For years, supply chain managers and specialists have been striving for more and more transparency of the supply chain. However, the current dynamics require something much more important: predictability.
Internet of Things
In itself, striving for transparency makes sense. Insight into events in the supply chain, and especially real-time, offers the opportunity to intervene earlier and better. But the drawback remains that people react to events that have already taken place. Eliminating structural flaws remains important, but in this age of IoT this is no longer enough.
Developments are moving so fast that reactive behavior is immediately punished. Correcting your inventory levels is becoming less meaningful by the time it has an effect throughout your supply chain. Where the cranking effects used to be transparent, it is now a sequence of waves that wash up on your ‘supply chain beach’.
The digital supply chain with a consumer who changes his preferences and needs at an increasing rate – which means that turnaround times are even further compressed – still unintentionally leads to more stocks in the chain for many companies. Despite frantic efforts by the supply chain manager to lower this, the 2016 State of Logistics Report shows.
The reasons for this are diverse; shorter turnaround times and the fact that consumers exercise their demand both physically and digitally. The increasing diversity of consumer demand is forcing companies to maintain a wider range. The multichannel concept also still leads to problems and cross stock optimization is difficult to achieve due to the different service levels.
Traditional retailers who fear for their market position increase their range and stock in order to continue to serve the customer. Finally, many companies have limited insight into their supply chain because it is often too long, up to 7 tiers. Transparency then becomes an impossible task and control of the supply chain is left behind. In short, many problems have a man-induced buffer, planning is still too often based on averages, statistical lead times, ERP systems that are difficult to adjust, so that every reaction is reactive by definition.
The most important factors in any supply chain are lead time, variability, costs, capacity and buffers. The development of these factors in your supply chain must be continuously monitored, understood and predicted in order to continue to meet your service levels.
However, it is myriad factors that influence here and even beyond. The economic situation of citizens, social media, the weather, price developments, traffic jams, climate disruptions, etc.
Traditional transparency largely ignores these signals and most Control Towers still lack the possibilities to convert such big data into usable supply chain information. The big challenge is what information is relevant, what are the correlations, what are the actual causal relationships. The problem with big data is what it is: big data.
During one of my projects at a cinema organization, the question was whether it could be predicted which visitor numbers a film would reach, on which days and at what times. The underlying goal is better coordination of visitor numbers at specific times in connection with the capacity of the theater and sales before the start of the film and during the break (stock issue).
A university group of econometricians had analyzed visitors for several years and tried to relate to the weather, season, promotion activities, time of show, type of visitor group, oral advertising and competitive films (in their own cinema or third parties).
In the end, no clear conclusion was drawn. Additional research from my project team brought up new elements: the success of the film in the US, the time of its release, and similar films in the same time period.
Example: films that are successful in the US appear to be among the top 10 in the Netherlands in more than 80 percent of cases (in much the same order). As the total number of visitors has been fairly stable for years, this was easy to calculate. If two similar films are shown in a certain period (quarter), they ‘eat’ each other’s visitors. Pre-sale of tickets offers direct insight into demand, but at box office sales people seem to decide at the last minute.
What the real example above shows is that there is a lot of information available, but you have to watch out for false correlations. Do not hesitate to search deeper until you have your finger in the right place.
Analysis and statistics
The big problem with big data is the amount of data that can hardly be analyzed by people, let alone that the right conclusions can be drawn. It requires advanced algorithms and artificial intelligence to analyze large amounts of information from social media, external and internal circumstances, make the right connections and feed your ERP or other systems with the results.
Organizations such as Google and Facebook use this approach to optimally serve their digital customers. To provide customer specific information at the right time and place. From advertisements to relevant media articles. Now the challenge also lies with supply chain managers. Your future is not transparency but predictability.