In 1961, the manager of a Metropolitan Life Insurance office in Rockaway drew a new organization chart. Alfred Oberlander put his six best agents in one unit, under his best assistant manager. An equal number of average agents went to an average assistant manager. The remaining low producers went to the least able one.
Oberlander insisted, by J. Sterling Livingston’s account in the Harvard Business Review, that he had never signaled low expectations to them. In his own account they were simply the agents “not considered to have any chance of reaching the half-million-dollar mark.”
The agents got the message anyway. Livingston put it down to the manager’s indifferent manner, and to the chart itself: they took their assignment “as equivalent to a request for their resignation.”
Their productivity fell. More of them left.
Three links in a chain
Livingston described what agents in that position tend to do. Salespeople who have been failing, he argued, protect their egos “by avoiding situations that might lead to greater failure.” They make fewer sales calls, or stop trying to close when closing risks another rejection, or both. That makes failure likelier, which is what everyone expected of the bottom unit in the first place.
Three years after Oberlander’s experiment began, Victor Vroom published a theory that names the break. In Work and Motivation (1964), effort hangs on three beliefs. Expectancy: if I try harder, will I perform better? Instrumentality: if I perform better, will it lead to the outcome? Valence: how much do I want that outcome?
The theory is usually summed up as a product, motivation equals expectancy times instrumentality times valence. As a formula it is loose, but it makes one point well: a zero anywhere wipes out the rest. An agent can want the sale badly and still not pick up the phone, if she no longer believes calling will work.
Effort follows three beliefs: I can do it, it will count, and the reward is worth having.
You won’t feed coins into a vending machine you think is broken, however thirsty you are. Effort is the coin.
Livingston had a factory example for the first link. A large electrical manufacturer found that production actually fell when quotas were set too high, “because the workers simply stopped trying to meet them.” The second link can break too. At Microsoft under its old ranking system, Vanity Fair reported, even meeting every one of your goals did not guarantee a high rating, because a colleague might beat theirs by more (see Forced Ranking). And the third breaks quietly whenever the prize is something people don’t much want.
What the studies found
Expectancy theory went on to hold, in one review’s words, “a major position in the study of work motivation,” and it was tested for decades. In 1996, Wendelien Van Eerde and Henk Thierry pooled 77 studies of Vroom’s models and their three parts, against measures such as performance, effort, intention and choice. The average correlations were “somewhat lower than reported in previous narrative reviews,” and the measurement differences they tested did not explain why results varied so much from study to study.
Two economists, Randolph Sloof and Mirjam van Praag, later summarized the meta-analysis’s warnings. Most studies relied on people’s own reports of their beliefs, often next to self-reported effort, which can inflate the link. Most compared different people, while Vroom’s idea was about the choices inside one person. And correlations can’t show which way the arrow points: belief may drive effort, or effort may drive belief.
Sloof and van Praag then ran their own lab experiment on the first link. Making the connection between effort and measured performance noisier did not change how much effort their participants chose, which they said casts doubt on that part of the theory.
So the three links are better used as questions than as a formula. Expectancy theory doesn’t tell you how much a weak link costs. It tells you where to look.
The bonus nobody chased
Say Tomás runs a support team and announces a bonus for anyone who closes 120 tickets in a week. His best agent closes about 80 on a good week. Nobody changes pace, and he decides the team doesn’t care about money.
Each link points somewhere different. If the agents think 120 is out of reach, more money won’t help, and a lower first rung or better tools might. If they doubt the count will be fair, because hard tickets take longer, the fix is in how tickets are counted. If they’d rather have a day off than the cash, the prize is the problem. One possible outcome of guessing wrong is a bigger bonus that nobody chases either.
Back in Rockaway, the middle unit’s assistant manager took the opposite line from the chart. She told her agents they had more potential than the super-staff and lacked only experience. By Livingston’s account, her unit then raised its productivity by a higher percentage each year than the stars did. The bottom unit had read its chances off an organization chart, and it lost people.