Why measure the performance of a supply chain? Because a supply chain is increasingly the sole competitive differentiator for an enterprise, and measuring its performance helps to improve it and detect and solve any problems. However, traditional enterprise measures of performance do not work for supply chains. To paraphrase a saying, the road to supply chain hell is paved with good intentions and bad metrics. Supply chains have other goals, cross over departmental boundaries, and now often extend outside a company’s perimeter too. Naturally therefore, a change of measurement viewpoint is necessary.
What Performance Should We Measure?
The simple answer is to measure what is important. Conveniently, in a supply chain just two things are really important: customer satisfaction and supply chain profitability. Whether a supply chain offers speed, variety, quality, cost-efficiency, flexibility, and/or any other attribute desired by the target market, the end results must be satisfaction and profitability. Within those two global goals, there are different sub-goals (such as speed, variety, and so on), sub-sub-goals (possibly on a departmental level now), and so on.
Understanding what kinds of goals and performance should be measured starts with understanding the value the company offers its customers, and the business strategy it uses to make sure customers keep buying from the company. From this understanding, supply chain metrics will then ideally be:
- A few essential measurements that cover all the strategic aspects of the supply chain
- Composed of relevant lower-level measurements showing factors mainly responsible for exceptional (good or bad) performance
- Available for management to proactively avoid problems and seize opportunities, not just react after-the-fact when a problem has already occurred.
One of the challenges in the agribusiness sector, as research from the Netherlands has shown, is to resolve the mismatch between famers and manufacturers, and between manufacturers and customers, about what is important and therefore what should be measured. Conversely, asking the right questions of the right entities can help supply chains do and measure the right things. IT vendor Dell, for instance, initially thought that the fastest processor speed was what counted in the world of computers. On surveying its corporate customers, however, the company discovered that product standardisation and reliability were often more important to its business clientele, and adjusted its offering and its metrics appropriately.
Shortcomings of Traditional Performance Measures
A minimal set of metrics that is adapted to a supply chain might include:
- Service metrics for measuring how well customer needs are met
- Inventory metrics, because of the significant impact inventory often has on performance
- Time, speed and flexibility metrics to measure how fast the supply chain responds to events
- Financial metrics to indicate profitability.
By comparison, traditional enterprise performance measurement (for example, pre-1980s, before the term “supply chain” came into being) has used a financial accounting approach and excluded the other types of metric listed above. Also, this approach is rooted in the past (what has happened) without offering much vision of the future (what will or is likely to happen). There is no provision for strategically important, non-financial measures of customer satisfaction or loyalty, nor of operational optimisation. Stakeholders get a short-term, financially-oriented view of the situation. They see profits and revenues, but receive no long-term visibility of value or competitiveness.
Variations on this accounting theme appeared to try to remedy the situation. Time and motion studies offered improvement initially in managing production lines and logistics, although sometimes efficiency was mistaken for effectiveness. Activity-Based Costing (ABC) attempted to link financial metrics and operational performance by defining activities in terms of component tasks and costs. In this way, performance of separate activities (e.g. two different product launches) could be compared in terms of cost-effectiveness. However, it was not before 1996 and the creation of the Supply Chain Council (SCC) by PRTM (Pittiglio, Rabin, Todd & McGrath) and AMR that SCOR (Supply Chain Operations Reference-Model) performance measurement arrived, specifically designed for supply chains.
Departmental Performance Measures – Necessary, but not Sufficient
Separate departments, which contribute to the supply chain as a whole, usually have their own performance measures. Sales and marketing will assess performance in terms of market share, revenue, sales growth, and customer satisfaction, for example. Manufacturing will monitor unit and labour costs, labour and plant productivity, quality and output, to name some common metrics. Engineering will evaluate according to product features created, time-to-market, effectiveness of design for manufacturing, and labour and materials costs.
While each department should be able to measure its own performance, many of the examples of metrics cited here do not extend from department to another. They cannot therefore cover the supply chain as a whole. They may even lead to conflicting departmental goals. For example, sales and marketing push for higher inventories and lower volume orders to keep customers satisfied. Meanwhile, manufacturing builds higher inventories but through longer, less flexible production runs. Logistics tries to reduce inventory, yet batch orders together for greater efficiency, possibly pressuring customers to wait longer for deliveries or order in larger quantities.
If each department only tries to optimise its own performance, the result may be functional silo-style operations and mediocre (or worse) performance of the supply chain overall.
Performance Measurement adapted to the Supply Chain as a Whole
SCOR is one of the best known performance measurement methodologies specifically adapted to the supply chain as a whole. Overall, SCOR provides a reference model for processes, performance and best practices for different categories of supply chain, each category being defined by the product supplied and the type of customer served. Performance is analysed for the key SCOR supply chain processes of Plan, Source, Make, Deliver and Return. Metrics are defined at three levels:
- Level 1 – the most aggregated level, measuring overall supply chain performance
- Level 2 – next level down, but still high-level measures that may span multiple SCOR processes
- Level 3 – a further level down; may or may not cover an entire SCOR process (Plan, Source, Make, Deliver and Return).
SCOR model users then use the specified combination of cycle time metrics (notably “cash-to-cash”), cost metrics (such as cost per shipment), service and quality metrics (correct deliveries made on time), and asset metrics (for instance, inventories). They can also compare their performance with SCOR benchmark data (as available) for their particular product/customer supply chain.
Other models have also been specifically designed for assessing supply chain performance. An example is ASLOG, a French model using a different set of metrics compared to those of SCOR. Some enterprises use SCOR and ASLOG together for finer coverage of the different aspects of supply chain performance.
Besides these “purpose-built” supply chain performance measurement schema, more general enterprise-level performance measurement models can also be applied. Balanced Scorecard is one of the more popular ones. When it is applied to supply chains, measurements can be made from the:
- Financial perspective – such as the costs of production and of warehousing
- Customer perspective – for instance, on-time delivery and order fill rate
- Internal business perspective – manufacturing achievement compared to plan, error rate in forecasting, for example
- Innovation and learning perspective – cycle time for new product development, as a case in point.
Performance Measures Beyond the Enterprise Boundary
Increasingly, several actors determine customer satisfaction and profitability in end-to-end supply chains, rather than one separate entity. Even if end-customers associate one particular enterprise with a supply chain, through the promotion of a product brand such as Apple, Nestlé, or Sony for example, all the players in the supply chain must perform well.
Correspondingly, metrics for supply chain performance must also extend beyond a single enterprise. Departmental measures for each internal operating unit and enterprise-wide measures across individual functional units have to be supplemented by inter-enterprise measures for processes that span several companies.
Often, the same performance measures already used enterprise-wide are applicable at the inter-enterprise level too. Examples include:
- Order-to-cash (order fulfilment)
- Concept-to-first-sale (new product development and market launch)
- Cash-to-cash (total cycle time from materials purchase to customer payment)
- Perfect order ratio (percentage of customer orders fulfilled without flaws or errors in the references delivered, the place and timeliness of the delivery, completeness and condition of delivered goods, and invoicing of the customer).
- Total landed cost to get products to the end-customer, including all the component costs of sourcing, production, inventory, warehousing and transport along the entire length of the supply chain.
This may sometimes be a challenge to brands that have stamped the supply chain with their identity, but where the company owning the brand does not directly control the other companies contributing to overall fulfilment. Franchise operations in the fast-food sector are just one example. However, it remains critical to be able to monitor, intervene in, and ensure corrections at all stages of the chain. This requires both appropriate end-to-end global performance measures and detailed local performance measures (at departmental levels in different actors in the supply chain) to then pinpoint root causes of any problems and identify options for resolving them. Third-party providers, whether in manufacturing, warehousing, transport or any other part of the supply chain should therefore already have adequate departmental and enterprise-wide measures of their own. They should also have appropriate visibility into the end-to-end performance of the supply chain to better understand and improve their contribution to overall performance. Some cloud-based software solutions offer the means for several providers and actors in a supply chain to work together better in this way.
Supply chain performance metrics may still lead to problems if they are incorrectly defined or misunderstood.
- Problems in definition of the measure. The case of Dell mentioned above concerned the definition of the wrong performance measurement, although Dell was smart enough to check with customers and choose a better mix of metrics afterwards. There is also the case where the performance measurement is relevant, but the definition may be ambiguous or faulty. Purchase Price Variance (PPV) is an example. For a given part used in production for instance, it is defined as [baseline price – current price] X volume for a defined period. However, the baseline may be taken to be a past data point, a past average, a published index price, or some other price datum. In addition, PPV does not reveal whether changing or reengineering a part in a process has cost more or brought savings, as PPV for the new configuration begins again at zero.
- Problems in interpretation of the result. In simple cases, the more costs are reduced and the more productivity is increased, the better it is for the enterprise. However, supply chains are typically not simple cases. Cost reductions and productivity increases are not always desirable outcomes, particularly when they adversely affect another strategically important parameter. The inventory-transport dichotomy is a good example. The simplest approach to transport costs would be to reduce them. However, this tends to raise inventory costs, which are proportionately greater than transport costs. In this sense, it makes sense to increase transport costs, on condition of course that inventory costs go down. In yet other cases, both costs might be deliberately increased, for example, if an enterprise is promoting same day delivery as a competitive advantage to its market.
Future Performance Measurements May Change Again
Finally, supply chain performance measurement should be revisited as part of a regular drill. Markets change and with them business objectives, programs, products and services. Modular approaches like SCOR can help businesses adapt to new situations. They help them to produce the right handful of key metrics to monitor overall performance, with lower-level metrics to help mid-management diagnose and solve any problems. At all times, it is also crucial to remember that supply chain performance measures are for measuring results, not actions, and for helping customer satisfaction and supply chain profitability both stay on an upward trend.