On Talent Density

Part 1

My hot take: the north star of HR/People departments should be to maximize talent density and sustain it over time. Everything else is secondary.

I define talent density as the % of employees that are top tier, superstar, smart, motivated, engaged etc. In other words, the concentration of “A” players in an organization.

Achieving talent density

At a fundamental level, there are only a few ways to achieve talent density:

1. Just hire “A” players (don’t dilute in the first place)

2. Get rid of non-“A” players (purify later)

3. Turn “B” and “C” players into “A” players (perform alchemy)

Method 1 is the simplest, but extremely hard. Method 2 is the easiest, but still very difficult to execute on. Most HR departments and managers want to depend on Method 3, but the ROI is limited and sometimes negative.

What’s an “A” player?

I believe that we can categorize people as “A” players (and “B” players, “C” players etc.) These categories are loose, but they do describe reality in a meaningful way. 

• “A” players are negative entropy

  • This means that an organization becomes less disordered for each “A” player added. In most circumstances, adding an additional person creates more complexity (from management to communication), and therefore increases disorder in the system

  • “C” and “B” players will add value, but come at the cost of additional managerial complexity, training, and context overhead. This means that the rest of the team will need to do more work cumulatively to have a “B” or “C” player on the team

• “A” players are not about being high IQ (though being intelligent helps). I like how Satya Nadella puts it: “do they create clarity, do they create energy?”

• I exclude “brilliant assholes” from my definition of “A” players. True “A” players are to me by definition team players—or at least they understand the importance of working in teams and can do so productively. “Brilliant assholes” are necessarily “B” players at best, since they will create entropy by disrupting workflows and causing tensions and create less clarity for the team

A few other assumptions

• Over time, “A” players tend toward only joining companies with “A” founders (talent-dense companies are largely a subset of companies with “A” founders)

• Generally, only “A” players can consistently identify “A” players (great talent identifiers are a subset of “A” talent)

• Talent levels tend to be self-reinforcing across the board (“A” teams attract “A” candidates, “B” teams attract “B” candidates, etc.)

A losing battle for most

By definition, not every company can have high talent density. If there are companies that have higher than average talent density, then there must be companies with lower than average talent density.

We can see skewed distributions of talent density across every industry. PayPal Mafia, early Google, and early Stripe are a few notable examples of talent-dense companies from Silicon Valley. Top law firms, the Big 4 accounting firms, and the Ivy League, all exhibit this agglomerating behavior in other domains. (Side note: I wonder if there’s some adherence to Zipf’s law in these distributions?)

So while it’s good that most companies aspire to be talent dense, not every organization can be. Network effects mean that existing talent density tends to attract more talent, and vice versa, meaning that organizations that currently lack talent density will be stuck in an intensely uphill battle, as they’ll struggle to hire and retain top talent until they reach a critical density of talent.

My revised hot take

The north star for People and HR should be to maximize and sustain talent density for companies that are able to effectively attract “A” talent.

For companies that cannot compete on talent density, the objectives for People and HR will have to be more multi-faceted.In fact, companies with lower talent density have much harder jobs to do. They will still want to attract and retain the best talent they can get, but unlike with

talent-dense companies that can rely on negative entropy “A” players to scale efficiently, the rest of the crop will face more pressure to implement processes, system-building, and culture development, to minimize entropy and disorder over time and at scale in their teams.

Fool’s gold?

The inability for most companies to achieve high talent density through recruitment is why Method 3 is the most discussed: transforming “B” and “C” players into “A” players. It not only feels the most noble and virtuous, but it is the situation that the majority of companies and managers find themselves in. With coaching, L&D programs, and other employee engagement programs, managers and People Ops departments hope to transform their current employees into superstars.

The problem is that trying to develop “B” players into “A” players at scale is either very expensive, or impossible. I’m not saying this out of cynicism, but rather from a sympathetic understanding that for most people, work will never be the top priority for them.

Many “B” are in fact totally capable of being “A” players, but are instead, unwilling: they don’t want more responsibilities at the cost of more stress, or spend extra hours at the office to become a top performer. They would rather, for example, prioritize time with their family and for hobbies. And no amount of corporate branding or workplace “empowerment” will change that. 

And then, some “B” and “C” players are actually incapable of becoming “A” players; and therefore, no amount of formal mentoring, training, or incentive programs will drive the type of high-multiple returns managers dream about.

You’re therefore left with the rare diamonds in the rough that are currently “B” players who can be earnestly developed into “A” players and future leaders. These are “B” players who truly have the potential and desire to become “A” players. Realizing that potential is one of the most rewarding and highest ROI activities that People teams can accomplish.

Still, successfully developing “B” players with “A” potential is tricky for a few reasons:

• Most companies take a blanket approach, meaning that company-wide policies that apply equally to all employees ensure that high-potential employees are relatively underinvested in, while lower-potential and lower-motivated employees are overinvested in. 

  • Because of concerns around fairness and the intrinsic difficulty of accurately assessing employee potential, most companies don’t invest in talent development resources efficiently outside of the organic benefits of promotion schemes.

• The payback period for talent development is long, so companies with retention issues are paradoxically less incentivized to invest in developing talent, even though L&D resourcing is often seen as a retention tactic. If the average payback period is greater than the average employee tenure, then it’s economically irrational to put resources into it

  • Excerbating this trend is the fact that developing employees into “A” players and raising their aspirations can actually increase attrition, since increasing one’s talent levels also increases opportunities, and there are many more external opportunities than internal opportunities at all but the very best organizations

• It usually takes “A” players to develop “A” players. If the company isn’t talent-dense (i.e. there’s not a lot of “A” players), it’s less likely that a high potential employee will have a model “A” player as a manager or mentor to learn from

Loose thoughts & questions

• The bar for “A” talent that I’ve set is high, but I have seen it in a few cases. I want it to be aspirational enough, feasible for enough situations, and also because I do believe that it is only at this high bar that we can start making meaningful categorical observations about talent levels

• No large organization is actually all “A” players. First, there’s generally regression to the mean at scale, and these regressions only move in one direction (once orgs lose critical talent density point, it can never be talent dense again outside of a massive shakeup). Second, 100% hit rate with talent selection is practically impossible, especially when in hypergrowth

• Organizations can be very talent dense on some teams, and less so on other teams—all within the same organization. What are the common emergent dynamics when teams of uneven talent density interact with each other? How does executive sponsorship of either side impact things?

• What are the common patterns of companies that achieve talent density through Method 1 vs. 2? Other than the fact that Method 2 would intuitively be more cutthroat (i.e. up or out practices, Elon’s Twitter, Netflix’s high performance culture)? Who does Method 1 best?