Incentives – why they almost never work

Home/Complexity/Incentives – why they almost never work

Incentives – why they almost never work

carrot-and-stickIn most organisations we engage with, the first thing that is considered whenever a behaviour change is required is an incentive scheme.  If we need better customer experience ratings, let’s incentivise the front-line staff; if we need people to share knowledge, let’s link that behaviour to their KPA’s; and so on, and so on.

The problem is that incentives almost never have the desired impact, in fact, often quite the opposite.  We often refer to Goodhart’s Law: When a measure becomes a target it ceases to be a measure.  We’ve seen this in operation more times than we can count.  The latest example I’ve come across is this article in the NY Times, which talks about how teachers are prone to cheat if their pay (or job security) is linked to the performance of their students on standardised tests.  In one of the projects we’re currently busy with, one of the key questions is around the measuring of the impact made on the education system by various interventions aiming to make significant improvements in schools.  One thing is certain, simply measuring the performance of the learners (e.g. how many pass matric) will not be sufficient in itself, hard measures like that is simply too open to gaming.

I came across another interesting article by Chip and Dan Heath on “Why Incentives are Effective, Irresistible, and Almost Certain to Backfire.”   They talk about a concept called a “focusing illusion”.  The essense of this is that humans tend to reduce the world to a one-variable equation, it distorts our judgements about ourselves and others.  The article mentions the following experiment:

“In another study, some college students were asked, “How happy are you?” and then “How many dates did you have last month?” The researchers found a pretty weak correlation between the level of happiness and the number of dates. But then (hilariously) the researchers flipped the order of the two questions. Suddenly, there was a strong correlation. Having just confessed to a lack of dates, students reported that their lives were joyless.”

When considering incentives it’s crucial to understand the complex contexts that those incentives will be operating in.  People are not one-dimentional beings, and they very seldom act or react in predictable ways.  Putting people under pressure to perform, especially in competition with others with “winner takes all stakes usually have unintended consequences.  People are very creative when it comes to breaking rules and finding loopholes in order to win or save face.  Apart from cheating, there are many other potential collateral damage, for example in a large investment firm, the training and mentoring of new employees were completely compromised.  Seasoned traders effectively ignored new employees and their questions.  Why?  Every minute spent talking meant a minute wasted that could’ve been spent on earning their monthly bonuses.

In South Africa, one of the industries that is riddled with incentive issues is the mining industry.  In general, mine worker’s bonuses depend on production.  The more coal or gold they manage to get out of the ground, the more money they earn.  Apart from production, the main focus of mines nowadays is safety, attaining the elusive “Zero Harm”.  So, now you have a bunch of people going all out to reach a production target, hampered by restrictive safety rules.  It’s not difficult to see everything that can go wrong in this scenario.  Now many mining houses are considering a ‘safety bonus’, I can only guess what the impact of that will be, but I doubt it’ll be good.

Incentives are often in tension with penalties of some sort, whether financial i.e. not getting a bonus,  or even losing your job.  I wonder how many employees are walking around feeling something like this …

Carrot & Stick

By |July 7th, 2010|Categories: Complexity|Tags: , |0 Comments

About the Author:

Leave A Comment