Data is a dish best served . . . with lasagna

Zach Coseglia Zachary Coseglia Culture Data Ethics Hui Chen

“This isn’t a story about politics or journalism (or lasagna). It’s about data.”

Like many of you here in the U.S., and I suspect a good number of our friends abroad, I spent last Tuesday night glued to the television (and my phone and at least one computer), awaiting and digesting the results of a tightly contested and divisive—but also unequivocally historic—presidential election.  As a side note: it was also a self-imposed, and well-earned, cheat day.  While the journalists, pundits, and analysts on TV (and, yes, those in my phone and computer, too) fed my political hunger, a homemade lasagna and some garlic bread fed my actual hunger—and my anxiety. And it was good. Really good. But this isn’t a story about politics or the candidates or lasagna. It’s also not really a story about television news or our (read: my) sometimes undisciplined hunt for information, insights, perspective, and reporting across multiple devices. This. Is. A. Story. About. Data.    

Now, don’t get me wrong, data—be it through polling or other research methods— aren’t exactly a cutting-edge development in the world of politics or political reporting. Far, far from it, in fact. But in previous election cycles, I can’t remember seeing such an intense use of data to unravel the story of election night and attempt to make sense of the vote.

Yes, America fell in love with Steve Kornacki years ago (for the record, I’m okay with anyone who wants to rebrand me as a “Chartthrob”); he was, after all, named one of the sexiest men alive all the way back in 2020 (and, frankly, I think he was robbed of at least two or three spots on that list). Over at CNN, John King has been scatting and doo-be-doo-doing at the Magic Wall since it debuted in 2008; but the wall only just became available by app this year.  538, the polling aggregator, blog, and analysis website founded by journalist and statistician Nate Silver (early loyalists, like me, know it as FiveThirtyEight), has also been around since 2008; it was a licensed feature of the New York Times in the early teens; and has been a part of ABC News since at least 2018 (before that, it was linked to ESPN, also a Walt Disney Company).  And if you watched ABC on election night last Tuesday, you surely would have seen 538’s current incarnation (sans Mr. Silver’s forecasting model) on prominent display. Oh, and just to put a finer point on the history of it all, modern presidential polling, pioneered by our friends at Gallup (or better said, by Mr. Gallup himself), has been around since the 1936 presidential election; and (ostensibly, much less sophisticated) straw polling dates back to at least 1824. Surely, you remember the 1824 election, right? That’s the one where Andrew Jackson won both the popular vote and the electoral vote; but didn’t win the presidency because, in the absence of winning a majority of the electoral votes, the House of Representatives elected John Quincy Adams instead. What a system we’ve got. Amiright?

Then, as now, journalists used the results to tell data-driven stories about the electorate and to prognosticate about electoral outcomes.   But let’s get back to today. Or last Tuesday night. Because as I scanned through pretty much every available news source in those prime hours of election night, I saw far less “talking heads” than normal; far less political posturing and dueling perspectives from each side of the political aisle; and far less awkward “person on the street” interviews in front of random polling sites.

Instead, data was the star: early returns were analyzed in real time; exit polls were dissected as soon as polls closed; sophisticated data visualization brought the numbers to life (I especially liked CNN’s use of bubble charts to show the size and political leanings of outstanding votes by county); and historical data was used to offer context—whether it was comparing 2024 to 2020 and beyond; or 538’s use of predictive performance metrics to show the level of support that a candidate likely needed in a particular county to get them on a path to victory. Put simply, it was several hours of swimming—maybe a bit of treading—in rich, wonderful, sometimes confusing, sometimes confounding, real-time data. And it got me thinking about a few things that are relevant to the work we do in the ethics, compliance and culture spaces.

Because this isn’t a story about politics or journalism. It’s about data.

Data is Everywhere, Folks . . . and it Kind of Always Has Been

Let’s start with basics. You know this. I know this. Your neighbor knows this. Data. Is. Everywhere. It was on election night, and it is every other day of the year, as well. Mr. GPT (that’s GPT, Chat) tells me that more than 90% of the world’s data has been generated in the last decade; and that we generate more than 2.5 quintillion bytes of data every year month day. Every day. Nothing earth shattering there. We’ve all heard some version of these stats—and whatever the actual number, it is so unimaginably massive that our brains can’t possibly comprehend it. Data powers our elections and political journalism. But it’s also the language of business.  

Manufacturers use predictive maintenance technology to get ahead of system failures, using real-time data to schedule proactive machine maintenance.  Retailers (and just about any business that, you know, sells stuff) use dynamic pricing models powered by customer demand, competitor pricing, and customer behavior data to adjust prices and to find the pricing sweet spot for their product. Customer-facing businesses use all sorts of publicly available and privately collected data (from social media posts to surveys to customer reviews) to gauge public sentiment about their products and services—and to inform marketing or other businesses strategies.  Subscription-based businesses have churn prediction models to identify potential risk of loss, and to reengage customers with special offers. We could go on and on and on.

These modern examples are exciting; but it’s not the data that makes them “modern.” It’s the sophistication of the analyses and the modernization of computing power that has opened doors to advanced analytics and machine learning. But here’s the thing: data has always been the language of business. Sales forecasting. Inventory management. Customer segmentation and analysis. And anyone who has ever managed a P&L knows that basic data-driven analysis and decision-making is part of the job (or maybe, it is the job)—to grow business and manage costs.  If I didn’t think I’d lose you, we could go way back. We could talk about Frederick Winslow Taylor, time study, management science and his influence on global manufacturing processes (fun fact to data visualization nerds: Taylor was a close associate of Henry Gantt of, you know, the Gantt Chart). For that matter, we could talk about Henry Ford—the godfather of the moving assembly line—for days. We could even talk about cave paintings and abacuses. Abaci? But we won’t.

Again, the possibilities, today, are nearly endless. But when it comes to ethics, compliance, and culture, we’re still figuring things out. And that’s okay. Yes, there’s a lot of exciting work happening in this space; but there’s a long way to go. And it’s not just about the tools we use or the opportunities that exist to innovate risk assessment processes, monitoring activities, and investigations; and it’s about more than building better reports, measuring programmatic effectiveness and impact, and telling authentic stories about organizational culture. At a much more foundational level, it’s about being able to speak the language of business. 

New Skills and Maybe Even Different People

Over the past couple of election cycles, we’ve seen this increasingly data-savvy set of players taking the spotlight—behind the camera and in front of it, and as contributors to legacy and new media. It’s a natural evolution of an increasingly data-centric political and journalistic environment. And we’re slowly seeing the same thing happen in the worlds of corporate ethics, compliance, and culture. 

Years ago, when I was still in-house at Pfizer, our Chief Compliance Officer challenged me to build a more data-driven approach to managing and identifying risk and building a culture of ethics and integrity. I had the gift of a wide remit, and the resources necessary to turn that challenge into a reality. I quickly realized, though, that the first step wasn’t enhanced data governance or investments in shiny new technologies: it was people. In a function that was led and dominated by lawyers (and to a lesser extent, other traditional culture and risk personnel), we needed a team that could turn our analytical ambitions into reality.  So, I hired an engineer, statisticians, data scientists, and visualization experts, who over time, successfully infused data. It wouldn’t have been possible without them. And in the years since, I’ve seen others take a similar approach. Slowly. But surely. It’s happening.

But it’s not just about bringing in new people. It’s also about upskilling existing personnel. Now, I’m going to get on my high horse here for just a second. To all my lawyer friends (and, yes, I still identify as a lawyer even if I’m not a practicing one), and even some of my colleagues in human resources: you have got to stop making comments (you think they’re “jokes”) about how you chose your career path because “it wasn’t supposed to involve numbers” or “math” or “data.” It’s terrifying to your business partners. This particular pet peeve has already been written about, and recently, so I won’t belabor the point. But whoever told you that math wouldn’t be part of the job lied to you.

Data Storytelling, Not Data Dumping

As much as I appreciated election night’s data-centric reporting, I also found it head spinning at times. And I’m a data guy through-and-through. The best reporting, in my view, supplemented raw data with thoughtful visualization and context: by comparing 2024 numbers with performance data from years’ past, by overlaying pre-election polling data with key campaign and world events (showing, for example, the impact of the assassination attempt on voter perceptions), by cutting-and-slicing data by core demographics to show shifting political loyalties, and by using exit polls to help us understand voter sentiment and priorities.  When it was good, I felt like I was being told a story. When it wasn’t as good, it could often feel like an assault of numbers and percentages and charts and maps.

This can be forgiven considering the real-time nature of election night returns. But it’s a complaint that I often hear from business leaders working alongside ethics, compliance, and culture professionals. They’re getting numbers and data, but not enough analysis and insight—those numbers and data aren’t telling a clear or compelling story. It’s what I call the curse of “so what?” We put data about compliance, ethics, and culture in front of our business partners, thinking we’re speaking their language; but they look at what we’ve shown them and say, “so what?”  

“We’ve had 50 investigations this quarter.”  So what?

“75% of our employees strongly agree with the statement, ‘Doing business with integrity is as important as how much business we do.’” So what?

“When asked, 95% of our employees rate ‘respect’ and ‘transparency’ as ‘very important.’” So what?

“We spent $1,500,000 in the last six months on suppliers who have ties to foreign governments.” So what?

These are all interesting and important truths, but without deeper analysis, insight, and context, they’re not actionable. When working with data, storytelling should be the goal. Storytelling about risk, programmatic effectiveness, human behavior, employee experiences, and the impact of it all on performance.

People are Complex

Another observation on election night: data can be easily (or clumsily) misunderstood—and we need to be careful about leaning on our assumptions or over-generalizing.

I won’t get too deep into this because, again, this isn’t a piece about politics. But if we take an overly simplistic view of people and their priorities, we might jump to the conclusion that certain voting blocs made choices last week that conflict with their own self-interests and political beliefs. Some exit polls showed that nearly 10% of self-described “liberals” voted for President Trump; in those same polls, 10% of people who said that abortion should be “Legal in all cases” and 49% of people who said it should be “Legal in most cases,” voted, again, for President Trump (on the latter, an equal percentage voted for Vice President Harris).  Likewise, in Arizona, Missouri, Montana, and Nevada, we saw the electorate vote to protect reproductive rights, while also voting for conservatives candidates. And in Vermont, where I spend a lot of my time, we re-elected our Republican governor by more than 50 points, well wider than the margin of victory for Vice President Harris (~32%) and by an even wider margin than our beloved Senator Bernie Sanders (~31%).

Now, in each of these cases, a couple things are true. First, if you do more than just quickly judge raw numbers, a story can be told to explain the results. Reproductive rights, for example, are no longer the partisan issue they once were. What’s more: data from this election show, and show convincingly, that other topics were simply weighted as more important to wide swaths of the electorate. And in Vermont, well, I’m not sure our Republican governor is actually a Republican. But that’s another story. The point is: when we dig deeper, there are stories in and around the data that need to be told. The second truth: people are complex, unpredictable, and sometimes, contradictory. This rings true in all our work, but especially in the organizational culture space.

We see human complexity at play in the business leader who regularly champions ethics and integrity, but who also skirts the rules to expedite a business process she and her team view as overly burdensome. We see human complexity at play in the low-level purchasing manager, who has no misconceptions about the Company’s written policies and procedures, but who feels trapped by his management team’s expectations of performance at all costs.  We see complexity at play when leaders say, “do the right thing”—but in a company of 10,000 people, spread all over the world, there are countless interpretations of what is “right.” And we see complexity at play when I tell a CEO that his culture has a strong culture of respect, but, also, that respect is an area to work on because it’s not currently working for—or experienced in the same way by—all. These are all stories that need to be told with data (qualitative and quantitative), nuance, and thoughtfulness.

Don’t get me wrong, sometimes, things are simple. But when they’re not, the instinct to oversimplify might lead us on a dangerous path: one where our biases and assumptions tell a story that isn’t actually there.

If you’ve stayed with me this long. Thank you. Everywhere I look I see lessons that can apply to our work; so, keep an eye on this space for more soon. And truth is, no matter your political leanings, last week was a stressful one. And writing this was definitely a form of catharsis. But not nearly as much as the Lasagna.

Previous
Previous

A Word on Monitorships From Our Co-Founder

Next
Next

What we do . . . an overview of our services