Re: Enterprise search - defining standard measures and a universal KPI #metrics #search

Matt Moore


I had a very similar conversation when I was at Panviva. This was a tool that documented workflows (esp. in contact centres).

Some of my colleagues said: “All organizational knowledge should be documented!”

My response was: “That sounds very, very expensive”

The amount of “controlled” knowledge you want to have will depend on the size, maturity and risk appetite of your organization. But generally it is between 0% and 100%.

This also links back to frameworks like Cynefin and the levels of knowledge control working in each domain requires.

Really, these aren’t questions about search but they are questions that search highlights. Enterprise search is a bit like turning on the light in a stoned teenagers bedroom. You may decide to clean things up. Or you may decide just to turn off the light.


Matt Moore
+61 423 784 504

On Mar 3, 2021, at 11:42 AM, Stephen Bounds <km@...> wrote:

Hi Murray,

I think you've put your finger on an irreconcilable problem with Enterprise Search. In an organisation, we can divide our knowledge artefacts into three basic buckets:

  1. Corporate knowledge - an explicit representation of organisational knowledge about processes, practices, values and/or decision making priorities which are meant to be acted upon by employees regardless of their personal feelings on the topic in question

  2. Pooled knowledge - a composite of information artefacts representing knowledge from a range of people and teams with some form of quality assurance or validation applied. Unlike corporate knowledge, pooled knowledge isn't definitively right or wrong but is shared to improve productivity and quality of outcomes. It is expected that people will use their own judgement on how to apply it correctly.

  3. Situated knowledge - information artefacts tied to knowledge primarily relevant to a single individual or team. Attempts to interpret this knowledge from outside the team without active mentoring or training is likely to fail due to missing context or capability factors.

Since "enterprise" search by its very definition assumes a single interface to access all three kinds of content, it's no wonder that users get confused about the purpose of results.

More to your point, I agree that it makes sense to have a class of corporate knowledge that is unambiguously definitive, because you want people to act as part of the whole, not individually. But removing all autonomy in knowledge search is a very "command and control" mentality and pretty brave in a world where localised knowledge assessment and decision-making is pretty-well agreed to be a more resilient option.

Stephen Bounds
Executive, Information Management
E: stephen.bounds@...
M: 0401 829 096

Join to automatically receive all group messages.