[IPAC-List] HR metrics

[Mark Hammer]

Delivering  HR metrics via survey data is pretty much what I do for a living.
I have been involved in several pan-governmental exercises to devise such
metrics.

The essential nature of the task is that of striking a balance between what 
is easily (and consistently, over time and over organizational units) 
measurable, and what gets at the heart of the matter.  The one is not 
necessarily the other.  On top of this, one strives to include both leading 
and lag indicators.  If you have plans for keeping your HR metrics system 
in place for a while, the lag indicators will assist in validating your leading 
ones over time.

Certainly what has struck our team about the general task is the extent of
measurement inequivalence that occurs.  For example, one may wish to 
use reliance on contingent workers as a barometer of extent of integration 
of HR planning and business planning, but some places necessarily have 
seasonal fluctuations in staffing needs, so contingent is a big part of how
they roll.  All the more reason to gather lots of contextual information with
your basic HR metrics, if only to make sense of things you end up seeing 
that don't make sense.

I think, as well, that it can often be the case that not enough thought is 
put into how easy it is to map a construct.  For instance, I have to deliver 
up data that broach on the fairness of hiring, and on the transparency of 
hiring.  I would suggest that "transparency" will likely require MORE
indicators/metrics than fairness, simply because it is a more diffuse 
construct.  So when you sit down with the decision-makers to do this, do 
not let the buzzwords drive what you do too much, or think that because 
you have a couple of indicators for this area and a couple for that one that
you'll be alright.  One really needs to think about how fuzzy the target is, 
and titrate the metrics in accordance with that.  Some things ARE easy to 
measure in 3 or 4 indicators, but not everything is.

Our own federal approach involves the integration of several sources/forms
of information.  One source is simply administrative data that you can pull 
from records.  Another source is self-report by the individual departments
and agencies themselves (essentially what they put together as "evidence" 
of how they conform to a given indicator/metric).  Another source is data
from other agencies like staff relations tribunals (i.e., how many 
grievances/appeals filed, supported).  The final source, that I'm on the 
hook for, is survey-based data.  Here we use a multi-stakeholder 
perspective, using indicators from hiring managers and from candidates.
We were considering including people in HR as well, but the numbers 
were going to be small and so inconsistent from reporting period to
reporting period that we wouldn't be able to derive anything at the 
departmental level.  Still, if one is interested in simply getting a handle on
process, overall, they're a good source of info.

As you might imagine, time-to-staff is a major source of interest in
management circles.  Unfortunately, because there can be so much 
variation in how people staff, or the order in which steps are done, it is
next to impossible to look at it through a process-improvement lens and
determine how much time this step or that step contributes. All we can 
really do is designate a starting point and an ending point, and ask about 
what slowed things down.  Unfortunately, the extent to which managers 
convey disappointment with how long a particular step in the process took
may not necessarily reflect how much it actually contributes to staffing
time, merely how much their expectations were defeated.

We can discuss more off-line if this is of interest or utility.

Mark Hammer
Ottawa



>>> "Pluta, Paul" <ppluta at hr.lacounty.gov> 2009/09/30 10:26 am >>>

We are currently working to implement department-wide metrics in order
to understand and get better control of our processes.  We want to make
sure we are measuring the right things in the right ways.  Has anyone on
the list worked on such a project and, if so, what metrics did you find
most informative in helping to understand and 
control your processes?