Discover more from benn.substack
The missing analytics executive
We should redefine the role of the chief data officer. For our companies, for our careers, and for ourselves.
There’s a treadmill at the mountain top.
Among the data leaders I know, many of their stories are the same. After rising up the ranks of their organizations, from junior analysts and data scientists, through positions as team leads and managers, and eventually to directors and VPs, their data careers stalled. Somewhere between senior management and the executive team, they found themselves caught at sea, adrift between the land they loved and the aspirational new world they left it for.
The shoreline behind them—the ones many of us leave, in the name of career advancement—are those of our intellectual homeland: The creative work that attracted many of us to the data industry in the first place. By making the leap into management, we gave up data’s puzzles and problems to lead teams, set OKRs, and manage vendor contracts;we gave up giving in to curiosity; we gave up, in a more subtle sense, the work that makes many of us us. We traded them away for pay and prestige, and for a sense of professional progress.
Our aspirational destination is a position of senior leadership. It’s in the CEO’s inner circle, in board meetings, in the room where it happens. It’s where strategic decisions get made, and where we aren’t handling requests from other teams, but making them. It’s where, as the cliché goes, we have an impact.
Most people never make it. Companies rarely have data leaders that sit among those at the top of the org chart. Analysts are typically told to drop anchor just offshore, one step below the c-suite, buried underneath some other function and some other executive.
In the rare cases where there is a chief data or analytics officer,the role is underwhelming. CDOs and chief analytics officers, unlike chief revenue, financial, marketing, and technology officers, are often second-class operational roles. They deal in risk management and tooling governance, and are more secretarial than strategic. Without the gravity of a large organization underneath them, these data executives play bit parts, pushed to the back of the board deck and relegated to the always-too-high G&A budget, an administrative asterisk next to the departments that are seen as making real products or real money.
Always stuck one door away from the inner sanctum, many of the data leaders I know bounced from company to company, looking for new ways in. But each time, they were trapped under the same mandate: Hire a team; rehabilitate broken data infrastructure; build a data culture. Across a range of company profiles—from enterprises “reinventing themselves” to tiny rocket ships, at companies growing quickly to companies imploding, with leaders eager for data and those skeptical of it—the outlines were the same, and the doors locked in the same place. The terminal point of an analyst’s career, it seems, is the data leader treadmill, hopping between different VP roles, always running but never moving forward.
Once in this loop, some people—a third of the ones I know, I’d guess—have stuck with it. The lucky ones do it out of a genuine love for management and a passion for helping others grow. Others are simply waiting it out, hoping, either through an official promotion or an informal advisorship,to get invited into the company’s highest echelon of leaders.
But most people didn’t last. Bored of the job and frustrated by its stubborn ceiling, a few returned to the ranks of individual contributors, surrendering their status and stock options for day-to-day satisfaction and progressively weirder job titles.
However, the overall majority of these data leaders—or, I should say, former data leaders—jumped ship entirely. Some transitioned into other executive functions, like marketing or finance, that still had room to grow. Others left for adjacent industries, most often venture capital. And many started their own companies, deciding that the only way to get a seat at the table was for them to create their own.
There’s a small tragedy in this. Of all the things we’ve figured out as an industry—the tools to build, the experience to create, the functions to fill, the pitches to raise astronomical amounts of cash—we haven’t figured out how to keep many of our best people. And without them, the persistent fight for influence will only be harder.
The executive analyst
The lack of analytical leadership on the executive team isn’t just a problem for analysts, however; it’s also a problem for the executive team.
At their best, analysts are curious investigators, observing the problems around them and proactively looking for opportunities and solutions. Their value scales with what they can see: We can’t solve problems we don’t know about.
But there are no analysts in the rooms with the widest and most strategic views. Board meetings rarely have analytical observers, and there is no official designation for “senior data advisors” to the executive team. Nobody is directly responsible for helping a company make its most important decisions, or for exploring its uncharted strategic opportunities. There is no role committed to this work, and no title that acknowledges its value.
These jobs are instead done haphazardly. Sometimes, VPs of data do it, periodically LARPing as individual contributors.But they rarely have time to do this work, and never have time to do it well. Other times, requests are handed down to senior individual contributors, who lack the context that’s needed to do anything more than narrowly answer questions as they’re asked. And most often, strategic exploratory work isn’t done at all. It simply falls through the cracks, too opaque and sensitive for people outside of management, and too creative and time-consuming for people in management.
A better chief analytics officer
A similar problem exists in technical organizations. Engineering leaders have to oversee both teams and technology, and those who are good at one aren’t necessarily good at the other. Companies solved this problem by splitting the role in two.
Traditionally,the VP of engineering sits at the top of the engineering org chart, and is responsible for the day-to-day operations of the team. They hire managers, manage directors, and direct the machine that builds the company’s technology.
The CTO, by contrast, is shaped like an individual contributor. They don’t typically manage large teams (or, in some cases, any team at all). Instead, they’re “an architect, a thinker, a researcher, a tester and a tinkerer.” Their job is to evaluate the organization’s most important technical decisions and to push its technological frontier outward. They derive their authority from expertise and influence, not an official reporting structure.
Despite the apparent discrepancies in title (CTO sounds higher than VP) and responsibilities (leading a department sounds more important than tinkering), the two roles are peers. Both are senior executives, and both often report to the CEO. The division of labor is a recognition not of hierarchy, but that there’s enough important labor in engineering that it needs to be divided: One role to manage, and one to advise.
Data departments should follow the same pattern. Rather than being led by a single ambiguously defined and overburdened CDO, data teams should have two representatives in senior management: A VP of data responsible for managing the team’s daily operations, and a chief analytics officer.
Much like a CTO, a chief analytics officer would be charged with working on their company’s most important and far-reaching problems, but without the management responsibilities (or organizational authority) of a department head. In exchange for their wide latitude and generous leash, they’d be expected to deliver impact commensurate with that of a CTO.
In some cases, this could be bold strategic research that uncovers a new market opportunity or an ambitious new model for forecasting future hiring needs. In other cases, it could be more superficially mundane—finally reporting on revenue correctly, say—but no less important.
To see how this is different from our current approach, consider a senior analyst who sees that the sales team is struggling to sell to a particular market segment. In response, they could build a new model for recommending which prospects in that segment are worth pursuing and which ones aren’t. This would be, by today’s standards, a good result.
An experienced chief analytics officer could respond differently. They know, from a recent board meeting, of the strategic importance of that market segment. The leadership team decided that the best way to improve its presence there is through key partnerships with several adjacent companies. The chief analyst suggests evaluating acquisitions as well, and creates a series of analyses to assess different options.
This latter work requires not only analytical craftsmanship, but also the ability to size markets, model sales dynamics, diagnose product metrics, and read financial statements. In a problem as multifaceted as an acquisition—especially if you have to generate the idea, not just respond to it—reasoning from first principles isn’t nearly enough. You need to have seen some things.
Moreover, this example highlights another advantage executive analysts have that others don’t: They have access to privileged material. In extreme cases—acquisitions, financings, layoffs—information is carefully guarded. These problems, however, are in no less need of analysis than others; in fact, the fewer people in the know, the more important the issue probably is.
The same applies to more routine problems. As Bobby Pinero argues, the best analysis often comes from curious analysts piecing together ideas from the conversations they overhear. By sitting in strategic meetings, even as little more than an observer, chief analytics officers have access to the company’s most valuable conversations—and can connect its most valuable dots.
A better career
For data leaders who want to be operational executives, this split lets them focus on making decisions and frees them from the uncomfortable burden of periodically being asked to be the board’s query mechanic. And for those who are good analysts but bad managers, this role is both cake they can have and cake they can eat. The promise of a chief analytics officer is the path for advancing their careers (and compensation) without having Peter principle themselves into middle management oblivion.
More broadly, the role acknowledges the importance of analytical work. It says great problem solvers, alongside operational leaders and great technologists, have a role on the executive team. If official advisory roles like these work for the president—if it’s valuable to have people in the Oval Office who aren’t elected leaders or cabinet secretaries running large organizations—surely they could work for CEOs as well.
Finally, this redefined chief analyst offers a way to step off the data leader treadmill without walking away from data entirely. Rather than having to chase impact and influence in other departments, great analysts can climb all the way to top by working on what they’re good at—and, usually, what they love to do.
Of course, such a role wouldn’t be without its perils. Chief analysts could develop the same organizational pathologies as bad CTOs who operate as ungovernable mad scientists, ignoring processes, working in their own sandboxes, and refusing to play nicely with anyone else on the team. But this problem doesn’t prevent us from hiring CTOs, and shouldn’t prevent us from hiring chief analysts.
In a less severe version of this problem, a chief analytics officer could also block other analysts’ access to meaningful work. I think this is a real risk, especially if they become the pet analyst of the CEO. But this isn’t insurmountable either. Chief analysts should proactively source much of their own work, adding strategic projects to the queue rather than plucking the best ones from the top of an existing list. Nor should they be loners. On projects for which it’s appropriate, they should recruit other analysts for help, with both working together in the trenches to solve a company’s most important problems. And just as engineering architects can operate as junior CTOs, senior analysts could forge a path to becoming chief analytics officers through parallel roles.
The biggest obstacle, however, is probably organizational. How many cooks can we add to the executive kitchen?
It’s a fair question, but the alternative—analytical games of telephone, missed opportunities, and stalled careers—is surely worse. Adding a quiet observer to the executive team is a small price to pay for consistently putting the full force of the company’s best analysts behind its most pressing problems.
There is, I admit, some irony in this proposal. During my time at Mode, I’ve had about a dozen jobs, including being a blogger, a PM, a support agent, and various stints running the data, marketing, product, solutions, and executive teams. The only consistent thing has been my title: chief analytics officer.
I tell myself that my eight-year rotational program is simply the life of a founder. But that may not be true. My spasms through Mode’s org chart may actually be a reflection of the general anxieties of those in the data industry who are searching for the right ambition for their careers. What I’ve been looking for, it turns out, may have been here the whole time. I just needed to rewrite it to make it my own.
For reasons I don’t entirely understand, analysts who move into management seem to give up some of their identity as an analyst. Design VPs, for example, still identify strongly as designers, simply in a new role. Data executives, by contrast, appear to lose that identity with each promotion.
There’s a deeper disease here, endemic everywhere but especially potent in Silicon Valley, where we worry more about how fast we’re rising and less about where we’re going, much less if we actually want to get there. I have a much (much) longer rant on this that that I’ll publish someday.
And often without compensation.
Many data executives are years removed from their days as analysts or data scientists. Speaking from experience, you rust.