In 2005, British psychologists Michelle Ryan and Alexander Haslam published a study with a provocative conclusion. The catalyst was an article in The Times claiming that companies with women on their boards of directors perform worse. Ryan and Haslam tested the cause-and-effect relationship—and discovered that it works in reverse. Women were not appointed before performance declined, but after. Companies experiencing a sustained drop in performance were significantly more likely to bring women onto their boards of directors than companies with stable metrics.
The phenomenon was termed the glass cliff—an extension of the glass ceiling metaphor. The ceiling keeps you from moving up. The cliff lets you up, but at the exact moment when the position is most precarious and the likelihood of failure is at its highest.
Subsequent research confirmed the effect both in laboratory settings and with real-world data—not only regarding gender, but also regarding other groups traditionally underrepresented in leadership.
The mechanism, apparently, does not come down to conscious calculation. Several explanations operate simultaneously.
First: in a crisis situation, perceptions change regarding what kind of leader is needed. A stable company is thought to require decisiveness and toughness, while a company in crisis needs the ability to understand people, withstand pressure, and manage the team's emotional state. These qualities are stereotypically associated with women. The appointment seems like a rational decision, even though it is rooted in a stereotype.
Second: in times of crisis, the willingness to deviate from the usual choice increases. As long as everything is working, a company replicates the familiar leader profile. When it is not working, a window opens for an unconventional solution. The window opens precisely when the risk is at its maximum.
Third, and most unpleasant: candidates from underrepresented groups are more likely to accept risky positions because they understand that other offers at this level may never come. Choosing between a poor chance and no chance at all is not a free choice.
The consequences extend beyond a single career. If someone takes leadership at a moment when decline is already underway, the probability of failure is high regardless of the quality of their decisions. And failure is then interpreted as confirmation of initial doubts—reinforcing the notion that the group handles leadership roles poorly.
What to do about this practically. Record the context of the appointment: was the position in crisis at the moment the leader arrived. Without this, it is impossible to distinguish a leadership failure from the failure of an inherited situation. Agree on resources, authority, and the evaluation horizon before accepting the role—this is not a privilege, but a prerequisite for being able to change anything. Evaluate the result relative to the initial trajectory upon entry, rather than against absolute metrics. And separately monitor under what circumstances unconventional candidates are appointed in your organization: if they consistently occur in crisis situations, this is a diagnostic signal of how promotions are handled during stable times.