Saturday, June 21, 2008

Pondering the wiki trough of death

Christopher Allen led an interesting discussion this past week at the Online Community Unconference on The Numbers of Trust (or something like that). He covered many of the topics, Dunbar's number, effective group size, research from WoW gilds, etc. that he used to discuss on his blog Life with Alacrity.

The discussion turned to the observed problem that wikis go through as they begin to scale. When they are just starting, and there are small numbers of people, wikis go through a rapid growth phase. The wiki-way of working is comfortable because the group is small and everyone knows and trusts each other. Ideas just explode. As more people come into the wiki, however, the level of trust begins to diminish and the growth begins to slow or stop. If the wiki doesn't completely fail at this point, if you can manage to keep it going long enough to make it through the "trough-of-death," then they bloom like nobody's business. Think Wikipedia, think Reed's Law.

So what happens to kill wikis as they transition through the trough-of-death? As the wiki grows and more people come, things start to get messy. This is where people want to try to "engineer" solutions. Often people want to control the new comers-- they don't trust them mostly because they don't know them. In our conversation at the Online Community conference we were calling this a "wiki-hole", but as I thought about it I think it is better described as a "wiki-trough". The trough is a dangerous place. Many (most?) enterprise wikis fail at this point.

These are some of the things that well meaning people try to do to get a handle on the messiness. They think they are doing things to help their wiki succeed when in reality they are accelerating its demise. These are all bad:

  • introduce a hierarchy-- bring over the old organizational structures
  • credentialism-- declare that only people with a certain expertise can contribute
  • authority-- put people "in charge" of sanctioning things as worthy
  • access control-- create permissions to view, edit, work flows...

All of these things work to diminish trust and enforce the old organizational ways of working. It's that definition of insanity at play, doing the same thing over and over and expecting different results... These old things conflict with the very values "baked-in" to the wiki: transparency, trust, flatness, inclusiveness... When you mess with these things you'll never make it out of the "trough-of-death." These things send people away at the very time you're needing many more people to join your effort. This is why I don't buy into the concept of an "enterprise class" wiki-- different values and different "features" for organizational use versus what people might use in a public setting. People are people whether they're working in an organization or doing it as part of a public project. People's motivations, desires, wishes, dreams don't change when they go to their workplace. We are not schizophrenic. There is not a work psychology and then something else for the rest of our life. Our work and personal lives are not separate. Motivation is motivation.

So the bottom-line... if you don't screw-it-up and try to over-engineer your wiki upon entering the "trough-of-death," you just might come out the other side and create some real organizational magic. You just have to have the guts to leave it alone when everyone starts clamoring for more control. If you try to fix it you will break it for certain.


jed said...

This is a problem I'm very interested in. Can you provide some details? For example, how many participants do wikis have when they enter the trough? Does the size (number of pages) affect when they enter the trough? If so, how many pages do they typically have when they get in trouble? Does the nature of the content matter? What are the typical problems that are most troublesome -- edit wars, irrelevant material, bad organization, ???

Anonymous said...

This is an incredibly insightful post and could really apply to any online social network. I've seen it happen on mailing lists often enough. People want to move to something bigger more spontaneous, but they can't relinquish control.

I think it's why you see a lot of these efforts flourish outside of organizations. The control imperative is easier to resist in that setting, and the culture really only exists around the web artifact.

Kevin Gamble said...


Great questions. I think those could be made into more full posts. I'll start working on them.

I do know that the content is rarely the problem. Bad organization on the other hand... This is directly related to over-engineering. There are as many different ways to organize as there are people. This can rarely ever be done in a way that will succeed. It needs to be as flat as possible.

More later.


Kevin Gamble said...

Thank you Bud! Your comments really mean a lot to me as well as being right on the money.


ChristopherA said...

I've begun to detail more on this topic in a series of blog posts on group size issues starting at