“In The Wrath of Khan (1982), Spock says, “Logic clearly dictates that the needs of the many outweigh the needs of the few.” Captain Kirk answers, “Or the one.” This sets up a pivotal scene near the end of the film (spoilers follow). … Spock replies, “Or the one.””
And so with data (and analytics) governance. The needs of the many data outweigh needs of the few data, or the one. This is why historical so many data governance programs start with:
- A major investment with a consulting firm
- Many interviews with business leaders to gauge what is ‘wrong’ with data and to identify all the problems with data
- An assumption someone has to ‘sell’ the work of data governance to business folks
- An assumption that the work of data governance is about data
- An assumption that the work of data governance is new work anyway – never before even worked well
- An assumption that a focus on reusing data and data standards or data standardization is central to success of a data governance effort
- A need for executive support
And these are reasons why:
- Such programs never survive engagement with every day business (the real world)
- They never last beyond 3-7 months
- They seem over engineered
- They don’t seem relevant
- They seem to slow (everyone invited to the first, bloated, meeting
- Firms repeat the effort every 18 to 24 months or less (since the problems remain systemic)
I was talking with a client that other day. We looked at this:
- You have 100 bits of data to worry about
- You have $100 to spend to govern your data well enough
So you plan on spending the $100 roughly equally across all 100 bits of data. That is the wise words of that consultant; that is how the vendors suggest you spend that money. You do need to spend all $100, after all. Surely that needs of all the data (the many) outweigh the needs of the few (any data)…. wrong.
A modern D&A program recognizes how we behave and also that:
- Not all data is equal
- Some data is more important than other data
- Some data matters most
So really you should spend about $70 on the 10 bits of data that matters most, $20 on ~20-30 more bits of data that letter between slightly and a lot less than the first 10 bits; and the remaining $10 on the remaining 60 bits of data. Those are roughly estimates but do you get the point?
The needs of the whole (business) will be met by meeting the needs of the fewest data well, and by meeting the needs of the most (data), least well.
Where is Spock when you need him?