Analysis helps performance.
‘We behave how we are measured’ is a statement applicable to supply chain professionals. Performance is being measured; but what if your performance could be improved through access to analysis of what is happening and projections of likely events? This is Supply Chain Analytics.
My previous post discussed the time it could take for new technologies applicable for Supply Chains to be generally adopted in organisations. This post considers techniques that require a lower capital investment and potentially a shorter time to adoption. A challenge will be the need to employ supply chain professionals who are comfortable with numbers. This could be the biggest hurdle.
An adaptation of the APICS definition of Supply Chain Analytics is:“The ability of supply chain professionals to analyse large sets of data using proven analytical and mathematical techniques (regression analysis, stochastic modelling, linear and non‐linear optimization, etc.), allowing the identification of patterns and correlations, perform comparisons and highlight opportunities”.
Supply Chain data consists of three types: transactional, time-phased (as in planning) and into the future, streamed (from sensors such as RFID and Internet of Things). The data can be analysed within the four elements of Supply Chain Analytics, identified by SAS, a software supplier in the analytics area, as:
- Statistical Analysis – Why is this happening? What opportunities am I missing? An example is to identify why there is a decline in production yield
- Forecasting – What if these trends continue? How much is needed? When will it be needed? With increased software capability, forecasts will be improved to incorporate market (rather than just sales) forecasts. These forecasts will incorporate structured and unstructured data. An example of structured data is actual sales by line item by period. Unstructured data is usually obtained externally, as a collection of data that requires re-construction to suit the organisation. An example is forecasts for sales, inventory and production of house-wiring cables produced and sold by an electrical cable company. Here, unstructured data might consist of: building approvals with a time lag to completion for each building type; possible housing estate developments based on developers’ land banks and commercial financing policies and weather patterns (for building sequence). The unstructured data would be aligned with the structured data prior to forecast models being applied
- Predictive Modelling – What will happen next? How will it affect my business? An example is to identify the effect on your sales if a competitor runs a promotion. What can you do to avoid cannibalization of your sales? Which products should you be spending your promotion dollars on?
- Decision Optimization –How do we do things better? What is the best decision for a complex problem? An example is optimizing your inventory plan to meet targeted service levels, considering: financial goals, supply constraints and available production capacity
Using Supply Chain Analytics
Numbers are the basis of Supply Chains, with analysis useful across a range of operational reviews. For example, the techniques could have been of value to the largest milk processor in Australia; it has announced that its profit for the financial year will be 50 percent of that forecast in July 2015. Based on media reports, the profit reduction has been influenced by:
- The company’s business plan required a move from generic products (which have their selling price governed by the global price of milk powder), to ‘dairy food’ brands, which carry their own value added price and margin. The introduction of new ‘dairy foods’ and achieving increased sales did not occur with the speed expected
- Inventories increased, with multiple short term leases for additional warehouse space
- The price for milk on global markets has been falling
- Sanctions imposed against Russia has caused European dairy exports to be re-directed to Asia
- The removal of quotas on milk production in Europe, resulting in increased production and exports
It appears the company was unable to reconcile and act upon data and information from multiple parts of their supply chains – political and economic events, global milk price trends, sales, actual milk supply, production capacity and inventories. This has cost at least $40m in profits, loss of reputation and retrenchment of the CEO and Finance director.
Encouraging new analysts
Analysis, often using spreadsheets, has been done for many years, but Supply Chain Analytics provides a higher level of mathematics and analysis to provide potentially deeper insights in a quicker time frame. This not only requires people who are comfortable with numbers but they also need to interpret what the numbers mean. When I was first shown the power of pattern recognition within predictive modelling, the analysts calculated that Tuesday was the busiest day for trucks to arrive in California from interstate. But being mathematical analysts, they were not in a position to appreciate what the pattern meant for planning by logisticians.
But where will Supply Chain Analysts come from? Searching the Internet for award level analytics programs and analysis course yielded only a few results in America and Europe, with fewer in Asia. Like me, there could be many Supply chain professionals that only recognised the importance of quantitative analysis when confronted with the need to do it. But where do experienced professionals learn the techniques of analySIS without enrolling in a specialised analytics Masters program? There are a few on-line courses provided by education and training establishments and software providers, but none provide a recognised certification – is thIS an opportunity for APICS?
Supply Chain Analytics has the potential to be implemented in organisations over a shorter time scale than hardware technologies; but the shortage of capable people is likely to hinder its acceptance. This could provide consultancies with an opportunity to increase their income streams.