Signal to Noise / Dunning-Kruger / Lifecycle of a forum

[John Horton]

When I started my forum 20 years ago it was not a business. It was a community I wanted. It was a place for a group of friends who were very committed to a sport to talk about technique and gear. There was some goofy banter, but a large portion of the conversation was focused. Many of us were nationally ranked-athletes. We all knew the skill level and experience of each other. The most vocal of us were legitimate experts in our sport. New members came looking for knowledge. The quality of the discussions was the draw.

As the user base grew the average user had less and less knowledge. By the time the forum reached peak traffic about 6 years ago, the most knowledgeable users found themselves overwhelmed with posts from overconfident knuckleheads. In the frame of Dunning-Kruger the user base was moving toward the left on the chart.

For the most part, I have thought of this problem under the paradigm of “Signal to Noise”. Fresh content is the lifeblood of forums but if the quality of the content is low it has to eventually burn itself out. You can’t just ban enthusiastic new members who constantly post low-quality content.

I have been attempting to fight this for years and it has been a slow losing battle. One idea I had recently is what if reputation was a metric calculated as  “Likes” divided by “Posts”. Gamification has a limited impact but at this point, it is the only idea I have. 

 

[JoelR]

I always get a chuckle from the choice descriptions in the Dunning-Kruger graph ... Valley of Despair ...  😁

Some thoughts:

1. Expert Content - Our communities, as a whole, need to do a better job of identifying and surfacing the very best content for new members. It's one of those situations where 1 superpost can probably answer 50% of all newbie questions.  We really need to do a better job of creating something like "Best of 2024" or "New User Guide to XYZ" that brings together the best information from our communities in an easy manner for new users.  We probably need to link those "best of" definitive guides directly in our menus or on the homepage.  This is one thing that I plan on evaluating in IPS 5, when we have some more interesting layouts for Sidebar Menus.  
2. Fresh Content - IPS is definitely ahead of its competition, all of whom show chronological feeds - they're not interesting, they're not helpful, they're just timely.  I would encourage all IPS admins to use a mix of automated feeds (eg. widgets like Trending Content or Popular Content) and manual promotion (eg. Our Picks promotion). It's time-intensive, but letting staff pick out expert and great nuggets for Our Picks promotion rather than letting them get buried in the community is important.    
3. Signal to Noise - You're definitely not the only one who has had the same idea. Multiple people have toyed with the idea of Likes / Posts.  The problem with Likes, however, is that it can incorporate all likes, which can include humor, funny, poo, spicy, interesting, etc.  My personal take is that IPS is moving in the right direction by moving away from reactions (which can encompass a broad and offtopic assortment of feelings) to Helpful, to truly identify helpful content.  You use Helpful in areas of your community of hard content; you use Reactions in areas of your community of social content.  

Would love to hear what others think!