Early Signals and Forecasting

00:08 Introduction
00:25 What is the idea behind forecasting early signals?
04:29 What is the issue in terms of the quality of data?
06:35 Are there any other early signals that we could look at? What about the weather?
10:05 What can you do if you are confronted with an early signal?
13:34 Are there any industries which are classically slow to react which you believe could benefit from the use of early signals?
16:57 Do you think that in the future we can get to a stage where the technology will allow us to have more confidence in the results coming from early signals?
20:25 What would be your advice to supply chain practitioners working in those industries where capturing those rapid spikes in demand and initial trends is so important?


With a product’s success so often relying upon the latest trends, the ability to capture initial demand and detect new crazes is becoming increasingly sought after. Here we discover whether the rise of social media means that forecasting these trends is now possible and what companies can do to react based on early signals.

The general idea is that, by leveraging large data sources, you can obtain a glimpse into the near-future - for example something that is trending on Instagram, such as a celebrity wearing a specific type of garment. However, for supply chain forecasting early signals based on external data sources, there isn’t really anything currently available on the market.

Extracting data from social media may seem well and good, however most of that traffic isn’t even caused by real people but by bots, therefore it’s complex to discern what is real. Navigating through all this noise makes any forecasting highly problematic and it does not translate well into telling a supply chain precisely what it should buy or produce more of.

Not to mention, that when a picture goes viral on Instagram, Twitter or Facebook, it’s highly unlikely that the exact products featured are referenced and it’s complicated to predict what potential customers will take away from the picture and products. For example with a pair of shoes, would it be more the colour, the brand or the shape? In addition, sometimes things go viral for the wrong reasons, with the multiple views and comments actually coming from a negative point of view.

To wrap things up, we debate whether extreme events, such as the Covid-19 crisis, could be predicted and the prime importance of forecasting engines. We also go into more detail about how much trust you can actually place in early signals and the inherent complexity of supply chains. Due to this complexity, transforming forecasts into decisions within a supply chain is something that can never be instantaneous.