When data drives growth, that’s a Hollywood ending. So where are the Moneyball sequels?
The graphic above tracks demographic representation across 12 law firms working for BASF Corporation, 2016 to 2021. On average, shares of BASF work grew +11% for diverse ethnicity partners and +46% for associates. For women, shares grew +24% for partners and +28% for associates. Additionally (not pictured), ethnicity and gender representation in firmwide leadership grew +10%.
What explains this blockbuster growth? To me, it’s the courageous leaders using data to achieve a shared mission.
Each year, BASF’s in-house leaders request diversity data from outside counsel. Parker Analytics (PA) then prepares individualized data scorecards and action items, which BASF shares with its firms. As the graphic clarifies, firm leaders responded, driving growth in diverse representation in BASF legal teams and the firm overall.
The most successful data projects I know include a shared mission—between clients and firms, firm leaders and analysts, etc. Achieving mission goals together enriches these professional relationships. So, in leading courageously with data you are powering a virtuous cycle.
To use data courageously you must do three things well: (1) lead with a single number, (2) build transparently, and (3) be flexible. When these features are present—as with BASF’s diversity program and others I’ll describe below—a data system for decisions can produce a Hollywood ending on par with Moneyball.
Disclosure: Obviously, BASF is a client of Parker Analytics. To its credit, BASF is sharing its learning with other legal departments and law firms. See, e.g., Phillip Bantz, “Q&A: BASF Legal Chief on Driving Diversity at Law Firms: Carrots Are More Effective Than Sticks,” Corp Counsel, May 10, 2022 (interviewing BASF General Counsel Matt Lepore on the company’s approach to diversity).
Building a data system for the mission
What does it mean to be data-driven?
My answer might surprise you: When data disrupts existing decision settings, it creates new tension, and even increases conflict. On topics like diversity and equity in law, data will highlight uncomfortable historical truths. Imagine seeing the marginalization of black lawyers at your firm quantified (-50 hours per month, all else equal, compared to majority lawyers).
I deeply admire law firm leaders courageous enough to confront uncomfortable truths. It helps to show vulnerability when developing data systems. So, for those with the courage to make a credible commitment, what’s the best path? To start we’ll need a system anchored by a single number.
1. Lead with a single number
To lead with a number, the analyst who is creating it needs client trust. Get leaders comfortable with a single, information-rich number (often the product of more complex computations). In legal diversity tracking, this number reflects a summary of progress, like a Diversity Index. In legal recruiting, this number is a predictive score about job applicants’ likelihood of success.
Building a system around a single number is powerful for a couple of reasons. For one, it is easy to lead with it—it’s one number! Less is more. Second, it makes it possible to approach the same decision consistently, using the same information. And whether decisions are consistent with this number in practice becomes a testable proposition.
In the most successful data systems I’ve seen, one or a few leaders get on board. They advocate tirelessly for this number, insisting it receives weight in every decision. They use their capital to evangelize skeptics. This makes people uncomfortable. It’s why leading them requires courage. Then come the payoffs.
Strong leaders are comfortable with change. And what’s the payoff: the ability to incorporate an information-rich number that predicts desired outcomes. Paying attention to this number and using it to decide brings long-run returns, like the lead graphic’s diversity gains.
In addition to a single number to drive decisions, a second key ingredient is for leaders and analysts to build transparently.
2. Build transparently
There are several ways I’ve seen transparency drive progress in building a data system.
At BASF, GCs Matt Lepore and Sneha Desai trusted PA to iterate the scorecards year after year. This trust has paid off in an effective addition to the scorecard: written, concrete action items informed by data. PA checks progress on these action items and reports to BASF and the firm the following year.
As Reed Smith’s Chief Diversity Officer, John Iino, shared with members of the Pittsburgh Legal Diversity and Inclusion Coalition, “BASF providing targeted action items based on my firm’s data and specific to our working relationship is a game-changer. Using the Scorecard, it’s a very short conversation with the relationship partners and management. Do this, the client will be happy, and the firm will benefit overall. I am a huge fan.” Like I said, a virtuous cycle.
A second example where transparency was part of leading courageously came in building a selection algorithm. Always, PA is transparent about results connecting different biographical profiles to success. Seeing these results can give rise to skepticism and insecurity—imagine a preferred selection factor, like law review, showing up in the negative column. See Post 258 (discussing this example and others during the early days of Moneyball in law). Those who lead courageously with data get past these specifics to see the big-picture value.
Recently, I shared results with a courageous partner who, because it was a transparent process, provided everyone with a moment of clarity. In the results, the biographical profiles on the negative side seemed incoherent. Having learned to interpret the results, the partner identified a general theme—the negative factors reflected a less-focused career progression (advanced degree, publication in a non-legal area, entrepreneurialism)—and it clicked for everyone. Lawyers with diffuse interests had difficulty succeeding in the firm’s fast-paced, high-dollar transactions environment, which requires laser-like client focus.
3. Be flexible
When it comes to data and potentially high-stakes decisions, sometimes the best approach is to be flexible. Speaking as someone who has been there, it takes courage to acknowledge, that despite a lot of work, you missed the mark with the first deliverables. In newer areas of work, like data analytics, clients won’t always know how to articulate what they need. If they’re flexible, leaders change the plan and the shared mission carries the project forward.
In one example, a large NY firm worked on a method for tracking associate progress. After some conversation and early statistical analysis, we’d not yet met client expectations. What I eventually realized was a different class of statistical models—growth rate models—better fit the lawyers’ intuitions around what matters. Where early iterations focused on predicting a static number, the number that resonated with firm leaders described progress in terms of a growth trajectory.
Figure 2 below summarizes hours trajectories using a growth curve analysis. It plots a summary for a composite group of associates, cumulative hours by month. Plotting progress for associates along this curve is a visual way to assess progress. A courageous leader can lead with these numbers to advocate for associates whose expected trajectories fall above/below typical hours growth (the blue line).
With the insight tuned to growth, we then analyzed the origins and nature of associate progress. Faster growth rates predicted long-term success, which is something that is worth millions in both money and emotion to those involved in recruiting and leadership. Yet, importantly, there was no evidence these rates differed across demographics.
We also learned hours growth rates were similar for all hires in the first year. A quick separation soon followed, however, indicating the second and third years’ growth predictions deserved attention (separating rates of growth occurred quickly, and much before the firm could collect and distribute comprehensive performance feedback).
Flexibility is important because the first or second iteration of data analysis is not always going to yield simple, elegant solution. However, each iteration does yield new insights. By faithfully following this process, long-term sustainable solutions follow.
Conclusion
In this post, I developed three criteria for leading courageously with data. The insights come from working alongside determined leaders. At BASF, leaders are collaborating with outside counsel and producing diversity gains. Diverse legal teams predict better legal outcomes for companies and increased profitability for outside firms. It’s a classic “moneyball” story, powered by leaders who use scorecards and shift the focus of relationships focus to the mission (a welcome change from focusing on the GC’s score on the back nine).
There aren’t many stories about data driving change told with the gusto of Moneyball. More data projects are likely to succeed when collaborators put greater effort into the human side of data leadership.
Who’s the cast of characters in this Moneyball sequel? It begins with a courageous leader willing to bet on a number and let it ride. I’d say Chris Luna, VP of Legal Affairs at T-Mobile, is a great model. Chris says the mere act of putting an item on an outside counsel diversity survey will drive change. It signals he’s focused here, and firms act when you follow up with the data.
Also, you’ll need to cast people who are invested in a shared mission—law firm leaders who want to drive diversity, or analysts who care about bringing opportunity to their organizations, and so on. Lastly, data here is a character actor, supporting and enriching the relationships that give us our Hollywood ending.
This is the main lesson for leaders of the profession: to use data effectively, we use it to focus on a mission and the building of relationships. Fortunately, at BASF and elsewhere, I have seen this done. So far, the winning formula appears to be to lead with a single number, work transparently, and practice flexible leadership.