She poured me another diet Pepsi, because it wasn’t serving Coke, and I mechanically gave her a chip.

I’ve been at this craps table for 10 hours. Maybe more. Doing the same thing. Bet on don’t pass. Roll. If lucky, lay the maximum odds. And let it roll until I win or lose. And do it all over again. And again. And again.

Craps is considered a social game. But I don’t remember any face. I don’t even recall throwing the dice. I clearly didn’t talk to anyone. I could have played alone; it would have been the same. I chose that particular table at that particular casino because the min bet is $5, which corresponds to my 20x bankroll management. I played the same strategy without any deviation because it is mathematically sound. Or is it.

For years, I’ve been obsessed with the idea of intrinsic and extrinsic incentives. I first learnt about the concept in my 20s from a professor in management, and I saw it materialised in the tech industry. How many people, myself included, started in tech because of the task itself: the fun of building something. And how many are we now performing these tasks because of the pecuniary outcome of it and suffering the over-justification effect?

I had that conversation with a friend recently. He works at one of the top tech firms in the world. And not only his company is really cool, and he worked hard to get there, but he is also working on some of the most insane problems in ML. But he isn’t happy anymore. He wants to change. He even considered quitting. The reason: he believes he isn’t learning anything new anymore. He believes that everything is boring. He believes that he is doing the same thing repeatedly every day. How is that possible when he is working on challenging problems that are literally not solved yet in one of the most amazing sectors doing the job he dreamt of doing as a kid?

“Given the opportunity, players will optimise the fun out of the game”, Soren wrote in one of his notes. And it is exactly what my friend experienced at work and what I experienced at craps. We both started with something fun. I started playing craps because I was interested in that mysterious game that looks so daunting but always gathers a pretty big and loud crowd. My friend started his career in ML because he was intrigued by the possibilities offered by machines, and he picked that company because he was given a platform to express himself through his code. And we both started experimenting with different things that are probably very silly at best. And it was fun. So fun.

Then, at some point, we started to optimise. We understood the system well enough to find a way to abuse it. Or so we think so. In my case, it isn’t hard to do the math and realise that playing the don’t pass gives a tiny little less edge to the house. In my friend’s case, he realised that he could get his job done with elements he had already created. We figured out a way to get to our goal with minimum effort and maximum efficiency in both cases.

Wait. What is the goal again?

As I type, I shiver at how easy it is to divert people from their initial goal. And in my case, I knew about it. I literally spent more than a decade designing and building products with different incentive mechanics in mind. I’ve been mentored, and I guided teams to never lose their north star. I talked and wrote in length about the importance of doing meaningful things and always considering true value… and I still fell into the trap and spent hours and hours doing the same boring task. Think about it: I wanted to play craps because of its social value, and I ended up playing the only way that goes against everyone at the table!

From a product design standpoint, the balance of intrinsic and extrinsic incentives is key because it impacts users’ satisfaction and retention. One of the biggest manifestations of such misunderstanding is when teams confuse product-market-fit with the go-to-market.

When working on PMF, it is all about building intrinsic value: the pleasure of performing the task itself. Is the tool pleasant to use? Is the service delivering a good experience? Is the game fun to play? It can go even deeper. Within a game, for instance: is it satisfying to click heads? Is it fun to collect coins? It is interesting to do a quest?

When it comes to GTM, it is usually more about the extrinsic value. How do I convince someone to try my product? How do I make someone sign up? How do I lead someone to my website? Again, we can go a few levels beyond and look into the product: how should I reward my user for clicking that button? How do I make users test that feature? How do I manage to get users to perform that activity?

Extrinsic incentives are also hard to scale. And if the incentive is money, it also makes it hard to defend. It is why churn in tech companies gets so high when all these engineers work on cool problems and are offered top packages. Employees stopped caring about the task itself because they optimise for their compensations: any MANGA is the same. They choose where to go like how I chose at what table to play: which one gives the better bucks per hour.

Is money the root of all evil, then? I don’t think so. I think the evil resides in bad design. Let’s take a game that has a scoring system. If every run shows players their scores and a breakdown of their mistakes, players will quickly learn how to optimise for higher scores. That could lead them to do things that divert from intrinsic motivation and leads to over-justification.

Now, let’s change that a bit. Instead of showing a score with that breakdown, how about only showing the score once they beat the game once and play that level a second time? Even better: how about only showing their scores if they get into the top 10%?

Imagine only showing achievements or trophies when you finished the game once. Or even better: never show the complete list. Not even indicate how much you’ve covered. Let the players explore by themselves.

How about changing the bonus system in some corporations so it is an actual bonus. In many companies today, employees are often destroyed when they don’t get a big bonus because they somehow believe it is part of their base compensation. So, they feel they’ve failed at their job. And if we take a step back and think about it, it makes no sense. The very definition of a bonus is that it isn’t part of the main set. And making bonuses such as it is considered part of the base leads to a lose-lose situation for both the employer and the employee.

I’m sure there are tons of much better ideas out there. I’m sure there are much more competent designers thinking about these problems at a much deeper level. I invite them to speak up.

As more and more people talk about crypto, a.k.a. “the financialisation of everything”, and more relevant to my industry: “play-to-earn”, we need top designers to think about the right balance between intrinsic and extrinsic incentives. Every blockchain game project openly talks about the importance of making a fun game. Sure. But “given the opportunity, players will optimise the fun out of the game”. Out of any game. Always.