In my last post I discussed problems with the shared drive and SharePoint as knowledge management solutions. With these systems adoption is high (everybody is in the pool) because they are simple, but due to lack of an information management strategy the content is often a mess.
In this post, I’ll discuss strategies for building successful KM systems that achieve high adoption while simultaneously providing access to organized content. In other words, throwing a KM Pool Party that isn’t a hot mess.
As we’ve seen, high adoption requires keeping it simple, resulting in a frictionless system, and what I call a KM pool party – everyone is in the pool.
The hard part – keeping the system frictionless without lawlessness.
Where many KM implementations fall apart is in the methods chosen to organize content. To keep things organized you must impose structure, and you need rules for adding or maintaining content.
In my experience, there are three ways managers of KM systems try to impose structure and rules:
- Forcing participants to play by the rules. In this approach users are trained (or really constrained) to use the system in proscribed ways. This can work when users have a massive and immediate gain by using the system. But most KM systems offer gains that are diffuse, varied and happen over time. So the play by the rules approach generally leads to low adoption rates and system failure, with the exasperated information professional screaming “why can’t they see the benefits?”
- Imposing organization and order after the fact. In this automated approach, order has two components: a harvesting component and a processing component. Content creation and maintenance is left unchanged, but the gathering component crawls the various repositories (the ECM, the shared drives, intranets, etc.), and aggregates data into a single centralized index. Value added processing (such as auto-classification) occurs as content is harvested and aggregated, resulting in two key benefits: First, users no longer have to search through myriad systems to find what they seek – the system enables discovery from one centralized venue. The second benefit is that value-added processes enhance findability and improve navigation, thereby allowing users to easily discover relevant content.
The upside is also the problem.
When we don’t ask users to play by the rules, content creators are unaffected, and we are able to build the solution with minimal disruption. It’s almost like we brought the pool to the party! Everybody is automatically in, but because everything is in, the signal to noise ratio is high, which is a big problem.
What’s the worst that could happen?
Imagine you need to send a briefing document to your team. You want to send them the original product launch plan for “Autonomous Drones,” which you know is in the system. But the system also captured both the re-drafted launch plan, which was dumbed down by Sales, as well as the cleaned-up version submitted to the board (scrubbed for regulatory review). Which one should you circulate to your group? What if the Sales draft comes up first, and because you don’t know about the other copies, you send this incomplete version to your team?!
- Information professionals organize, curate and publish the content. This approach offers many benefits. First, we don’t have to force change upon participants (e.g. asking them to play by the rules and/or submit content via a complex process) since information experts are going to clean up the content before publication.
Second, we avoid problems related to harvesting everything from everywhere, because humans know how to de-dupe and so forth. We end up with high quality content and a low signal to noise ratio. Third, during content curation, trained information professionals can classify, use tags, and add value to the collection, making search and discovery fast and efficient. Hey – this sounds fantastic!
Failure to scale
The problem with this approach is that it simply does not scale for very large data sets and/or highly dynamic data. It just too expensive to make a person do this for all the content.
The right solution…
For many organizations the answer lies in a combination of these three approaches while adjusting the mix depending upon the context and goals. Try these simple techniques:
- When you can, and for formal documents (e.g. the final version of client deliverables in a consulting firm), get your users to play by the rules, and create a formal submission and meta-tagging process. Remember – garbage in means garbage out, so identify the most critical documents and create formal ways to capture that content.
- Where big data sets and dynamic content rule, turn to automation to capture and add value to your content. Remember, when trying to drink from a fire hose, you need automation to help turn the blast into a water fountain.
- Use information professionals where they can provide the most impact. What is the highest value content we have, and what value can humans add that machines cannot?
If you get the mix right, you really can have that elusive and nearly mythical KM Pool Party where the content is well behaved and users find what they seek.
Are your KM initiatives ending up like a pool party or a hot mess? Let me know.
Knowledge management best practices include methodologies collected by KM expert Stan Garfield in his book Proven Practices for Promoting a KM Program.
Knowledge management best practices include ongoing training strategies that leverage a variety of channels e.g., webinars and self-paced courses.
Selling KM is a journey; getting buy-in is an ongoing knowledge management best practice permanent process requiring a continuous improvement strategy.
Knowledge management best practices include making content easy to find through multiple methods and channels, curating collections of useful content.