Awarding the Right Behaviors in Digital Design

Digital badges to encourage behavior are a fine idea in concept but riddled with issues in practice. I talked about badges at UXPA International last week (slides here), including a whole host of reasons why they go awry. One of the biggest ones, in my opinion, is that designers may choose to award digital badges for behaviors that aren’t really critical ones for obtaining meaningful outcomes. Instead, they reward behaviors that are easy to measure (like clicks or check-ins). The result is a reward system that doesn’t actually lead to results.

The classic example of digital badges is Foursquare, which had themed badges for bars, restaurants, stores, etc. At one point they offered 3rd party health badges which were given for completing Runkeeper challenges or staying in a Health Month bracket. However, it was very easy for people to “earn” these badges without legitimately completing the health behaviors they supposedly commemorated. In 2012, Foursquare retired the health badges.

The quote I like to use about what happened to Foursquare’s health badges is from Scott Rigby. He writes, “One reported reason for ending these badges is precisely what would be predicted by self-determination theory: users began to ‘game’ the system, looking to circumvent the health behavior simply to get the badge.”

So what behaviors should badges reward? Despite my comment above, clicks, logins, checkins, and the like aren’t a terrible choice as the basis for some badges. These sorts of metrics serve as leading indicators, early signs about whether or not an intervention might eventually produce outcomes of value. Often, the most meaningful outcomes–like reduced cost, or better health–take a lot of time to achieve. They can’t be measured until months after someone starts changing their behavior; they are lagging indicators. But you can measure their participation, and if that’s poor, you know that the long-term outcomes probably won’t be there either.

This diagram shows one way to plot out the data you can collect at various post-intervention time points.
This diagram shows one way to plot out the data you can collect at various post-intervention time points.

An outcomes map like this is one way to organize the outcomes you’re trying to drive with an intervention. Because both leading and lagging indicators are critical steps in making your intervention successful, it makes sense that you’d want to reward both. And given the time-based nature of the outcomes map, it also makes sense you’d want to give more badges or rewards for leading indicators early in the process, and gradually transition to an emphasis on lagging indicators. However, the problem is that because clicks are so easy to measure, it’s common for people to keep focusing on clicks at the expense of other more meaningful data.

Unintended consequences If you’re not careful about defining the behavior you’re trying to change with badges, you may end up changing the wrong one. Foursquare’s fitness badges, for example, were not actually causing people to get more physical activity. Another example comes from when India was under British Colonial rule. The British government was concerned about the number of poisonous snakes in Delhi and wanted to enlist the citizens to help eradicate them. So, they offered rewards for the dead snakes that people could bring to them. What do you think happened?

People began to breed snakes so that they would have a bigger population to kill and get rewarded for. The rewards for dead snakes actually ended up making the problem worse. That’s come to be known as the cobra effect, and it’s something to pay attention to when adding rewards or incentives to the programs we design.

Positive examples? So is anyone doing a digital reward system well? I think so. An example I use a lot is the Starbucks app, which people can use to order and pay for their Starbucks items. Starbucks pretty clearly wants to encourage the behavior of spending money at Starbucks. So what does it do? It rewards people who spend money at Starbucks, with more products from Starbucks, which we can infer they enjoy because they earned them by buying Starbucks. The most avid users gain status (which is the analog for a badge in this example) and it gives them even more coffee power. It’s a lovely cycle where the rewards and the behaviors are well-aligned.

And it’s quite successful–not only does it have the highest number of monthly average users on Android and Apple of any food brand in the US, its biggest problem seems to be that stores can’t keep up with app-driven demand.


Rigby, C. S. (2015). “Gamification and motivation,” in S. P. Walz & S. Deterding (eds.), The Gameful World: Approaches, Issues, Applications. Cambridge, MA: MIT Press.