IBM Study Reveals Effect of Gamification Withdrawal

IBM Study Reveals Effect of Gamification Withdrawal


We’ve all read something along these lines before: “Company X announces new internal social network with gamification features”.  With big names such as Samsung and Salesforce using this technology, there is clearly value to be had in having these gamified social networks. Inclusion of badges, points, and leaderboards all lead to increased user engagement but what happens when you take away these features from the network?

IBM Researchers Jennifer Thom, David R. Millen, and Joan DiMicco conducted an experiment in which they attempt to answer: “How does the removal of gamification features affect user activity within an enterprise social networking service?” While the introduction of the incentive system “dramatically increased the overall levels of content”, the paper’s findings suggest that users who are engaged with gamification in these networks had more activity than those without it and that the removal of these same features resulted in about 50% less activity. However it should be noted that the top users often wrote brief and salutatory posts to the extent of “hi” and removal of the point system decreased posts written in this manner. In a competitive and gamified environment, does increased activity through gamification actually increase meaningful engagement? Or does it merely just provide a reason for users to post terse comments?

IBM is careful to mention that the study is done within the context of their own work environment and that the effects of gamification can vary among cultures. For example, IBM found that users in India were far more likely to engage with users outside of India compared to American users who mostly stayed within their country.  Whether this was a result of the Indian users being based in the service sector of the company or a result of geographical difference, it shows a difference in the way the IBM social network is being utilized by different employees. Companies considering the integration of gamification into their social networks should understand if the introduction of game mechanics will be appropriate for the work environment it is going to be a part of.

You can read the full paper here



Need help with behavioral science and gamification? Get in touch with our boutique consulting agency Dopamine.


  1. This really highlights the problem with the extrinsic reward systems in gamification, if they aren’t well thought out – if you reward people for posting absolutely anything, that’s exactly what they’re going to do. When you remove that reward, they aren’t going to do that as much anymore. The trick is rewarding _precisely_ the behaviors you want to encourage – nothing more, nothing less – and carefully considering what unintended consequences might result from those rewards.

    • Richard,

      An interesting observation. How would you suggest to encourage “meaningful” posts as apposed to terse comments? Would it be helpful to require a certain number of characters in a post or give more points for more characters?

      My company is working with Zynga and I’m looking to come up with some potential solutions before the problem exists within our organization.


      • Well that is really the trick, isn’t it? If you want to encourage insightful comments, that’s exactly what you should reward. Ideally, we’d use something like AI to determine which comments were insightful and which were not.

        Given current technology, this is more difficult. One approach might be to use votes from other users to determine quality – this type of gamification has been in use in many websites. The problem is that in doing so, you give users a clear path from low points to high points, so anyone that wants to “game” the system still can, e.g. by creating fake accounts or getting friends to click up their posts without merit, etc. Any time you formalize the rules, you create a way for users to “win.”

        Frankly, this is one of the reasons I think Foursquare is successful – there are a number of badges where the precise earning conditions are never clear to the users. It is up to them to explore, try new things out, etc until they earn those badges. But of course… that means the system is rewarding exploration. If that wasn’t your goal, you’d need to think of something else!

        In websites where the target is increased participation, these kinds of general gamification systems are very effective because the organizational goal is matched to the goal structure – the site wants more activity (posts) in order to appear active and boost their SEO. The site rewards more activity. So it is a win-win. But if your real goal, as might be the case with a private social network site full of employees, is to encourage deep conversations, increased collaboration, etc – then that’s what you need to reward.

        • Yes, that is mainly related to the stuff of unexpected vs. expected rewards (e.g., a badge). While an expected reward is of extrinsic nature a unexpected reward (where it is not clear where it exactly comes from) is of intrinsic nature.
          You can read about that in detail in Deci et al. 1999 (which is by the way citet by the IBM guys (citation 2)).

          You can easily observe that in real life when your expect an incentive due to good job performance which is cool but not sustainable or if you boss come to you unexpectly and tell you that you did a great job. However, a gamified system should have both, expected and unexpected rewards. Nonetheless, the intention of the expected reward should have meaning in order to stimulate mastery, purpose, belonging or whatever job oder personal related resources.

          In our studies with the Technology Acceptance Model we found exactly the same effects in strucutual equation modeling.

  2. I agree with Richard above. My biggest gripe with gamification is that there seems to be no connection between the extrinsic and intrinsic take aways one gets from just gaming a system. A reward to has to mean something besides just getting a badge especially if it is about behavior modification which is at the core of gamification. If I were to create a gamified activity in the organization I would look for the end goal — which should be more than just increasing productivity (which is what you get out of it as the business leader) — that would benefit both me and the player. Think of Clay Shirky’s The Promise, The Tools and The Bargain.

  3. I am assuming John is making a joke here, by asking if we can tie “meaningfulness” to “number of characters”. Perhaps we should question whether “meaningfulness” and “incentivized” might be mutually exclusive. If we NEED to incentivize it, doesn’t that — by definition — make it NOT meaningful? That does not mean we cannot incentivize other things including quantity and other forms of increased activity. But really, if it truly meaningful then it carries its own intrinsically-motivating “reward”. If it is NOT meaningful, no incentive is going to change that. On the contrary, an incentive HAS been proven to change what IS intrinsically meaningful to something less so.

    But I am certainly LOLing at the “number of characters == meaningful” idea.

    • “If we NEED to incentivize it, doesn’t that — by definition — make it NOT meaningful?”

      Definitely not! Consider the number of people that join gamificiation-based weight loss websites to motivate them. The outcome of that process is already meaningful; the process itself is just unpleasant. They just need more motivation to convince themselves that the outcome is “worth it”. Most of our study of intrinsic/extrinsic motivation focuses on single time points; the long-term effects of (and interplay between) intrinsic and extrinsic motivators are less clear.

      I would agree, however, that there are limits to this.

    • The benefit of point systems is that they break down a long process of action-longpause-reward cycles into shorter cycles of action=reward. They help the “player” see the incremental progress and value every little step along the way.