The Christmas break offers the opportunity to read children’s stories to grandchildren and this year I ploughed into Thomas the Tank Engine and got reacquainted with the Troublesome Trucks! If you aren’t familiar with the Trucks, let me just say that they are always causing trouble, they like to play tricks on the engines and are often the cause of accidents!
Now don’t get me wrong. Children’s books are not the foundation of my idea of management but the stories rang a small bell. We all have our “Troublesome Trucks” and it got me thinking about some of the difficulties that we commonly experience in using performance measures.
I refer specifically to the difficulties often arising after we have gone through a systematic approach to deciding upon workplace performance measures – measures that really support the business strategies. The first Truck is called:
There has been a shift in use of workplace performance measures in Facilities Management based upon costs per square foot or metre towards costs per FTE or space per FTE (FTE of course referring to headcount – counted in “full time equivalents”). This is entirely reasonable given that the output of all workplace service costs and services is consumed not by buildings in particular but by the activities of people housed in them. But this raises the question of how the FTE number is arrived at, who is included in the number and how an accurate figure can be arrived at. What we really want to know is how many “heads” are supported by the building and its services.
Naturally the HR department has records of employees (particularly for payroll purposes) but it don’t necessarily know how many other people work within the organisation if those people are on short term contracts, work for suppliers and/or have been contracted locally by business departments. Even if we DO know how many FTEs are being serviced, we also have to think about the process by which changes in FTE are reported and how often? And whether the figures are available by location and by business unit.
Unless we agree across the business what figures to use in FTE-based measures and the ownership of FTE data, there will always be the potential for disagreements on workplace performance measures and potentially misleading information will be generated.
The next Truck is called:
These are performance measures that have not been designed with sufficient precision so that they lead to misleading information and erroneous decisions.
For example, we had one client that measured utilisation once a year and during the year, took views from line managers as to whether they felt the usage was higher or lower than the most recent measurements. The idea was to see if the managers felt things were manageable or whether the space was really under pressure. Although not a refined process, they felt this gave them a good enough feel for the space usage.
However, practical experience regularly shows a substantial bias towards people believing that their space is more highly used than is actually the case. Regular and frequent measurements of actual use are the only sensible way in which this performance can be understood.
Differences that arise from seasonal factors such as the business cycle need to be taken into account as well and possibly factored into the calculation of the performance measure before decisions are taken. For example, monitoring the use of utilities such as gas and electricity should be done in relation to identified drivers of demand such as outside temperature and daylight hours, together with expected occupancy, which itself will be driven by any seasonality in the business cycle.
A final Truck – called:
It is common in our industry to have KPIs and SLAs that are reported upon monthly. Many of these are financial measures and often reported as 12 month rolling averages which smooths out seasonal variations and hopefully shows long term trends.
However, while this seems a simple approach, it can give a false sense of security and hide workplace performance problems which can lie undetected. Often there is misunderstanding associated with the concept of “natural variation” – if you believe your average lies within the range of natural variation, you won’t take any action. But what if the natural variation has been wrongly determined?
The use of control (XmR) charts to monitor industrial or business processes can give a much more accurate idea of what is going on. They measure the moving range (mR) between individual (X) consecutive data point measurements (hence creating an XmR chart). This allows the natural variation values to be set more accurately.
This blog isn’t an appropriate vehicle for describing the technique in detail, but using this approach will show when one month’s figure is really different to others and hence when it is meaningful to base decisions upon it (by highlighting how it relates to the natural variation that is always present when measuring processes).
For further reading, Stacey Barr has a good summary: http://www.staceybarr.com/measure-up/three-things-you-need-on-every-kpi-graph/ and
a lay description of using XmR charts is well covered in Understanding Variation by Donald Wheeler ISBN-10: 0945320531