In 2008 the world went through a severe financial crisis. A large part of our economy collapsed, in large part because financial institutions had created a pretty unsustainable system of mortgage-backed securities.
It was pretty bad.
Fingers were pointed; new rules put in place; people yelled at.
“Oh them financial folk, they’re all the same!” we said. “They only care about the money. Evil, evil people! They will be the ruin of us all!”
And then, we all went on with our lives. Crossing our fingers that this wouldn’t happen again.
What we may not realize is that what happened in the financial industry is not just a fluke, it’s pointing to an inherent problem in our approach to motivating people. And it’s in no way specific to financial institutions. “Financial folk” are not “evil,” we made them this way. We held up a carrot for them to run after, and so they did. And then another one, and another, and another.
And, while perhaps to a less extreme extent, the same is happing in your organization. We use carrots all over the place. We have “pay for performance” programs. We tell our people “good job!” We “align incentives with our company goals.” We promise people a promotion if they do X. We call out people in our all hands, implicitly telling the rest: “be more like them!”
Have we ever stopped and asked: Is this working? Are there any hidden costs to this? Is this sustainable?
Nah! No worries. This is a good practice. Everybody does this. We’re not bad people like them financial folk.
I’m sure we will be just fine.
While the understanding of the topic of motivation in psychology has grown a lot in the last few decades, common thinking is still dominated by what is referred to as behaviorism, the state of the psychological art in the early ‘70s.
One of the biggest influencers around behaviorist theory is B.F. Skinner, an American psychologist, author, and psychology professor at Harvard during the ’50-‘70s.
Behaviorist theory claims that all behavior — both of animals and humans, although most experiments used animals such as pigeons — can be explained through operant conditioning.
The model is nice and simple: we can influence people’s behavior through incentives. If they do something we like, we use a positive incentive — a reward. If they do something we don’t like, we use a negative incentive — a punishment. By setting up the right incentive structure, we can control behavior.
Skinner’s belief in this model was so strong that he claimed that free will is essentially an illusion — instead, anything people do is a result of previous actions and their consequences. In other words: the fact that this very evening I sit behind my keyboard writing these very words, can be completely explained by a series of actions I have taken up to this point, and the consequences (positive or negative) that resulted. Essentially, I can be reduced to a set of behaviors that trigger when certain events occur.
Thus, according to behaviorism, there is only a single type of motivation: extrinsic motivation. Motivation incited from the outside, through incentives.
While this mental model is rather nihilist, it is elegant and simple.
Perhaps that’s why it stuck.
Even today, most of our “best practices” around how to motivate people are based on behaviorist ideas.
However, this was the state of research in the 1970s. Progress in understanding human motivation did not stop there.
In the 1970s-1980s, a new theory emerged called Self-Determination Theory (SDT). Research on self-determination theory found there is not just one type of motivation — extrinsic motivation as theorized by behaviorism — but two:
Key studies that led to emergence of SDT included research on intrinsic motivation. Intrinsic motivation refers to initiating an activity because it is interesting and satisfying in itself to do so, as opposed to doing an activity for the purpose of obtaining an external goal (extrinsic motivation). A taxonomy of motivations have been described based on the degree to which they are internalized. Internalization refers to the active attempt to transform an extrinsic motive into personally endorsed values and thus assimilate behavioral regulations that were originally external.
Self-determination theory concluded there are two types of motivation with very different properties:
- Intrinsic motivation is the natural, inherent drive to seek out challenges and new possibilities. Intrinsic motivation can grow naturally, under the right circumstances. It doesn’t need constant reinforcement, it persists unless actively pushed out.
- Extrinsic motivation as described earlier comes from external sources. Compared to intrinsic motivation, extrinsic motivation is easier to control from the outside (hence extrinsic), but it is also short lived — it requires constant reinforcement.
So, according to SDT we are not just extrinsic-motivator trained neural networks on legs. Wouldn’t that be something?
SDT concluded that intrinsic motivation comes from three basic needs:
- Autonomy — our ability to have control our own acts and life.
- Competence — our need to control outcomes and experience mastery.
- Relatedness our will to interact with, be connected to others (a sense belonging), and contributing to something bigger than ourselves (purpose).
Intrinsic motivation is present in all of us from birth. Every kid is naturally curious, wants autonomy, wants to master new skills and be close to people around them (initially parents, brothers and sisters).
However, as children grow up, their intrinsic motivation is under attack. It is under attack from the outside: from extrinsic motivators.
The more we — as parents, teachers, or later in life: managers — use incentives such as rewards and punishments to control, the more we’re in trouble.
And in trouble we are.
The effects of rewards and punishment
Alfie Kohn, in the ’90s wrote a book on the topic of punishments and rewards entitled Punished By Rewards. In its 450 pages, out of which about 40 are just references to all the research and experiments he’s citing to back up his claims, Kohn demonstrates six things about the effects of rewards and punishments.
And to make this a bit more buzzfeed-y: “…and you will never believe what he found!”
Before we get into the findings, let’s clearly define what we mean with reward and punishment.
A reward or punishment is something in the shape of: “If you do this, you will get that.” Obviously, in the case or rewards “that” is desirable, in the case of a punishment it is not.
Examples of rewards include:
- “Pay for performance:” bonuses and commissions (you may be tempted to slot in salary here too, but I’ll get back to that later).
- Other monetary incentives (e.g. referral bonuses for recruitment).
- Praise (“well done!”)
- (Performance) ratings
Here are Kohn’s findings summarized briefly. In the book these are backed up with a lot of research and nuance. I recommend you go read the book, I’ll give you a cookie if you do. That was a joke you’ll appreciate in a minute.
First, of course rewards do have effect. Of course they do. They are effective to get somebody to do something that they would otherwise not do.
Second, extrinsically motivated work (work motivated by rewards) is of worse quality. Because the thing we’re after is the reward we want to get, the means to get there plays second fiddle: “I want the reward, what is the shortest and easiest path to get it?” Note that one way this “worse quality” can materialize is that it completely disregards long-term effects (such as what happened in the 2008 market crash).
Third, the effect of rewards does not persist after you stop the rewards. Following the first finding: the reason they do the thing you ask is the pursuit of the reward. If the reward goes away, so does the pursuit. You may hope that rewards may somehow push new behavior into people’s systems in a durable way. It does not.
Except… in a way it does. There is one effect that persists. Sadly.
Finding four: extrinsic motivators eat away from intrinsic motivation.
Here’s an experiment that was conducted (and replicated in various forms over time):
Students were invited to solve a set of puzzles. Half of the students were given rewards (let’s say $10 for every puzzle solved), the other half weren’t. After the experiment ended, the students were told to wait for some time in the room. Guess what happened? The students that were rewarded for solving puzzles, instantly dropped the puzzles once the experiment was over and didn’t touch them again. The group that was not rewarded, kept playing. This is one of many experiments that would show that not only do extrinsic motivators stop working the moment you stop using them, they have a detrimental effect on intrinsic motivation that lasts.
In other words: the worst you can do to somebody who’s naturally motivated to do something, is reward them for it.
There are two possible explanations why this may be the case:
- Because of the “if you do this, you will get that” you implicitly communicate that “that” is more desirable than “this.” This may reduce your overall interest in “this.”
- People will realize you are attempting to control them, and this backfires into disinterest in the activity.
Fifth finding: any limit on the quantity of rewards given out, results in less cooperation. If you say “one of you can win this award,” it’s every man or woman for himself or herself. People will elbow their way towards the reward. If you care about cooperation between people, inviting this type of competition is a highly effective way to kill it.
Sixth, rewards emphasize an unbalanced relationship — a clear power structure. Since you’re the one in power to reward, you’re effectively saying “I am the one in control of you, because I decide if, how, and for what you will be rewarded. Bow before me, underling!” We really have to think if emphasizing this power imbalance will lead to a productive relationship with our people.
In short: rewards are bad news. Stop using them.
So if this has been known for over 50 years, why are we still relying on rewards so heavily today?
First, they are easy. We can apply them easily, often by throwing money at the problem. Our people don’t write any unit tests, how do we fix that? I know! Let’s fire them if they don’t! Or, less aggressively: let’s give them a small bonus every time they do.
Second, they work… short term. Often we cannot clearly see, or bother to check, the longer-lasting impact, nor the quality. What will the quality of unit tests written out of fear of being fired be, would you imagine? What long lasting effects of putting such punishments or rewards in place can we expect? Usually, we don’t check. Unit test coverage is going up — mission accomplished!
Sure, occasionally there’s a visible blip like the 2008 financial crisis we kicked off with where people were more focused on getting their bonuses and commissions than… I don’t know… keeping banks running sustainably. I wonder why that happened 🤔.
Third, this is what we’ve always done, everybody else is doing it! We can not be the only one not using rewards, people expect them. They would quit without them!
No more salary
Speaking of quitting… what about salary? Salaries are rewards! Should we now all become volunteer organizations?
This is a subtle topic. Indeed, if you stop paying people, they tend to stop showing up. How should we think about that? Do we need to rely on this type of extrinsic motivator after all?
We can definitely describe salary as a reward: “If you work for us, you will get a salary” following that same “if you do this, you will get that” formula.
Here’s my thinking:
In practice, at least in our line of work (software engineering — I know this is different in some other professions, such as sales or many jobs in the financial industry), salaries is are not very directly tied to our performance day to day. Of course, there is a connection, but mostly in the extremes: if we do very poorly, we are likely to get fired and thus get no more money. On the other extreme, if we perform very well, we get a raise.
In those two extremes, salary changes could function as punishments and rewards, respectively. Although we could also look at them as an adjustment to what is a fair amount to pay for your services. For salary, a perception of being fair is critical.
The impact of the prospect of a salary raise is different based on the person. Personally, I think my need to do the best I can at work is intrinsic, and not really motivated by the perspective of earning more money at some point. I know some other people feel this differently.
Yet, most of the time, salary is just background noise. It needs to be there, but it doesn’t act as a motivator day to day. The day I get my salary I do not feel more motivated to work than the day before, do you? This is likely a hint that we don’t perceive our salary as a reward in practice. And it’s this perception that matters.
So… what instead?
If only there was a cheap replacement for rewards, amirite? If there was, we probably would’ve switched long ago. But, nope, there isn’t.
When you think about it, it would be rather ironic if we could increase intrinsic motivation from the outside — and thus extrinsically. That won’t work.
However, what we can do is create an environment that nurtures intrinsic motivation. This starts with eliminating anything that actively kills intrinsic motivation, such as rewards and punishment. The good thing about doing this is that is saves us a lot of time we would otherwise spend designing and operating incentive programs, evaluating people and so on. There’s opportunity cost to everything. We can now spend this time differently.
How should we spend it? Self-determination theory, as mentioned, tells us what fosters intrinsic motivation:
- Autonomy — our ability to have control our own acts and life. We can support this by reducing control, and leading with context instead.
- Competence — our need to control outcomes and experience mastery. Give people the space to improve their skills, feel competent and to always be sharpening the saw.
- Relatedness our will to interact with, be connected to others (a sense belonging), and contributing to something bigger than ourselves (purpose).
On that last item of “relatedness,” let me share something I’ve been doing lately as a replacement for praise at work.
I’ve been referring to this as context hugs. It’s an evolution of something I hinted at some time ago:
Then I’d tell them what I thought was the significance of their talk: “I think a lot of people learned something about CloudFront because of your talk, CDNs are not something all are exposed to day-to-day and I think it’s healthy to be more aware of this part of our stack.” It’s not feedback, it hints at the significance of the effort. Context as a gift, if you will.
I now do this frequently.
Rather than using praise: “Good job on that delivering that feature!”
I connect it to purpose: “Our customers will be so happy to finally be able to use that feature, they have been asking for it for years!”
It’s a simple concept, and it only requires a bit more effort to implement than the default “Awesome job!” “You’re awesome!” “Good code typing!” Pat on the back.
The theory would suggest (although I have no tangible proof yet) that this should nurture intrinsic motivation. It helps the receiver remember why they skipped a night of sleep to get this done in time.
I know that reminding myself about the purpose of my work got me through dark times at work, so it likely will have an effect on others as well.
Stopping the use of rewards cold turkey will be hard. It’s a practice that is so deeply ingrained into all of us, we cannot just shake it off overnight. The previous section gave some ideas about alternatives, but there are some other things we can do to “control the damage.”
If we are not in a position to remove rewards entirely, we can at least deemphasize them. Don’t focus on them too much. You can mention they’re there, but not hold them up as a carrot. No “we should really complete this project, because bonus!” but “we should really complete this project, because purpose!”
No more “step it up people, performance reviews are coming!”
If you are somehow involved in the development of reward-based systems, raise awareness of the costs and risks. First, do no harm. It’s conceivable there are valid reasons to introduce a reward, but explore every avenue to come up with alternatives and prepare to pay the cost down the line.
No reward is free, and likely money will not be the biggest cost.