved did not find that story in the data. That scientist was removed.
In another case, it appears there is a scientific team within a company that was operating under an assumption they were free to publish scientific research regardless of how the research portrays the company and its work. Turns out the researches may be more censored than they realized.
These are complex stories with multiple sides, but what we see in both appears to be a conflict of interest. The data or research paint one picture. The people in charge do not like sharing that picture*.
A business would be foolish to think this conflict of interest between data-derived facts and people in charge is never going to be a problem for them. It will. At times the problem will be conspicuous with clear discrepancies emerging in the metrics and KPIs as the manipulation of facts proceeds. More often it is very subtle, such as KPI reports or tracking efforts simply being delayed or reprioritized in perpetuity. I’ve seen both types first hand and so have many of the experienced data professionals I’ve spoken too.
The best defense against this is simple: separate the teams performing a function (e.g. Sales, Marketing, Engineering, etc.) from the teams assessing how that function is going. If the truth about your company is important to you, make sure data containing that truth can be analyzed and surfaced by analysts without fear.
In short, don’t shoot the messenger!
*This, by the way, is why tenure exists at research universities; to protect the job of a professor who publishes something unpopular with the university or its donors, not so a professor can be lazy as is often assumed.