We’re interrupting our normal schedule this week to bring you some fresh thoughts about someone else’s fresh thoughts!
There is so much in there to comment on, so much simple truth!
Given that here, we are amongst tech-heavy people, I’m going to ignore all the good things about KPIs, presentation, value, and so on, and go straight for the things that strike me most from a technical perspective.
11. All data is incomplete
I’m pretty sure this is totally obvious to all of us, right?
Under normal circumstances data is incomplete, but taking into account the fact that we must live with 36 and 40, we should actually pull our heads out in despair.
That is, until we realise that data doesn’t have to be complete to be useful.
12. All data is inaccurate
Again, we all know this.
We also spent the last decade or so trying to convince our stake holders that no matter, we can still use it.
We talk about “10% is totally normal” and about testing.
Our stake holders, on the other hand, either flatly don’t believe us (“how can you seriously expect me to be ‘bad data-driven’?”), or they use this as an excuse.
Whichever it is, this hurts.
27. Most detected anomalies in the data should not be investigated
This one made me hesitate for a second, but Tim is right!
You have 11 and 12, above, meaning there will be more issues and anomalies in your data than you’d like, plus people on the Internet behave like drunken doofuses, so even if the data is right, you shouldn’t try to catch up with someone who repeatedly tries to run up a water slide.
36. All analytics implementations are flawed
And there are so many reason why they are that it is really unlikely, no matter how good you are, or how hard you try, that yours should be an exception.
Live with it.
37. Most analytics implementations are good enough to get value from in their current state
I think it is important to remember that value is not the same as clean data, good data, or “the right data.”
Value has to do with your stake holders, their wants and needs, and the willingness of a bunch of people to react to what analysis comes up with.
40. Digital analytics data collection is built on a hack of technology/standards intended for other uses
51. Analysts should understand the realities of martech
That is all.