- Thomas Thurston
Do great leaders improve a business’s odds? What does the data say?
I use algorithms to predict if businesses will survive or fail. In this line of work, probably the #1 most asked question is “how much do you look at the leadership team?” Said another way, does the quality of a business’s leadership change its odds of success? It’s a venture capital cliche to bet on the “jockey” (leadership team) rather than the “horse” (the business itself). Most businesspeople, gurus, authors and consultants feel strongly that great leadership makes a huge difference, especially in early-stage businesses, but what does the data say?
Answer: leaders may matter, but only a little bit.
Statistically speaking, there’s some data to suggest great teams can increase a firm’s odds of success. For example, a Harvard study found entrepreneurs with prior exits more likely to have future exits – compared with entrepreneurs who failed before or first-time entrepreneurs (who tended to perform the worst). Yet if you look closely at the study, the difference in outcomes between the best and worst teams was a mere 12%. In other words, having founders with prior startup success can increase a firm’s odds, but not by much (it still doesn’t account for 88% of the whole equation).
This conclusion isn’t new. Even in the 1970s the impact of leadership was studied in 167 companies spanning a 20-year period, concluding that the business itself and its industry have a much bigger impact on sales and profits than leadership.(i)
Large samples of CEOs again showed business success to be mostly determined by factors beyond any leader’s control.(ii)
More research found that changing the CEO has no statistically significant
Another recent study looked at 8,400 businesses, slicing and dicing leadership teams and success criteria in a variety of ways. It found no statistically significant correlation between good teams and success.
Still not convinced? A big round-up of leadership research going as far back as 1977 found that, while leaders can have some impact, their actions rarely explain more than 10% of the performance difference between the best and worst organizations.(iv)
Right about now I imagine some folks are shaking their heads in disbelief. We all know good managers are better than bad managers. We know some leaders are toxic and can destroy any business they so much as glance at. Many of us have personally felt the incredible difference a great leader can make. So how do we square our intuitions and experiences with the cold, hard, statistical facts?
We need to realize our personal intuitions and experiences are probably dealing with tiny sample sizes. Even if we have a new leader every 2 years, for 50 consecutive years, we still don’t have enough data (in our personal experience) to meet the minimum sample size requirements for statistical confidence. No matter what we experience either directly or vicariously, let’s face it – our intuitions may only be dealing with a small part of a much bigger, counterintuitive equation. We also have to admit our intuitions may not always do a great job canceling out tons of other variables that add signal or noise to our personal world-views.
What decades of empirical research has found, and what we’ve also found (as data scientists who research these issues), is this:
The business itself, not the leadership team, is the largest predictor of firm survival or failure.
Leaders who’ve been successful before are more likely to be successful again, but only by a little bit (around 12%).
Bad leaders can lower your odds a lot more than great leaders can raise them. So success is less about finding great leaders and more about avoiding awful ones.
Rather than worrying too much about how great your team is or isn’t, the most important thing is to have the right business model and leaders that are at least good enough to not mess it up. As a CEO it’s more than a little humbling to think I could be replaced without changing my business’s odds too much. Yet as a scientist I have to accept this, like it or not, and in a weird way it makes me feel a little better about things. Hey – as long as the business model is sound, I don’t have to be a rockstar – I just have to be good enough to not completely blow it. Kind of a relief. At least it can help me stay focused on the right things (the business, not myself).
For all the die-hard leadership worshipers out there, and there are tons of you (from the typical VC who cares more about the team than anything else, to Jim Collins, to the lowly MBA with a stack of leadership books in the closet), you may have a small point. However you’re grossly overweighting the impact of leadership when you should be focusing on other parts of the business instead.
Anyone who says “bet on the jockey, not the horse” hasn’t done their homework. It’s easy and fun to attribute success or failure to human-centric causality. Humans love to anthropomorphize stuff. But in this case it’s mostly wrong.
If you want a better rule of thumb – one that squares with the data – Warren Buffett said it best: “good jockeys will do well on good horses, but not on broken down nags.”
(i) Lieberson & O’Connor, Leadership and Organizational Performance: A Study of Large Organizations, American Sociological Review 37 (1972)
(ii) See Pfeffer & Blake, Administrative Succession and Organizational Performance: How Administrative Experience Mediates the Succession Effect, Academy of Managemnet Journal 29 (1986); Pfeffer & Sutton, Hard Facts, Dangerous Half-Truths & Total Nonesense; Profiting from Evidence-Based Management, (2006); Pfeffer, The Ambiguity of Leadership, Academy of Management Review 2 (1977)
(iii) Carroll & Hannan, The Demography of Corporations and Industries, Princeton University Press (2000)
(iv) Pfeffer & Sutton, Hard Facts, Dangerous Half-Truths & Total Nonesense; Profiting from Evidence-Based Management (2006)
#BusinessModelSimulation #GrowthScience #Impactofteams #leadershipdata