Author’s note: Data in the context of this piece is in reference to business intelligence or analytics being used for internal strategic purposes.
People brag about being Data First.
They talk about their terabytes of data and their product managers that know SQL.
Being Data First isn’t cool anymore.
Having 100 dashboards is a waste of time.
Just because we can’t measure it doesn’t mean it’s doesn’t exist.
Most of reality is unmeasurable.
Knowing how to read a chart isn’t a substitute for experience.
I am Data Last.
Here are the three laws of Data Last:
All data is wrong but some is useful.
Data is not critical.
Data must not override common sense.
All data is wrong but some is useful.
Data is an artifact of some human process.
Code is a human process, too.
I guarantee the tool you used to capture the data was made by a human.
Humans mess up all the time.
Your data quality is not bad.
You either don’t know how the data was created or how it was captured.
Understand the context, unlock the value.
What’s the fun in only reading the last page of a novel?
Data is not critical.
Critical means you can’t run your business without it.
Data is important, but your website won’t crash if your KPI dashboard doesn’t load.
Data doesn’t pay your server costs or your employees.
If a broken dashboard prevents you from doing your job, then you don’t know how the business works.
If you don’t know how the business works, you shouldn’t work there.
What about data we send to a regulatory board or serve to customers?
That’s product, not data.
Product is critical, data is useful.
There is no such thing as an urgent data request.
Data must not override common sense.
The human brain is a deep learning model trained on five million years of data.
Instinct is the output of this model.
This model is bad, but it’s the best model on the market.
Someone had the idea to fill in instinctual gaps with data.
This was good, but some went too far and got rid of common sense, too.
Data can supercharge instincts but can never replace them.
Do you really need an A/B test to help you decide between a green or blue button?
If a dashboard told you to jump off a bridge, would you do it?
There were successful businesses before big data.
Is there a Data Middle?
I’d like to go there.
In the meantime, you can find me at Data Last.