I’ll start with an admission – although I’ve spent the majority of my career as a supply chain professional, I’ve never been formally trained in the field. Like many others in the industry, I’ve had my share of on-the-job training and continuing education (APICS, for instance), but my formal schooling was in chemical engineering. So I came to the supply chain field as somewhat of an outsider – which may not have been such a bad thing. For example, my chemical engineering background helped me think about supply chains in terms of inputs and outputs, and dependent and independent variables. Perhaps because of this outsider’s perspective, I’ve never been shy about challenging the status quo, or pushing back against conventional wisdom. And one aspect of the conventional wisdom in supply chain management that has always bothered me is the framework of metrics that we use to measure, operate and improve supply chains.
I know what you’re probably thinking – aren’t supply chain metrics pretty well established by now? Haven’t we beaten this topic to death over the last several decades talking about KPIs and balanced score cards, and adding increasing sophistication to the measurement of forecast accuracy, service level and cost? Yes, this is indeed true – over the years, much work has been done on metrics in order to define, measure, analyze, improve and control supply chains. Frameworks like SCOR have been introduced to provide a formal basis for modeling and quantifying supply chains. Armies of management consultants have come up with ways to integrate supply chain metrics into the balanced score cards that are used to run companies. Processes like Sales & Operations Planning have been designed to align supply chain performance with financial, operational and commercial performance. I have no disagreement with any of this – on the contrary, each of these developments has been beneficial in moving the field forward, and making supply chain management essential to the success of businesses of all kinds. My more fundamental, nagging issue is this: are we critically looking at the metrics we use, and asking ourselves (1) if they are the right ones for our business, and (2) if they are driving the right behaviors? And I suspect the answer to those two questions is “No,” in many cases.
Let’s go back to basics for a moment. The need for metrics is underpinned by something Lord Kelvin is purported to have said, “If you cannot measure it, you cannot improve it.” I’m paraphrasing what he said, of course, but while his basic message is simple enough, it doesn’t really tell us a whole lot about what to measure. We all know that there are measurements that unambiguously point to how we can improve things, but there are also a host of others that either offer no clue to what actions should be taken, or provide information well after the time to act has passed. And then there are others that are purely diagnostic in nature, that tell us something about the health of the system, but are not relevant to improvement, necessarily.
This is where the notion of leading, lagging and diagnostic indicators comes in, and the best way to explain it is through some examples. Let’s say we are trying to measure a student’s academic performance. Grade point average is an obvious and important metric, but it bears noting that GPA is an output measurement, or lagging indicator. It’s a dependent variable, not one that we can directly manipulate to improve outcomes. So if we want to improve a student’s GPA, what should we measure and work on? It could be time spent in class, or in doing assignments, but more fundamentally, we need some way to measure how well the student understands the material (one could argue that this was the original intent of grades and GPA). One way to do that would be through assessments done as and when the material is being taught. When a chapter in the textbook, or a segment of the course syllabus is completed, the student takes an assessment, and the scores on the assessment provide a measure of how well the student understands the material. This metric in turn could drive improvement on the student’s path to the final grade on the subject. So a metric that measures understanding of the material is a leading indicator — an independent variable, an input that can be manipulated to drive better outcomes. Next, there are diagnostic measures – metrics that measure the health of a system, or provide a baseline for the system as a whole. For example, if a student is preparing to take the SATs, he or she might take a diagnostic practice test at the start of the process to establish a baseline and identify problem areas that require particular attention during the preparations.
Now that we have done the groundwork to understand the different types of metrics and what they are good for, where does this take us? Should we focus on leading indicators for our supply chains? Lagging indicators? Diagnostic measures? Not surprisingly, the answer is that we need all three. Output measurements are critical for any system – to understand performance, we need to measure what the system delivers. So, for a supply chain, we need lagging indicators that capture the reliability, responsiveness, cost effectiveness, efficiency and flexibility with which goods are delivered to the satisfaction of the paying customer. Many of the familiar supply chain metrics that we know and love fall into this category: forecast accuracy, perfect order, lead time, inventory turns, cost-to-serve, and return on assets. And how do we drive improvement in these output measures? As we have discussed before, we need leading indicators that we can influence before delivery happens. What metrics could help us do that? Basically those that measure how well we source, make and move products in readiness for delivery. Are suppliers delivering raw materials on time? Is capacity available to make the product when needed? Are work centers able to complete production as scheduled? The metrics that can answer these questions include actual vs target supplier lead times, plant availability, schedule adherence, and actual vs takt time. The last of these, takt time, is a metric borrowed from the world of Lean Six Sigma – it is the rate at which product needs to be manufactured in order to meet customer demand. The third and last element of the metric framework is diagnostic measurement. Again, I find myself strongly influenced by the Lean Six Sigma approach to process improvement – and from this perspective, cycle time is a particularly compelling measurement for understanding the health of a supply chain.
Now that we have our leading, lagging and diagnostic indicators, what remains is to put them together for a particular supply chain and business. Remember, we cannot drive the right behaviors and get improvement if we don’t have metrics that make sense for the business. For example, forecast accuracy at the mix level may not make sense for some products. It may be better to direct those resources towards a just-in-time operation that replenishes components to a reorder point, and assembles to order based on known (not forecasted) customer demand. Similarly, many fast-moving consumer goods operations are moving to a delivery metric based on “sell-through” rather than “sell-in” (meaning that delivery is considered complete not when the product is shipped to the retailer, but only when the product is sold off the retailer’s shelf – which is what ultimately matters).
So, there you have it – not a new set of metrics, or a new type of balanced scorecard, but rather an attitude of critical appraisal towards what is measured and why. In summary, then, here are the three main points I’m trying to make –
- View every metric with healthy skepticism. Focus on the ones that make sense for your supply chain and your business.
- Ask yourself what the metric is for, whether it’s a leading, lagging or diagnostic indicator, and what behavior it’s intended to drive.
- And then critically examine whether the chosen metrics in each of those three categories are meaningful in the context of making the business successful.
Because we shouldn’t forget that there is a dangerous flip side to Lord Kelvin’s maxim – if you measure the wrong thing, you won’t improve what really matters!
© Tharuvai Ramesh