Acmepedia: A Case Study in Wikis, Facets and KM

Enterprise wikis are still squarely in the experimental phase of being accepted into KM practices, but there’s already enough success to say they’ll be here for the long run. We also already know they introduce substantial new challenges around disorganization, security, authority and trust. And we need to be sure they don’t become yet another information silo. Facets and tags provide elegant solutions to those challenges, as we’ll report from our own internal case study.

Wikipedia Worked. What about Acmepedia?
The Wikipedia is one of the great successes of the past five years of the Internet. It reminds us that the Internet connects people, not just machines, and that we have only begun to envision how those new connections will change our world. The millions of volunteer editors of the Wikipedia have already produced the equivalent of a 750-volume encyclopedia, with an accuracy measured to be similar to that of traditionally edited texts. And though it has real problems with vandalism and inaccuracies, its benefits make it indispensable.

Given this success, any enterprise has to wonder how a wiki encyclopedia of its own business content—call it the Acmepedia—might change the practice of KM. The comparisons of Wikipedia to Encyclopedia Britannica, and Acmepedia to enterprise knowledge management, are many. After all, KM follows an editorial process that ensures authority and expertise. The outcome is gold-standard information produced at great expense, along with huge gaps in coverage, articles that grow stale and an inability to capture the informality of real-life knowledge transfer.

Do the lessons of Wikipedia apply in the enterprise? Since the field is new, we could speculate, but instead, I can share the results of our own internal experiment, and some of our early adopters. It can work, fantastically well, provided you solve some material challenges. The key differences, each of which I’ll unpack in turn, are:

1. Wikis are an addition to KM, not a rip-and-replace. They will become yet another silo if they’re not designed to complement your existing enterprise packaged applications.

2. Faceted navigation and information access can be the key to crossing different content silos.

3. Although facets are part of the solution, they also introduce new requirements: how do you categorize everything so it can be found again through faceted navigation? With enhancements, tagging and folksonomy provide an answer.

4. Authority and trust impose different constraints in the enterprise than in the public Wikipedia. Facets offer dramatic changes here too.

Another Silo?
Our first experiment with wikis came organically: they cropped up in each department, alongside their cousins, blogs. Technically, our IT department could already offer our workers most of their functionality via commercial products for CMS, collaboration and an intranet. But a small change in ease-of-use can make a world of difference, and wikis and blogs flourished in a way our old collaboration software never had.
We quickly ran into a real problem. New silos of content were forming, content was being unknowingly duplicated, and it was impossible to find anything across silos. So next, IT standardized on a single wiki platform for everyone, and we "ate our own dog food" by extending our own Endeca information access platform to unite those silos (more on the latter below).

That simple IT solution exposed a big business process question: which content should migrate to wikis, which should stay in the commercial collaboration product and which should be left in packaged applications? For example, does an informal note about a transaction belong in the CRM system or a wiki about that customer? Our pragmatic solution was to embrace the silos, leaving three complementary "buckets" of content:

• "Curated" content: Only a utopian would expect white papers, legal documents and HR forms to be produced by the proverbial million monkeys banging on a wiki. These would continue to live in a CMS, complete with version control, workflows and the like.
• Packaged applications: Data and notes from CRM, HRM and ERP packaged applications were some of our most valuable content. We would sensibly leave it where it was, but use information access to integrate it.
• Wikis and blogs: These group collaborations proved best at capturing the conversational nature of emerging topics, discussions, threads, opinions, ephemera and niches. And they required some business process changes to align them with communities of practice.

The Wikipedia itself is not a silo. It’s always used as a complement to the wider Internet, your own desktop and the physical world, with links and references out and in. Acmepedia should follow that same pragmatic approach, and serve alongside tried and true processes.

Facets and Search Unite Silos
Information access platforms have proven their ability to break down content silos. For too long, data and content have been captive to the application where it was created or stored. But workers just want answers. Just as virtualization has abstracted storage and processing from specific physical systems, the solution is in information virtualization. This takes chaotic content from any source, delivers it to the worker at the moment he needs it to support a decision and adds value by organizing it. This includes charts, summaries, visualizations and navigation so he has the context to make sense of it all.

Faceted search and navigation is one of the key innovations that puts a flood of information into context. The result? When I search or navigate across, say, the name of a financial services customer, I might recall thousands of results, but summarized so I can filter by facets like regions, dates, departments, content types, source systems and beyond. These systems are now one of the fastest growing software categories, which is evidence of their utility. But equally telling is our internal result: adoption of our Acmepedia, "Endecapedia," is extremely high and climbing. Put simply, it works.

Wikis:Content:: Folksonomy:Categorization
Disparate silos will never "automagically" organize themselves. But just as the Wikipedia upended our old assumptions about how easy group content creation can be, we also need to change our assumptions about how to organize wiki content so it’s amenable to faceted search and navigation.