The Cobra Effect: why managing by metrics backfires

With the growing momentum of the “No Kings” protests, activists have increasingly turned to the 3.5% rule—Erica Chenoweth’s observation, based on over a century of historical data, that once a protest movement mobilizes 3.5% of the population, it achieves its goals within a year. As a result, many have begun to treat the 3.5% threshold as a primary objective.

Not so fast. Even Chenoweth herself has cautioned that the rule is “a descriptive finding, not necessarily a prescriptive one.” Her subsequent research has shown that at least one uprising—in Bahrain—reached a 6% participation rate and still failed. In fact, most successful movements never reach the 3.5% level at all. 

So hitting the threshold doesn’t guarantee success, and failing to reach it doesn’t guarantee failure. The 3.5% rule is a great example of the Cobra Effect, also known as the Cobra Paradox, or Goodhart’s Law, which states that when a measure becomes a target, it ceases to be a good measure. Metrics can be useful signals, but don’t substitute for a clear vision and a strategy to achieve it. 

The Cobra Effect explained

The Cobra Effect gets its name from a campaign during the British Raj to reduce the cobra population in India. Authorities offered a bounty for dead cobras, paying anyone who could prove they had killed one by bringing in the carcass. Yet the policy failed. In fact, the incentives didn’t reduce the cobra population at all. 

What happened was that entrepreneurial Indians began breeding cobras in order to kill them and collect the bounty. After all, who in their right mind would risk their lives by going around hunting deadly snakes in the wild when it was much easier to raise them in captivity and kill them in a much safer environment. 

Something similar seems to have happened with respect to uprisings against authoritarian regimes. In his 2014 bestseller, The End of Power, political scientist Moisés Naím examined the success of uprisings like the color revolutions and theArab Spring, as well as how startups were upending major corporations and argued, “power is easier to get but harder to use or keep.”

Yet in his 2023 book, The Revenge of Power, he shows how many of those same players adapted, using similar tools, such as the Internet and social media, to reconsolidate their power. Today, autocrats like Vladimir Putin and Xi Jinping, as well as tech oligarchs like Elon Musk and Peter Thiel, are enjoying a heyday.

As soon as a system is established, players begin gaming it. 

Erica Chenoweth is a serious researcher, but very different from figures like Timothy Snyder, Anne Applebaum, and Ruth Ben-Ghiat, who feature much more prominently in the media. Those scholars study authoritarian regimes, while Chenoweth studies how they are overthrown. For reasons that aren’t entirely clear, there is very little overlap between the two groups. 

Chenoweth’s field of interest was pioneered by Gene Sharp, whose key insight was that there are sources of power that support the status quo and that these have an institutional basis. It really doesn’t matter how many people you get to go into the streets. As long as that support remains in place, nothing will change. 

To understand why, consider the history of marches on Washington. The first one of major significance was the Woman Suffrage Procession in 1913. It was a disaster, quickly turned into a riot and greatly discredited the movement. Since then, there have been roughly 300 marches, almost all of which were either largely ineffective or total busts.

The one major exception—the 1963 march where Martin Luther King Jr. gave his “I Have a Dream” speech—proves the rule. In that case, President Johnson was trying to push major civil rights legislation through Congress. That march, which included not only activists but a broad coalition of religious and labor organizations, was designed to support those institutional efforts. 

We see the same dynamic in organizational change. When we work with clients, we map the sources of power—both internal and external—that can affect the outcome of the change initiative and design the change program based on that. Your institutional targets determine your tactics, not the other way around.  

Managing the cultural mechanisms of change

One reason that institutions are so important is that they determine the norms and rituals that shape behaviors. For example, if employers adopt the norm that a college degree is required to get a good job, then that influences the behaviors of parents, students, and teachers. You can complain all you want, but unless employers change their minds, that won’t change.

Rituals then develop—such as the college admissions process, on-campus interviews, and summer internships—that encode that norm into the culture of the professional class and the broader society as a whole. Once graduates enter the workforce and have kids of their own, they start college funds to perpetuate it further. 

Or consider how the giants of the PC industry struggled to adopt the Dell Direct model that bypassed the reseller channel and sold customized computers directly to customers. The model gave Dell a clear cost advantage by eliminating dealer markups—and even more importantly, it allowed the company to receive payment before paying suppliers.

Still, none of the incumbent industry giants, such as Compaq and HP, managed to adopt it. It wasn’t for lack of trying. The advantages of Dell’s model were highly publicized and well known. Firms launched initiatives to emulate it. But the cultural norms and rituals that drove the behavior of stakeholders supported the existing model and nothing ever changed. 

Now you can see why Chenoweth warned that her research was “descriptive, not prescriptive.”

Yes, historically the 3.5% threshold was significant, but without an institutional shift to reshape norms and rituals, the status quo will prevail. 

Manage for mission, not for metrics

Humans don’t always react well to abstractions. We want our leaders to be specific and concrete about their expectations. That’s why we so often try to boil things down to a single metric that can be optimized. That metric then acts as a proxy for success, something everyone can understand and work toward. 

So activists adopt the goal of getting 3.5% of the population. Economists focus on GDP to evaluate prosperity. CEOs adopt shareholder value as the measure of their own performance. Yet people marching in the streets does not, by itself, create institutional change, just as GDP does not encapsulate the well-being of a society and a stock price does not tell you how well a company is serving its customers, employees, and other stakeholders.

Yet those metrics often create perverse incentives that shape behavior. An activist focused solely on populating protests is unlikely to do the kind of institutional work needed to bring change to a society. A government consumed by GDP will likely miss important indicators of societal well-being, such as health and happiness. A CEO rewarded for increasing the stock price is likely to focus on quick fixes like layoffs and stock buybacks rather than doing the hard work of improving operations. 

That’s why it’s important to manage for mission, not for metrics. Or as the Polish activist and journalist Adam Michnik put it:

“Do you believe in free speech? Then speak freely. Do you love the truth? Then tell it. Do you believe in an open society? Then act in the open. Do you believe in a decent and humane society? Then behave decently and humanely.”

In other words, if you want to make an impact, think about the end state you want to achieve. There will never be a shortage of influencers, gurus, and other hucksters who try to sell you on a simpler answer in the form of a concrete metric supposedly backed by some form of pseudoscience. Don’t listen to them. 

Anybody can create a number and call it an achievement. To truly accomplish something worthwhile, you need to define a mission and attract others to it.