Why Rewarding People for Failure Makes Sense: Paying "Kill Fees" for Bad Projects

The notion that companies ought to reward people for failure and punish them for success is, at best, a dangerous half-truth.  A high failure rate is a hallmark of innovation.  Whether we are talking about products, new companies, or new business processes, there is little evidence that aiming to reduce failure rates is a useful strategy.

U.C. Davis Professor Dean Keith Simonton, who has spent much of his career doing long-term quantitative studies of creative genius,  has concluded that a high failure rate is a hallmark of creative geniuses -- he concludes that the most creative people -- scientists,  composers, artists, authors, and on and on -- have the greatest number of failures because they do the most stuff.  And he can find little evidence that creative geniuses have a higher success rate than their more ordinary counterparts; they just take more swings at the ball. Check out his book Origins for Genius , perhaps the most complete review of research on the subject. 

The upshot of all this is that the most creative people -- and companies -- don't have lower failure rates, they fail faster and cheaper, and perhaps learn more from their setbacks, than their competitors.  One of the biggest impediments to faster and cheaper failures is that once people have made a public commitment to some course of action and have devoted a lot of time and energy to it, they become convinced that what they are doing valuable independently of the facts.  My colleague and friend Barry Staw at the Haas Business School has devoted much of his career to studying this process of "escalating commitment to a failing course of action."  Barry shows through a host of experiments, field studies, and case studies that such irrational devotion can be extremely destructive and remarkably hard to stop once it starts.

One antidote to such misguided commitment is provide people incentives for pulling the plug as early as possible on failing projects. Merck, the giant pharmaceutical firm, is doing a host of things to improve their innovation process these days, and following Staw's research, Peter Kim, the new head of R&D has instituted what they call "kill fees"" at Merck, paying out serious dollars to scientists who pull the plug on failing projects.  As BusinessWeek reported:

'An inability to admit failure leads to inefficiencies. A scientist may spend months and tens of thousands of dollars studying a compound, hoping for a result he or she knows likely won't come, rather than pitching in on a project with a better chance of turning into a viable drug. So Kim is promising stock options to scientists who bail out on losing projects. It's not the loss per se that's being rewarded but the decision to accept failure and move on. "You can't change the truth. You can only delay how long it takes to find it out," Kim says. "If you're a good scientist, you want to spend your time and the company's money on something that's going to lead to success."'

If you blend together research suggesting that failing faster rather than failing less often is essential to innovation, that an action orientation is essential to innovation,  as well as research suggesting that so-called experts aren't very good at guessing which new ideas will succeed and fail, you can see why I proposed in "Weird Ideas That Work" that creativity is sparked when organizations "reward success and failure, punish inaction."  It may sound really weird, but in addition to the evidence that supports it, Merck seems to be doing it. And so do a lot of other creative organizations.

When I  really want to get executives upset, I sometimes propose that they reward failure MORE than success when they are managing creative work.  I am not sure if I believe it is a good idea, but having the discussion can be pretty interesting.

          

Comments

[Cole Taylor]

So Lavinia would kick Edison out the door ... wow.

I found a typo :

what they are doing valuable independently of the facts.

Sorry, it's a sad obsession, I know.

Nice article, however.

Cole

[Alan Rimm-Kaufman]

Hi Bob --

Nice post.

Mike Moran (search guy @ IBM with new book out) calls this "do it wrong quickly".

The agile development community warns against committing too deeply too quickly, YAGNI.

All the same idea: to hit homeruns, you need as many at bats as you can get.

Related ideas: http://www.rimmkaufman.com/rkgblog/2007/10/16/rule-of-three/

Cheers --

Alan

[Chuck]

Isn't the reward for fast failure simply greater successes?

Edison failed quickly so that he could find a solution more quickly. Is there a way to just get people more directly connected to that sort of ownership?

[Steve Roberts]

Working on a successful project is a reward in itself, and brings other rewards - especially career advancement. Killing a project that is not going to succeed is real work - difficult to do & it takes courage. Why not reward such a valuable act ?

[Lavinia Weissman]

At one time, Procter and Gamble, while led by Disney's current CEO, John Pepper, rewarded brand managers for retiring brands. In this case, it was about getting out of date products off the shelf instead of explicitly rewarding failure.

There is also opportunity to reward for innovation of ideas in cycles of leading change -- from submitting ideas to executing them.

Possibly in my opinion the greatest leadership challenge today is "executing new ideas" and "inviting imagination."

Innovation occurs in fewer than 10% of work environments. Most companies do not invite this and want to only hedge their bets on the sure thing, so they turn to labs like Ideo.

Part of the talent challenge today for recruitment is inviting people to learn to become talented and the complaint there is no talent to me needs to be retired and investigated. One remedy I believe is in this world of virtual work and one line responses is to learn how to invite "learning" or what you call failure and support a dialogue of work effectiveness which won't grow out of virtual communications without first building trust and designing deliberate events and synchronistic communications to build a foundation to begin.

So what are conditions for learning and how to you determine and build trust to invest?

[Herman Najoli]

Bob, I think outrageous success should be enthusiastically rewarded and mediocre success should be thorougly punished. First time failure, on the other hand, should be cautiously rewarded. Repeated failure should be slammed and consistent failure should be drop-kicked out of the organization.

[sean]

Does that first line have it backward?

The standard approach is to reward for success and punish failure.