Positive feedback loops and numerical indicators of progress

There are a number of online systems that incorporate feedback systems to give you an unambiguous numerical indicator to measure reputation or value in a system that might otherwise be very tough to quantify. Amazon provides book rankings and review statistics, eBay has an elaborate reputation engine, and games like Second Life have in-game currencies that reward players for successful play.

When you are designing such systems, take note that there can be at least as much or more incentive for people to get status and approval in the system than any kind of monetary rewards. You don't need to provide explicit monetary feedback to generate interest, because the market will in some cases provide its own feedback to reward people who have behaved well in your game. Sellers with good reputations can command higher prices for their auction goods, as only one example.

Systems with a lot of opportunities for positive feedback are also just a lot more fun. Consider the varieties of pedometers that you can use to measure step counts. Some just tell you how many steps you have gone; others go so far as letting you compete in national walking leagues or let you tend to a virtual pet that likes to be exercised. BJ Fogg calls this Persuasive Technology and notes that computers can motivate you to do all sorts of things, for good or for ill.

UPDATE 5/06: Think "unambiguous numerical indicators of progress", and contemplate for a moment multiplayer games where different players are playing different games and thus there is ambiguity in the scoring systems.

Technorati Tags: ebay, feedback, flood, pedometer, reputation

Comments

[BJ Fogg]

The bigger picture here has to do with how a computer (a machine) can dish out positive reinforcement. Some people find rewards in numbers that represent social status or achievement (such as eBay's feedback system). Others may be reinforced by number of incoming links to a blog post.

I've outlined how a computing system could try different types of reinforcers in order to find what works for each individual. The next step is to schedule those reinforcements (like slot machines, but smarter) to make the behavior turn into an addiction.

I rarely share this research, since I think it's sort of -- well -- dangerous. Does that make me a bad academic?

Maybe.

BJ