Hi everyone.

As we’re entering the holiday season, I am doing the traditional exercise of looking back at the year (and thinking ahead to the next one). And looking at almost a year’s worth of this newsletter, I realised I hadn’t really touched topics surrounding culture — company culture and how we work to strengthen our data teams and our data practice in our media companies.

So for this week and the next one, this is what I am proposing we take a look at.

But as far as next year, don’t just leave me to my own devices: Drop me a note with what’s on your mind and your ideas for our programming for 2023!

All my best, Ariane

Anchoring data team culture in the mission to inform

News of tech layoffs are a big culture shock for Silicon Valley but also for industries further afield, like ours, where we think of tech’s unshakeable job market as a bellwether (DotCom bubble bust, anyone?) 

When you’re not a hot start-up soaked in VC money, our experience of hiring for technical roles is to fight an uphill battle with more roles than people to fill them. And while this is true across the spectrum of technical roles, this is particularly acute in data. 

But hiring for data in media was never just a numbers game. 

Sure, it’s tricky to compete with Google’s millions, but media does manage to hire for technical data roles. And while the current wave of layoffs may be good news for your recruitment team, there are still plenty of roles across the industry that will be offering more foosball tables, more cash, more company-paid team-building trips to some skiing destination.

So yes, not just a numbers game.

So the question is: How do media companies create team cultures that are attractive and meaningful for data folks? 

The war for talent is made of several distinct battles. Yes, a competitive compensation package is going to matter. We will leave this one for a different newsletter though (or maybe 10 newsletters). But if the recent developments at Twitter have shown us anything, it is that folks will absolutely leave their well-paying jobs if company culture doesn’t align with where they are. The tech labour market isn’t quite so broken that money takes all.

So what does news media offer that can win over that in-demand data person? 

A great mission. And, hopefully, a company culture that is evolving in a manner that puts data in key decisions.

“The easy thing about working for The New York Times: We are a mission-driven company,” Kendell Timmers, the SVP and head of data and insights at The New York Times, said at the Monte Carlo Impact conference last month. “We seek the truth and help people understand the world. You’ll notice this mission is very journalism centric. But everything about the mission and the values have a data equivalent that we can use when we think about how we hire.” 

(Sidebar: We’re lucky to have Kendell as one of the advisors of the Smart Data Initiative here at INMA.)

And to augment this, consider that this is something I had personally observed throughout the tech organisation at The New York Times (she says, a 14-year veteran of the joint): I can’t tell you how many developers I knew at this organisation who joined specifically for the mission to inform society. Engineers who had also worked on their college newspapers. Data engineers who took sabbaticals to go on Trust & Policy fellowships in academia or join advisory bodies to contribute to the great work on privacy. 

You’d think this is just at The New York Times, but many of these folks eventually leave The Times to go and join other institutions precisely because they want to support these other organisations, understanding that being mission-aligned is a big driver in their own motivation and happiness. 

So what are these mission-driven values that you’d weave at the organisation-level but also in the data team? 

“Curiosity, respect, collaboration, excellence — these feel very common to to any organisation within the company,” Kendell said.  “So how does that work in the data world? Independence originally felt like a very journalistic focus. But in the data sense, we focus on unbiased analytics and fact-based recommendations. So I’m not going to provide something that’s the answer you want to hear. We need to provide the answer that’s the actual correct answer. Otherwise, we can’t make great data-driven decisions.”

A place for diversity of experience, perspective, and debate

This point is really key — and probably a hard moment of transformation for many a legacy media company: When we say, “We want to be data-informed, data-driven,” do we mean it?

How often is the data person in the room asked to join a project informed with a shiny PowerPoint, and the decisions made in the room are actually already made and not particularly informed by these data insights? 

If you’re the person in the room making a lot of the executive decisions, go back to your last big project and contrast the data you’ve been given with the decisions that were made. 

When Kendell says, “We can’t make great data-driven decisions,” there are two cultural angles to consider: 

• Whether the data team has a place of safety and independence to speak from.

• But also whether the rest of the company has a meaningful culture of accepting and leaning on data that may contradict long-standing assumptions.

This is a point I had heard in an interview with Robin Berjon (who was leaving his role as the VP of data governance at The New York Times and is now at Protocol Labs). He noted how much culture building was necessarily — from the data team and also from outside the data team. And that data wasn’t perceived to be a place to get your project rubber-stamped with “got data to back it up,” but rather a team that would constructively add value to whatever is being built or optimised.

“You need to have built basically a culture and a systematic awareness of what to do so that people come to you. And also you really need to make it clear that, you know, you’re not the data police, and you need to have a reputation for solving problems and making people’s lives easier,” Robin said.

“If you build that culture, then people will actually line up to come talk to you — just because you making their lives so much easier. And, you know, a lot of that has to do with being some kind of repository of institutional knowledge.”

But Kendell Timmers of The Times connected it to values that are also expressed in company values, writ large — those we convey to our ultimate constituents: our users.

“Integrity, building the readers trust in each other’s trust, from the reader standpoint, considering privacy and algorithmic ethics in our work,” Timmers said. “That’s a really key part of integrity, curiosity — for everybody in the company welcoming up debates and different viewpoints. I think from a tech standpoint, a big focus and curiosity has been eager to try new methodologies and tools, new approaches, and also being welcoming of people who came to the analytical groups from different backgrounds.”

This is where culture and hiring start to dovetail — both in terms of making it more likely that hired team members flourish and contribute fully, but also in creating wider, richer pools. Particularly when it comes to diversity (a cruel problem in data), we can’t expect to diversify our workforce by plumbing the depth of the same small, shallow pools of candidates. And the cultural work to both prime the company to extend its understanding of what makes a great candidate and the cultural work to make your company a great place to work are very much two sides of the same coin.  

Next week: hiring, remote work culture, and the data team.

Further afield on the wide, wide Web

Meanwhile, in Internet rabbit holes: AI auditing. The complexity of which is greatly complexified when, in the first place, the data you want to audit is not easily obtainable — like the recommender systems that create the news feeds of our favourite social apps.

AlgorithmWatch offers a really interesting overview of the work going on in this space — both some of the methods researchers are using to try to obtain data sets and some of the challenges they are facing as social platforms push back on the work being carried out.

AlgorithmWatch itself last year had been summoned to take down its data scraping tool, though later regretted they didn’t go to court to push back on the injunction. It’s like David vs. Goliath, but also, David wears a blindfold, most of the time.

About this newsletter

Today’s newsletter is written by Ariane Bernard, a Paris- and New York-based consultant who focuses on publishing utilities and data products, and is the CEO of a young incubated company, Helio.cloud. 

This newsletter is part of the INMA Smart Data Initiative. You can e-mail me at Ariane.Bernard@inma.org with thoughts, suggestions, and questions. Also, sign up to our Slack channel.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .