Now that your reputation system is integrated with the rest of your site, you need to evaluate what the users are using it to accomplish. Hopefully, the value they are creating is in alignment with your goals. You won't know for sure until you measure the results. And after you understand how this organism is behaving, you may need to make adjustments to the reputation model, or even to your business plan.
This chapter will discuss strategies for evaluating the efficacy of your reputation system, detailing specific common metrics. It will describe when to start tracking and evaluating behavior, how to compare before-and-after figures, etc.) It will also recommend corrective actions you may be able to take for commonly encountered problems.
As we've noted at the end of Common Reputation Models, the incentive structures that reputation systems create and/or facilitate are often perceived differently that you intended by some users. There are often external incentives to abuse the reputation model to achieve an end of personal merit. For example, leader boards often lead to the creation of robots in an attempt by some to get the notoriety of being #1 on a large public site. Yelp.com banned the accounts of business owners who were writing positive reviews of other businesses using collusion techniques outlined in A Grammar for Reputation - direct profit is probably the strongest external motivator that drives undesired behavior. And that's before we even mention spam.
This sort of abuse mitigation is critical to the health of your reputation and proven techniques are detailed here.
|← Application Integration, Testing & Tuning (SUMMARY)||Case Studies (SUMMARY) →|