Quantitative / Qualitative Paradox in Supply Chains

00:08 Introduction
00:25 What is the idea behind the Quantitative/Qualitative paradox in supply chains?
02:58 What is a Qualitative judgement?
06:14 Without having a Qualitative understanding, is there a limit to what we can do with data?
08:11 What are the Qualitative steps that you would take at Lokad?
12:46 Do some of your clients get frustrated due to the fact that the focus is on this interpretation of the events, instead of it being on improving the numerical results?
15:43 How often should supply chain practitioners be revisiting the Qualitative approach they are taking?
20:16 What about that idea that you need to fall in love with the problem, not with the solution?
22:19 To conclude, why is it so important to discuss the difference between the Quantitative/Qualitative paradox?


Better quantitative results in supply chains are frequently obtained through better qualitative, if not highly subjective, perspectives rather than from better numerical methods. Most the breakthroughs at Lokad, that ultimately lead to quantitative improvements, were of a qualitative nature.

In supply chain, there is often this quantitative/qualitative paradox. At Lokad, our approach is the “Quantitative Supply Chain”. People frequently challenge this approach and wonder why there isn’t a simpler way to measure everything in pure, monetary figures. However, it’s far more complicated than that. It actually takes a lot of qualitative understanding to make sense of any quantitative figures we can give. Usually naked figures, i.e. “your company will save 10 million dollars”, are very misleading. The paradox is that although we have a highly quantitative methodology - numerical tooling, technologies for crunching numbers at scale, advanced statistics; all taking place on a programmatic platform - usually the only way to make sense of our findings is to make qualitative judgements.

One of the most important things to understand is that every measurement you can make is highly subjective. You could think that all personal judgements and biases can be removed, but this is practically impossible. Supply chains are part of a huge ecosystem of software, machines, people and processes; therefore judgement calls are required. As the saying goes: “turnover is vanity, profit is opinion, cash is king”. Depending on what you look at, is it something “real” or something that requires interpretation?

As Lokad advanced, beginning with a classic forecasting approach and producing only one straightforward number, we realised that something was missing and that was uncertainty. Our initial perspective was incorrect because we were dismissing any possible uncertainty. We began to see the subtleties and nuances in the numbers we were producing and this required us to take arbitrary judgement calls. It’s a naive idea to think that data will somehow just reveal itself and how it can be best used for your company, unlocking some kind of magic business improvement.

It’s frustrating because supply chains are complex - it would be nice to provide an easy, round number and just be done with it. Yet, supply chain is more often than not a game of balancing conflicting economic forces: the cost of stocks, the cost of higher production, the cost of more supplier flexibility, the cost of having shorter lead times vs. more stock and longer lead times… To sum all this up in one number is practically impossible. All this doesn’t mean that we should seek complexity just for the sake of it. There are many actors in the supply chain industry that do this, putting an over-emphasis on exceedingly technical solutions simply to appear savant.

To wrap things up, we talk in more detail about what makes a KPI worth investigating and whether its numerical value can qualify as surprising. In addition, we discuss Jeff Bezos and Amazon’s “Day 1” philosophy and how Lokad evolved from focusing purely on forecasts to understanding the importance of investigating probabilities and how this fits into the quantitative/qualitative tug of war.