A Neighborhood Watch concept...

[Eric Snyder]

In the Ottawa, Canada area, we have a mature online "gifting" population of probably 50,000 spread across a large number of different groups (eg. Freecycle, FullCircles, ReUseIt, Freestore, etc.). The focus is local / community, rather than global / subject.

And, of course, in such a large population and in these kinds of economic times, there will be those who see opportunities to take advantage of people's generosity. We have a handful of frequent offenders who, over the past 5+ years, have done some nasty things; been discovered; banned; re-emerged with a whole new identity; and done it all over again.

This morning, one of our long-in-the-tooth members spotted some suspicious behavior and IM'd me about it. Using Grouply (grouply.com), I was able to quickly do a cross-group analysis, and determined that yes, a person with the same email had posted 27 "stuff wanted" postings across 5 or 6 different groups, giving 4 or 5 different locations.

By their very nature, moderators are focused on ONE group and probably wouldn't notice this suspicious activity. But, looking for cross participation, one quickly comes to the conclusion that something doesn't smell right.

So, I was thinking, we could set up some kind of Community Watch mechanism... similar to the Neighborhood Watch idea. We could provide an email address for suspicious activity reports. An individual could quickly do a cross group analysis and send out an early warning "scamalert" message to all the 40 or 50 mods to put them on a higher level of vigilance.

It could easily be done by a volunteer in a couple of hours a week, I think.

Has anyone done anything like this? Any thoughts? Suggestions?

Eric Snyder
Ottawa.Fullcircles.org
=================

[Dave Govan]

Sorry for the delayed reply....Thanks Susan. Clients can implement it in a matter of a couple of days and Moderators could see a User's ReputationScore or an abstraction of it displayed as a symbol such as a green, yellow, or red traffic light, right next to the User's posted content. As you can imagine this can be very useful for pre moderation assignments. I still read / skim all of the postings but I know which ones not to speed read. :) Also, sites can reward their best users by letting Moderators point up positive behavior such as a User helping another User or self moderating, etc.

[Antoine GELLÉ]

Hello Eric,

When I read your post, I immediatly thought about one of the features of IMSafer, a computer-based moderation software which allows parents to receive alerts about potentially dangerous online chats their kids may have. This feature allows parents to report users with inappropriate attitudes in order to warn other parents should their kid start chatting with one of the flagged users.

Even recently, the Kenny Glenn story shows how quickly Internet users can get together to stop inappropriate behaviours, even though the consequences can be a little bit akward.

Being in France, we would never have any kind of campaign of Name and Shame against an individual, especially against a teenager. But this is just one of the many differences between "old" Europe and Anglo-Saxon countries, we both have a lot to learn from the other.

I look forward to learning more about ReputationShare though, it sounds like a promising idea !

[Dave Govan]

Antoine,

I would like to clear up one point in your comments; because naming and shaming is specifically not a feature in ReputationShare. The service is a central depository where participating services report incidents of positive – or negative – behavior. As such, the service provides aggregated behavior information, not judgments. Each site interprets the information according to their own terms of service, and adapts their responses accordingly.

Using this information, sites have several opportunities, but I’d like to highlight just two aspects because one shows how sites may use the information behind the scenes, and the second shows how sites may surface reputational information to other users.

1. Learn early in the registration process whether a would-be user has a sterling reputation or has displayed negative behavior in one or more categories. This information enables sites to make informed decisions:
a. If the user has a sterling reputation, the site may want to really welcome this user as they are likely to be a positive contributor, and bring similar people onboard, etc.
i. The site may also determine that the likelihood of this user being disruptive in any fashion is very low, and therefore need little, or no attention by moderators – which makes moderation more effective and less expensive.
b. If the user has one or more negative incidents reported against them, the site can:
i. Understand if the incidents occurred in areas of concern - If the incident is bullying, but the site is an ecommerce site with no user interaction, they may not care. However, if the incident is charge-back fraud they may care a great deal.
ii. Consider the date of the last incident - If the last incident is old, sites may treat it very differently than if the incident happened 3 times in the last 2 days.
1. NOTE: ReputationShare has a built in aging mechanism, which is critical as it enables users, who have made poor choices and then improve their behavior, to repair their reputations. The impact of negative incidents on a user’s score is reduced over time if no new negative incidents are reported, or if positive incidents are reported against an active account. This is particularly important for youth who may act out at times.
2. Determine whether ReputationShare runs only in the background for their moderators, or if they want to surface any reputational information about users to other users. Depending on the type of site, and the site’s values, terms and conditions they may choose to surface some information. For example a site may:
a. Choose to let users establish parameters for whose with whom they choose to interact.
i. A dating site may allow users to customize their ‘partner searches’ to not only include age, location and interests, they may also include search criteria like I am not interested in dating anyone who has a reputation score under (xxx where the user sets the bar), or who has been reported for Cyberstalking, etc.
1. Adding this selection criteria in no way names or shames, potential ‘dates’ who have these attributes are simply filtered out.
ii. A family safety /parental control service may allow parents to help their children transition from a highly protected level of interaction with only friends, to a broader interaction with online users by providing interaction criteria like “my child can interact with anyone whose reputation is over xxx, has no incidents of cyberbullying, sexual grooming…
1. Again, adding this selection criteria in no way names or shames, but helps families establish parameters to help increase the likelihood of positive interactions.
b. Choose to let other users see in real-time the reputation score (not what incidents generated the score) of other users.
i. An MMOG could choose to surface the reputation score of other players so gamers could in real-time make decisions about their interactions with other avatars. What this would show is that avatar ‘x’ has an overall reputation of ‘y’.

I hope this clarification helps put your concern to rest, please feel free to contact me directly (here’s how) if you have additional questions.

Best Regards,

Dave