We often hear about a rare or “lost” book that has been found hidden away in a library or an unknown masterpiece that has been discovered in a museum’s collection. They were always actually there, of course, but no one knew about them. And so no one specifically searched for them. When discovered, however, they were immediately recognized for what they were.
This analogy can be applied to data analysis – but only to a certain point. Decisive aspects of data analysis are unfortunately more difficult. So can you imagine your surprise if someone told you facts about your supply chain you had no idea about – and then went on to show you the relevant data from your own IT systems? The data is there – a fact we are aware of. But the data can only disclose the information we need if it is compiled and analyzed correctly. We need to know what we are looking for. To make use of another analogy, it’s like a jigsaw puzzle: it’s much easier to put all the pieces together if you have an idea about what the picture is supposed to look like at the end.
Data availability is no problem at all: current data is available in the ERP- systems, from attached or separate reporting systems or data warehouses that contain data in “cubes”, which then technically also allow for data entanglement based on selected criteria. Historic data is all stored – demand & supply plans, purchase orders, sales, deliveries, goods received and issued, financial data, master data. Data storage is also no problem and – provided you don’t need constant access to all the data –cheap as well.
Analysis software and tools are very much in vogue and available with a large choice of features for data manipulation and visualization as well as with options tailored to specific industry sectors. So how is it even possible that I discover something new? Tools assist with operations and presentations, but selection and linkage of the right data must also be supported by expertise. If you have a problem in your supply chain, asking for “a data analysis” is the wrong approach, and won’t get you any closer to really fixing it. You can run many types of analyses, but you have to start with target-oriented hypotheses.
To begin with, there must be a fundamental understanding of how one’s own supply chain processes work, coupled with the question “how should things be running in this case?” Based on the discrepancies identified between “as is” and “to be”, the experts involved then develop hypotheses about the discrepancies, which can then be verified or negated through analysis of the applicable data. Once the discrepancies are confirmed, “to be” scenarios can be created to quantitatively assess the impact of changes. Next, the root cause(s) of the problem(s) must be identified and corrective measures developed to set the changes in motion.
Optimizing an entire supply chain end to end is a much more extensive undertaking. To begin with, an “as is” status must be described based on the main performance statistics and key parameters. Although we make use of existing data in this case as well, the focus is much broader at the beginning. The approach here should be to start on a large scale, and as we zoom in to deductively pull relevant data together. The result is a quantitative and – very important at this stage – unbiased description of the status “as is”.
Before defining potential improvements, a couple of questions needs to be answered:
- What is my company’s strategy, and what is the role of my supply chain in that context?
- What are the operational manifestations of our strategy, and how does my supply chain need to be set up in order to offer the best possible support?
- What targets do we need to set, and what target values do we aim for?
The more detailed the answers to these questions are, the more specifically can the discrepancies between key performance parameters (optimal status) and the present “as is” status be assessed based on the results of the data analysis.
Of course, single analysis results should not be viewed in isolation: supply chains are complex structures with numerous interdependencies and – due to their construction – overdetermined systems. We cannot optimize everything as a whole, but instead must prioritize. So, if analyses reveal quantitative potentials, it doesn’t necessarily mean that they have been overlooked so far. It is possible that they only came into being through other optimization processes that were of higher priority, or that it was not possible to optimize them together with other processes. Thus, while formulating hypotheses and modelling possible “to be” scenarios, it is imperative to always take the end-to-end supply chain into account in order to avoid “local optimization” that is detrimental to other processes (or at least to be aware of any possible negative impacts).
In summary, be bold enough to discover and activate hidden potential in your supply chain – in house or together with experts in the field. The data is there, go find it and make use of it!