In this blog, I’m going to highlight my support for major Enterprises using internal wikis for internal knowledge management and I will also call out some things that I think need to be thought hard about to make them what they deserve to be –Wikipedia quality content for internal information.
As a quick aside, the concept of an internal wiki and hence not sharing things in the public domain “open-source style” might sound a bit evil. It’s not, it is absolutely necessary for companies to need to keep certain information private, it just so happens that I believe wiks (more commonly associated with public sharing) have another use case.
Pits not to fall into (sprawl and deception)
In my view you need to consider a few things to maximise the benefit of a global Enterprise wiki.
On a large project that I personally spent some time on and that lasted quite a few years, the project hosted a wiki (moinmoin as it happens for the wikistorians out there). As you’d expect in anything that lasted for so long variety of people interacting with it, we learnt a lot about what did and didn’t work.
The main problem was what I call wiki sprawl. Perhaps I can misquote Pareto (aka 80:20 principle) here and state that over time, “80% of wiki pages end up containing inaccurate, part duplicated inconsistent information, whilst 20% of the pages contain 80% of the good quality information.”
Granularity of pages was a contributor to this. Wiki pages can be similar to working with aggregate-oriented design in the NoSQL world. What is the most common unit of information that will be accessed in one go? What should be normalised through different “aggregate types” (interlinked pages) and what should be denormalised (lumped onto one page). Without getting this right it can be very difficult to create consistent information and even more difficult find it.
Another problem was things like formatting of pages. You could take the view I might be being a bit anally retentive on this point, but genuinely when pages were well laid out (good summary, appropriate use of sections and subsections, highlighted code blocks etc.) they were a lot more usable and it was also a lot more tempting to contribute to them for the benefit of others.
Sprawl started to take over. Anyone who created a page called for example UsefulLinuxCommands without first searching and discovering the existence of a similar page HelpfulLinuxCommands was definitely causing proliferation when they proudly part-duplicated, perhaps even part-contradicted the advice from the pre-existing page.
Over time the problem compounded itself. When someone comes along wanting to create a wiki page for the “greater good” diligently searches for any pre-existing pages and finds a list of 10 possible pages to trawl though before they can be sure their information is not already captured, I can sympathise with them creating a new page. Especially using their fancy flavour-of-the-month page formatting style.
It’s not all doom and gloom. I can hardly suggest this is the fate of all wikis. We’re all aware of what I am assuming is the World’s largest and greatest wiki – Wikipedia
How to avoid pitfalls (not so moderate moderation)
As we all know, the success of Wikipedia cannot be attributed to the technology. Whilst the underlying software application MediaWiki is certainly more powerful than moinmoin (RIP), it is predominantly the community of people that have turned it into possibly the best single source of information on the planet (excluding indexers to other information like Google).
So how did Wikipedia get so good?
Most people know that it is supported by thousands of moderators protecting content accuracy, legality, conformance to the site policies, and of course fighting duplication, inconsistency and sprawl. It’s also widely recognised that the moderators (Wikipedia seems to call everyone Editors) are volunteers. So how is this distributed, almost anonymous, self-organising team successful?
From some rather limited research, I’ve observed some contributing factors. But the truth is, having only updated a few pages myself, I don’t fully understand the formula.
- Firstly anyone can edit a page, even anonymously. Therefore the barrier to contribution is extremely low. Obviously however, this also has an implication on information quality.
- Traceability of change. The site does its best to make it easy to see exactly what has changed, when, by who. This is through a combination of Watchlists related change pages, the Talk communication channel etc.
- Editor rank. Editors start off with a low rank which limits what they do in terms of revoking and approving changes. Through a meritocracy based on quality and quantity of contributions, people graduate from Autoconfirmed, to Adminstrators, Beurocrats, Abtration Committe, and finally Stewards. (The next position “Founder” is reserved!).
- Some pages become protected. Which is a process where higher ranking editors (perhaps with local status achieved due to their contributions to a particular page) can control and approve edits to protect vandalism.
- Finally, one of my favourites, Wikipedia is policed by numerous tools or “bots”. These all extend the capacity of human editors to moderate content.
So what does Wikipedia teach Enterprise about internal wikis?
I think the following very positive points:
- it is possible to do this at global scale
- it is possible to do this with volunteering (or in our case people going beyond their day jobs)
- this mechanism for structuring information to be stored in one place is so far the most powerful solution on the planet.
But don’t mustn’t ignore the critical success factors:
- Updating information has to be a pleasurable experience (good bye SharePoint)
- Review and governance processes need to be extremely well thought out and able to evolve over time
- Some tooling to help track and moderate change
- Perhaps a sort of “quality” critical mass beyond which
I’d be interested to hear if anyone has read this book about Wikipedia and/or has experiences good and bad to share.