By Simon Ellis, Practice Director, Supply Chain Strategies, Manufacturing Insights, an IDC company
"Our business is getting buried in data, we don't see obvious immediate uses, so we 'round file' most of it" said a former colleague of mine five years ago as he pondered the increasing amounts of data available to the supply chain. As true as that statement might have been then, it is even truer today. Supply chains are facing a blizzard of data and are challenged to figure out how to collect it, turn it into useful information and then actually act upon it in some kind of practical way. Whether it is the reduced number of "eye balls" available in the supply chain to look at data -- and the corresponding increased reliance on 'no touch' business processes -- or the fact that there has been so much focus on "sense" but not as much on "respond", supply chain professionals struggle to know exactly what to do with the data once they have it.
A question well worth asking is: "if I had this piece of data regularly available to me, what would I do operationally differently than I do today?" This question gets us quickly and squarely to the supply chain challenges faced by aspirations of "customer-driven" or "demand-driven". It is not enough to just have the data and the analytical capability to extract key insights: Your supply chain must also be able to respond in some useful fashion or the data is academic. It is similar to the goals of having "supply chain visibility" -- a useful capability, but if in the end you are not able to do anything with the enhanced visibility, it becomes irrelevant.
Demand-Aware, Fulfillment Driven
It is for these reasons that at Manufacturing Insights, we tend to think more along the lines of being "demand-aware, fulfillment-driven". We believe the practical application of fulfillment-driven techniques will ultimately prevail over the lofty demand-driven dreams to deliver supply chain efficiency. Based on our research, we expect that supply chain planning capability will elevate demand-shaping techniques and move more to a fulfillment/supply network-driven rather than demand-driven approach. Companies are finding the challenge of moving from a capacity optimization approach to a "buffer-inventory optimization" approach difficult to overcome given the cost trade-offs. Even in a perfect world of complete instantaneous capacity to handle the largest possible demand spike, the "pulsed" nature of supply chains (particularly distribution) requires buffer inventories. Better visibility can reduce unnecessary elements of inventory, but this pulsed nature remains, necessitating some level of buffer inventory. A true demand-driven approach also requires significant supply capabilities (for example, more rapid cycling of SKUs/manufacturing flexibility and capacity), which, at least in the short to mid-term, businesses have been reluctant to do.
The Data Lifecycle
Just because companies are still struggling to most efficiently use the massive amounts of available data, doesn't mean that there is not some relief on the horizon. We see a number of influential software vendors -- Oracle, SAP, IBM and Teradata among them -- furiously developing demand signal repositories (DSR). These demand signal repositories offer a "one-stop-shop" for demand data in an organized, efficient database. There is also the RFID-inspired, electronic product code information system (EPCIS), which holds the promise of similar capabilities for RFID and other sensor-derived demand data.
At Manufacturing Insights, we have developed a relatively simple way of looking at data acquisition and usage. In Figure 1, we illustrate this "identify-capture-analyze-affect" approach. We portray this model graphically as a circle because of the natural feedback loop that we see in our research and in many industry conversations we have about data and data governance.
Clearly the first step is to identify the various kinds of data available, and an initial assessment of how each data element can be used. Capturing the data, in some efficient way, is the next logical step -- today we see customer portals where suppliers can extract the data they want, institutional data sources, like Nielsen and IRI, as well as category management and plan-o-gram data sets. Then, the data must be analyzed and turned into actionable insights. This is where companies begin to struggle as overall supply chain headcounts fall annually and processes move more to "no touch". Here is where analytical capabilities in applications, like DSR and EPCIS, can be useful to manage to the proper exception levels. The last step, of course, is then to do something with the information you have extracted -- affect the supply chain in some positive way.
In the center of the model is master data management - and it is there for a reason. If data synchronization efforts of the last five years have shown us nothing else, they have highlighted the often dismal state of data governance and data accuracy. A technologically-facilitated, sophisticated data exchange process is ineffective if it is based on a foundation of inaccurate data. This is clearly not "rocket science", but it is worth observing that as the supply chain becomes more reliant on automated processes and demand data to run efficiently, it will be less able to tolerate poor quality master data.
The Way Forward
This article has really only scratched the surface of data, data proliferation and the emerging business processes and software applications that are becoming best practice. I do not imagine anyone would argue the fairly obvious premise that supply chains will accelerate their usage of demand and supply data to continue to drive out cost while improving speed and service. How "best-in-class" supply chain organizations do this over the next few years will be interesting to watch, and will be a continuing area of research coverage at Manufacturing Insights.