Beyond One and Zero
The madman cried: "Data is dead! And we have killed her."
This is part two of a three-part series exploring attitudes towards data. 1
Do I think data is important? Absolutely.
Do I think you should use data? You betcha.
Do I like getting paid to work with data? Totes my goats.
But do I trust data? Do I believe data? Uh, next question, please.
You see, I’ve selected data I can’t unselect. I’ve dredged the swampy morasses where they say big data lives. I’ve peeled off the semantic leeches from my legs, as they try to bleed me of the little domain understanding I started with. I’ve seen fisherman spear the bobbing carcass of a table, plate and dress it up, then serve it at the quarterly board meeting.
So if I dislike consuming your data, it’s not personal. I prefer to wash, clean, tutor, discipline data. To teach it to be well-behaved, know its limits, treat others with due respect. Give me the chance to work with data, and I’ll jump right in. But to eat it? No thanks.
I am a pessimetricist: an advocate of building data-driven systems, even as I find it hard to uncritically participate in them.
A colleague says, “We had 424,000 daily active users yesterday.” The pessimetricist thinks — hopefully, he does not say this — “Actually, you had in excess of 424,000 HTTP requests from devices associated at least temporarily with unique user accounts registered in your internal systems over a 24-hour time period that survived a number of arbitrary assumptions in your data processing systems that passed muster six months ago but which haven’t been re-evaluated meaningfully since.”
The pessimetricist is like the post-relativity scientist. Obviously, she believes in the truth of classical laws of physics. Obviously, she also believes they are wrong. She exists in a measurement multiverse: until she looks in the box, the metric is both right and wrong. What if a new feature caused an event to double-fire? What if a person has two different email addresses? What if our event bus was down briefly? What if…?
It’s a creative tension that, when healthy, leads to rebuilding, redefining, restructuring. The pessimetricist prospects for semantic oases that can be stabilized, terraformed, made viable for life. But it won’t be home, they won’t be comfortable there, because they know other worlds await.
The optimetricist and pessimetricist are both measurement advocates, but they lie on opposite ends of a social investment axis. The optimetricist is driven by a desire for consensus; the pessimetricist by a need to make a complex reality knowable.
Not everyone cares about data like this, which means it’s time for a 2x2:
The first dimension is an individual’s concern with social consensus: how concerned is the person with the semantic consent of others? The second dimension is a person’s investment in describing the world quantitatively.
The skeptimetricist is pro-social but anti-measurement. They need to work with other humans, to have a shared understanding of the world. Yet, they are hesitant to trust the metrics put before them, whether for procedural — this number isn’t right — or epistemic — this number isn’t right — grounds.
Sophisticated skeptimetricists spout off on the limitations of quantification. They advocate qualitative analysis and a “holistic understanding”. They gravitate towards comparisons, case studies, and process more than numeric results.
Skeptimetricism is understandable, especially for a business leader. The world he sees is complex, but prolonged skeptimetricism is untenable. At some point, he must take ownership of how his team views the world. It is too inefficient to organize otherwise. A dashboard will be demanded. The leader faces an either/or:
Impose “the dashboard". He champions his own quantitative view of the world, thereby becoming an optimetricist.
Depose “the dashboard”. He pays lip-service to metrics but pushes a personalized narrative, thereby collapsing into nihilmetricism.
The nihilmetricist is no crackpot. Her denial of quantification is not on semantic grounds, like the pessimetricists who say, “There is always another way to calculate it.” Her denial is not social, like the skeptimetrics who claim, “These metrics don’t capture the nuance of our enterprise.”
The nihilmetricist refuses to submit to quantification.
If the numbers don’t “work out” as needed, she can change the game easily enough. Her slide deck may be full of charts — all of them carefully curated in a private Excel workbook and copy-pasted into the board deck. She is no system-builder, creating a ship that will work of its own accord. She is a human force that builds towards consensus — her own.
With a wave of her hand, the nihilmetricst annihiliates entire quantitative realities: “We no longer care about costs for this project.” Like Atlantis, an entire city of financial metrics is submerged, to live on only in whispers and legends.
Or, she brings new worlds into being: “Our north star metric is now community engagement.” A dozen dashboards crop up where before there were none. Yet, these worlds are not binding, not to her.
There is only success and failure, and although data is a fact of modern life, the nihilmetricist is not a sheep, not one to subject herself to some pseudo-reality defined by others. The nihilmetricist — at least the one who makes it out of bed in the morning — is a hawk. And everyone knows hawks are to be watched, admired, feared.
And sheep? They’re for counting.