Note: this is Part 9 in a series of blog posts about becoming data driven in business. You may want to read the prior parts before reading this essay.
“This means there can be different operational definitions for the same thing. It is not a matter of being right or wrong, but rather a matter of agreement between parties involved.”
This is the essence of the post for me, and applies in so many different areas. It’s not right or wrong, but just getting agreement on what we are talking about. And people don’t realize they have implicit assumptions of what is “normal” and therefore don’t communicate what they are doing because it’s “obvious” and then all sorts of confusion ensues.
This comes all the time in management or leadership, where the leader asks for something, but doesn’t specify concretely what they need and how they will use it, so their team scurries off and delivers what they think is being asked for, and it doesn’t serve the intended purpose (that was never communicated!). This is how your failure mode #2 occurs where data gets repurposed because an analyst gets asked for something, and they grab the data they already have, because nobody specified how it would be used.
In summary, clear communication where you unpack your assumptions and intended usage leads to better results, not just in data but in any interaction between people (at least until we have mind reading).
Actually when I was writing this piece, I was thinking about the bit in our podcast interview when you said “the first big project was to determine what revenue [at Google] meant.” God, what a series of drudge-work meetings that must’ve been. Drudge work, yes, but also so important.