I recently came across a series of posts about competencies on Zapaterismo, a blog written by a gentleman with a good deal of experience in the competency management field. Basically, the gist of the series is that competencies currently are too much work, but that they can be made much easier with the application of the right approaches and technology.
In part 1 of the series he notes a myriad of problems with traditional processes — The tools are by and large poor so the mapping of competencies to jobs is tedious and error prone. He notes that rather than being a drag-and-drop exercise, this work tends to involve makeshift processes using Excel spreadsheets and Word docs. He also notes how taxonomies are brittle and destined to become quickly “outdated and nonsensical” because users don’t have flexible ways to interact with them.
According to Zapaterismo, the traditional process works something like this:
Ideally, several job incumbents and subject matter experts are interviewed using Job Analysis Questionnaires, and the interview results are analyzed to reveal key competencies for that position. This is done individually for every job, and ultimately natural naming conventions and taxonomies begin to emerge to support a larger competency architecture.
In part two of the series, he outlines what he calls the “Enhanced Top-Down Approach,” which, to nutshellize, seems to keep the analysis portion above in the hands of a central OD, Learning, or Talent Management department, but provides them with better tools:
The difference is that OD or whoever else is empowered by enhanced technology that allows for easier and more dynamic modeling. For example, org chart views of competency model assignments, the ability to cluster, drag, and drop competencies onto jobs, and other add- and edit-on-the-fly features made possible by web 2.0 technologies.
I completely agree with Zapaterismo’s vision. The mechanics of sharing, integrating, and updating content should be a lot easier than the current rather backward state-of-the-art. In my earlier post, “SOA Lightens Up in 2008″ I mentioned RESTful web services approaches and syndication as particularly applicable in competency integration. Since competency content integration is a specialized publishing use case, the RESTful approach likely has everything you need. You’d leave the development of taxonomies, competency models, position competency models as the value-added exercise it has always been. The key to making proprietary content easy to integrate might be exposing the pieces of a model or taxonomy that you want to share or allow to be updated as RESTful resources. For purely example sake, resources like:
http://example.com/taxonomy/{TaxonomyID}/node/{NodeID}/competencygroup/{CompetencyGroupID}
http://example.com/job/{JobID}/competencymodel/{CompetencyModelID}/competencygroup/{CompetencyGroup ID}
The above examples come out of my wild imagination. The developers of the competency content would work with the service designers in figuring out what resources need to be exposed to deliver competency content to applications and to allow the creation and update of competency models. This would be the technical “gorp” going on in the background. As Zapaterismo suggests, you might certainly have something like a rich internet application (RIA) or desktop application over top allowing drag-and-drop style user interactions and modeling.
Here’s the key — What would these RESTful resources generate/consume? While I don’t think it is useful for HR-XML or learning standards organizations to try to come up with all purpose meta-models for competency models and taxonomies, the HR-XML 3.0 library (draft here) is a great source for the competency bits-and-pieces (data types) those RESTful services might exchange. What kinds of bits and pieces? Things like PersonCompetency, PositionCompetency, PositionCompentencyGroup, PositionCompetencyModel, ReusableCompetencyDefinition, CompetencyEvidencePolicy, AssessedCompetency, CompetencyResult, CompetencyGroupResult, BehavioralIndicatorScoreDetails, etc. These are neutral, flexible, and extensible. They are well-thought out and support important options, without being gigantic (they are also drafts if you disagree and want to join us to influence their direction). In addition to HR-XML, I also mentioned in my earlier post how some competency content fits nicely within the Atom syndication format.
So in summary, I definitely share Zapaterismo’s vision. We have the technology. It is time to leave Excel-spreadsheets and integration via common-separated values (CSV) behind. While managing competencies will always be challenging, it can be a lot easier than the current woefully unsophisticated state of the art. Moreover, improving integration doesn’t rest (no pun intended) on everyone agreeing on a gigantic competency management framework. A lot can be accomplished with RESTful application development approaches and some standardized bits-and-pieces for moving data around. This isn’t over-the-rainbow stuff. It is here for customers who demand better and service providers willing to move their integration approaches forward.