[HOWTO]How to design a simple token economy for a small feature?

Hi, guys, I am struggling to design even a very simple token economy for a small feature. Every time I feel that there is a need to incentivize people for doing GOOD, I feel that I should also have a way to Penalize them.

For example;

Imagine you have a social networking app where people can “like” very good content. The more number of likes signifies - the content is more good. You decide to pay, say $1, for every like to the curator.
The problem here is, the moment people see this feature, they will ONLY come to the platform JUST to put a “like” so that they can earn money. thinking the more they like, the more they earn.
Now to solve this you need to think a way to penalize them, but how would you do that?

So Incentivizing people for doing GOOD is very hard. My question is, how to learn to work on these kinds of problems? Any paper or any reference would be highly appreciated.


I am not sure if you’re familiar with Goodhart’s law (link for wikipedia) but it is usually known as:

When a measure becomes a target, it ceases to be a good measure.

The problem above is that you want to make the measure so precise that it would penalise being a target. I think you’re thinking in the wrong direction.

One of the examples @uwiger came up for State Channels use case a few years back was a paywall mechanism where you pay for the amount of content you actually consume. Think of a movie with a great trailer but once you start watching it you realise it is not as good as you’ve expected it to be. You’ve already paid the full price though, so you’ve been cheated. If on the contrary you had been paying for the amount of time you’ve actually watched it - this would have been a great measure if you enjoyed it or not. Even more - if I see that people who’s movie taste I respect have watched this movie to the end - I might as well watch it. Here you go your social network and the incentive is put in the right place: where people put their own money.

This can easily be translated to academic papers (where you can see for free only the summary) or any paywalled content, really.