How to Feed a Data Professional
Columbus, Ohio. 4:00 PM. Every. Single. Day.
Snaaaaaacktime!
Mom, Dad, its snacktime!
Nack! Nack!
It’s 4 ‘o clock!! That’s snaaacktime!!
Nack!
Then, three voices in unison: “WHAT IS FOR SNACK!!!”
To my children, snacktime is an experience. Mealtime is a chore. Snacktime is full of possibilities. Mealtime is a mundane reality. Snacktime is loose, informal, fun. Mealtime is a butts-in-seats, parents-saying-words snoozefest.
It’s not like they’re getting ice cream and Pixie Sticks. There is something about the ad hoc nature of snacktime that any food item— a whole carrot, for example — transubstantiates into something greater than its organic reality.
In other words, my kids love snacks.
As Responsible Parents, though, my wife and I have to hold a firm line, lest the border between snack and mealtime merge into a boundary-free blur. If they had their way, the turning of every hour would yield a hit from the snack pantry: Pop-Tarts for breakfast, parfaits for second breakfast, PB&J for elevenses… you get the picture.
Punctuating the day with slower, deliberately crafted meals is not only a way for parents to stay sane, but it’s an opportunity to collect the family, share stories, to strategize about the coming hours and days. It is a ritual that celebrates body, mind, others, and the Bigger Picture of what we’re trying to build as a family. So while I might not be able to build a snack/meal food classifier, I can certainly spot the difference in practice.
Delivering a perfectly crafted data asset is a great pleasure. A scatter plot that creates a surprising and compelling connection. A singular number that warrants massive, bolded font. (You type it in caps lock even though you know it’s just a number). A completed spreadsheet of quarterly metrics requests, packaged tidily with red-yellow-green color-coding, like a pack of Starburst. Delighted stakeholders take these and throw them in their slide decks, on their Wiki pages, on their car bumpers.
In other words, everyone loves a good data snack.
It’s a great exchange, but if snacks are all that’s on the menu, the organization is not going to grow into its most vital self. Instead, it results in a shallow snacktime data culture.
A snacktime data culture is transactional. Someone in the business says, “It’s time for a chart!” Someone else says, “I can make a chart!” And boom, the chart is put on the table and consumed. No etiquette, no entree, no ad hoc conversation, no coaxing/pleading to “try new things”. Just a nice slice of data pie meant to satisfy the needs of the hour.
A hypothesis: without active leadership — a Responsible Data Parent in the room— data teams will face constant pressure to become snack shacks. In industry, there is a downward force to push data work to the lowest common denominator. We need to choose boring technologies, keep it simple, and be storytellers, not probability peddlers. Everyone already has too much to do, so can’t we just cut the analytical hand-wringing and put the bottom line up front, please?
This is a healthy pressure. More clarity leads to faster decisions. It is a reductive pressure nonetheless. As with “on-the-go meals,” there is something more than the food quality that gets lost in translation.
I spent years working in the data version of a snooty high-end restaurant: the academy. And while the academy has its share of problems — including the publish-or-perish culture, toxic work environments, a lack of real-world relevance — it cannot be assailed for lack of substance or ritual.
Building a research lab is very much like building a startup. You hustle to build infrastructure, you hustle to hire scrappy employees, you hustle to carve out a niche in a sub-field of a promising focus area of an established discipline (and hope that it will remain well-funded).
The difference is that the “product” delivered by the professor — and the academy — is a constellation of arguments meant to elucidate The Bigger Picture.
Unlike in industry, there is never a question of how central the “data” is to The Bigger Picture — although, when we talk about data in the academy, it is not a reductive techno-oriented conception of data. It is more akin to evidence in a trial. It is always up for examination. What separates the experienced researcher from the inexperienced is mostly in the interpretation of the evidence, not the technical ability to collect it or shape it.
A “data snack” is never sufficient support to “rest the case” on, though it may be plenty gratifying. It is always just one piece of evidence, and you can be assured that it will not be accepted without interrogation, both internal and external. The data is always just “part of a well-balanced breakfast”. This skepticism is healthy and helps to keep the focus on The Bigger Picture.
However, if industry data culture faces a downward pressure to become snack time, then the academy has its own upward pressure to devolve into competitive slap fights, like a dinner with surly teenagers.
Mom: Thanks for this great meal, dad! Kids, how was school today?
Annie, pulling journal from backpack: Soooo much fun. We learned about bar charts, and I made this neat one: it compares how smart James is relative to his peers. See, this is James here… *points to smallest bar*
James: Annie, that’s the stupidest chart I’ve ever seen! Your axes are not even labelled and I don’t even see the methods in your caption! You are going to fail that class so hard that —
Dad: Rooms! Now!
Ah, the academy. Rich in context, deep in content, entirely dysfunctional.
Like Responsible Parents, data leaders need to ensure their teams are getting a regular and balanced diet. Malnourished data folks get bored. Bored people quit.
Fortunately, I think there are tangible ways industry leaders can nurture their analytics professionals, and it doesn’t involve implementing Bayesian analysis in Looker or moving from Google Docs to LaTeX. It simply involves balancing the real-world demands of your industry with the higher-order needs of data professionals to participate in The Bigger Picture.
It’s standard practice for a data team to focus on building out modern infrastructure, landing a couple of key data products, and stewarding relationships with a few lines of business. There will be hand-wringing about centralized or decentralized org charts, and what data positions to hire — all hard, important questions. But not enough.
My hunch is that data professionals crave a connectedness to The Bigger Picture that sets them a bit apart from other professionals. They want to see the line between a data product and a business strategy. They want to understand the categories of problems they face, and bring in strategies from other companies to tackle them. They want to animate the business through experimentation.
The good news is that these strategic rituals are by no means the domain of the data team alone. They are open to all, and interesting to all. That is precisely what makes them part of the Bigger Picture. There are many ways the organization itself can sate this appetite: an executive laying out an evidence-based strategy during a company all-hands; a team lead discussing best practices from other companies for solving a thorny problem; a project manager blogging about their learnings from some failed initiative. All of these are rituals that feed the hungry data professional.
As a team with a wide lens and a craving for information, the data team can play a key role in amplifying these conversations and creating frameworks around them. Calculating Key Performance Indicators for strategic projects is the most obvious way in which the data team does this today, but it is just the surface of what is possible.
Data snacks are great, but they don’t last long and they don’t build community. If you put a feast on the table, though, you’ll have your neighbors, your colleagues, your extended family showing up. Uninvited, even. And you’ll still get a chance to share your delicious pie.
A fantastic read, as always!