Often when we think about evidence-based management we start with very simplistic models.
For example (and a giant tip of the hat to Gerry Crispin & Mark Mehler on this) with Source of Hire it is easy to suppose we have 3 sources of hires: job fairs, corporate career page, job boards. Rather than just invest blindly we gather evidence: Number of hires/source of hire. Then we can easily pick the Source that generates the most hires / dollar invested.
The trouble with this example is that the model assumes that candidates encounter one source of hire and then do or do not apply. In fact, candidate behaviour is complex and perhaps they first become aware of the company at a job fair; check it out on line, and then apply via the career page or a job board. A simplistic "evidence based" view that job fairs don't bring in candidates would be (in this made-up scenario) completely false and lead to a very bad decision to stop investing in fairs.
The takeaway is that we need to make explicit our model (i.e. how candidates interact with the source of hire) and test it before we start gathering metrics. It can be very tempting to skip that step and "get on with the work".
The second takeaway is that we need to be able to describe the more complex reality in a way that guides us in gathering evidence that we can then use for decision making. How do you do this? That's a question I'll put to the EBmgt brain trust.
One free tool that is a bit complex but might be helpful can be found at Insightmaker.com
Here is a model someone has made about recruitment & retention of optometry students.
http://insightmaker.com/insight/1761
I wonder if this kind of modeling is helpful or whether it just appeals to our inner-postgrad and we in fact do better to stick with the backs of envelopes.
