In my previous blog – Get Dirty with the Data 2 – I noted that cost is a lagging indicator – the outcome of a business process. The performance of that process is what drives its costs, so we need good measures of the effectiveness of the process and those factors which increase cost but add no value for the customer. So what should those measures be ? That depends quite a lot on the nature of your processes, but here is a starter set of performance measures to think about.
Firstly we need to measure the total flow time. For companies that make “call off” orders, then the number of days from order entry to delivery is a useful measure. Raw material inventory days, plus Work in Progress, plus finished good inventory days is also a good measure of total flow time; while the average time (with standard deviation) from project (job) initiation to completion (or customer sign-off) is used in make to order or custom environments.
Then we must calculate the total productive time. In physical processes, this should be fairly easy – it is the total value adding activity time (total of all cycle times), excluding rework, inspection, or other activities which do not add value for the customer. In service environments, and some manufacturing processes, productive time is not as easy to calculate because of variations in customer requirements. In such cases you need to undertake some data sampling.
Next we need to be able to measure the “waste” in the process – that is all the non-value-adding delays and impediments which cause the total flow time to extend beyond the “productive time”. Typically four categories of measure are used to highlight wastes in the process – quality, delivery, throughput and involvement.
First there is quality. Rework, scrap and other delays all cost money and do nothing for customer satisfaction, so we should measure the process “First Pass Yield” (or “First Time Through”) quality. It is often also appropriate to measure the time spent on rework, or the value of scrap (where the materials and components are expensive). It is also useful to measure downtime where equipment efficiency is a problem.
For most organisations customer service is measured by the percentage of deliveries “First Time in Full”, or a similar measure. Throughput is normally measured in terms of units per FTE (full time equivalent), or units per worked hour.
One thing you should not measure is “earned hours” or any other variance based on absorption. Earned hours is not a true measure of the productivity of the process because it tells us nothing about the difference between flow time and actual productive time, or about the wastes in the process. Earned hours is a measure of overhead absorption rather than of process efficiency, and, therefore, focusing on earned hours will encourage behaviours that improve absorption – such as production of large batches. Such behaviours will actually increase flow time by adding more waiting, inventory and, often, quality problems into the process.
The aim of process improvement is to reduce costs by taking out delay and waste, to create extra capacity to do more profitable work. This means that we need to be able to measure improvements. This is partly done by analysing the trends in the data and, for this reason, it is best to take the measures weekly (or even daily if possible) so that we get rapid feedback on the impact of process improvement activity, and prompt notice of problems arising. Of course, improvements are made by trained and motivated people, so a measure of involvement is a useful indicator of the effectiveness of improvement efforts. A skills matrix is a good start, providing evidence of flexibility and strength in depth; but we also need to see evidence of improvement teams actually working. The percentage of people involved in two or more structured improvement activities (cumulative) is a good gauge of the extent of an improvement culture. Some organisations measure the number of days (or hours) booked to improvement activities but I do not favour this approach – it often encourages the booking of time to “improvement” regardless of whether or not it involves a structured methodology and a clear purpose.
So a small number of indicators, measured frequently, can give us a very clear picture of how effectively a business process is working, and how much potential there is to bring total flow time nearer to actual productive time, and, thus, create the capacity to do more profitable work. If we then connect this data with the financial performance of the process we can start to see how performance drives cost and which improvements will deliver the most financial benefit.