Byron Drew discusses how meteorologists are helping to bring a blind spot into focus within the FMCG supply chain.
During World War Two, radar operators discovered that local weather produced radar echoes and masked potential targets. At the time, these echoes were an unwanted nuisance and techniques were developed to filter this type of scatter. Soon after the war however, the value of viewing rainfall in real time rapidly became apparent and surplus radars were commissioned specifically to track rainfall. Today, radar meteorology is critical to short term weather forecasting and helps to protect the lives and property of millions every year.
In much the same way, weather driven demand volatility has previously been viewed by some as ‘scatter’, masking the truly valuable information behind it. This sentiment is changing fast. As improved forecasting and observing systems allow users of weather data unprecedented understanding of past and future weather conditions, volatility associated with weather is fast becoming a key source of insight, and even predictability.
The influence of weather on products on supermarket shelves is profound, from global supply, through transportation to storage all the way to the consumer’s decision on whether or not to buy. Some variability is easy to understand anecdotally: the amount of BBQ goods sales increase around warm sunny weekends and the amount of cold-remedy/Vitamin C increases around the first cold turn of the winter. However, the influence of weather extends well beyond this anecdotal relationship and in many cases this key source of predictability is ignored or at least not thoroughly understood.
In the past, without consistent data on sales or consumer behaviour and with only a handful of reliable weather observations across the UK, producing meaningful relationships between weather and demand was difficult or impossible.
Furthermore, once this relationship was understood, using weather forecasts to predict demand or wider behaviour was often discounted because of the relatively low perceived quality of weather forecasts.
A quantum leap in understanding
In every respect, the science of using weather to help predict human behaviour has advanced rapidly in (relatively) recent years. Widespread use of powerful IT systems, capable of capturing historical sales and other data, has allowed users analyse historical demand simply and efficiently. A quantum leap in weather observation data (including the previously mentioned weather radar and private weather observations) has allowed users to understand past weather in ways we never imagined.
Alongside this deeper understanding of past weather, our understanding of future weather is deeper and more accurate than ever. The rapid advance of computing power (and particularly cloud computing) has allowed numerical forecasters to create weather models that are more hyper-local and more accurate than ever, these models can now show us which borough in London will be warmest tomorrow or just how far down a mountain slope the snow will extend. Global collaboration has also allowed us to share resources in centres like the ECMWF to build weather systems capable of accurately predicting trends 2 to 4 weeks into the future and provide valuable guidance on seasonal trends up to 7 months ahead.
Today, a 5 day forecast is as accurate as a 1 day forecast was 40 years ago, and a 3 day forecast was only 20 years ago. This means that in one generation, we now know as much about what will happen on Saturday by the previous Monday as we used to know on the Friday.
Arguably as important as all this is the changing attitudes of the public and professionals within the sector to weather data and forecasts, which is now providing the collaborative framework between the Meteorological community and the FMCG community to drive insights, and ultimately improve business practice.
Meaningful business decisions
While the advancement of science is interesting, what is more exciting is what this allows a user to do with the information. In the past, a rather vague understanding of past conditions and potential future conditions led to few decisions being made based on a weather forecast. Now, users can use the deep insights they have learned from their data, apply it to a very accurate weather forecast and make real meaningful business decisions. A forecast of an upcoming warm weekend causes demand analysts to order more BBQ’s and Pimms.
A manager can roster more staff at a bowling green on the first sunny weekend of spring or a store manager can move the Vitamin C to the front of store to coincide with the first cold snap of winter, increasing opportunistic sales.
The simple example of lettuce helps show the depth and breadth of weather’s influence. A mild March (like the one we had in 2017) will produce an early UK lettuce crop. Warmer weather will increase refrigeration costs during storage. Spring rain and wind can causing flash flooding or bridge closures, which will delay or interrupt distribution.
Impacts on buying behaviour
Even when the produce reaches the customer, a warm sunny day will significantly increase the customer’s likelihood of buying. With today’s data, we can model lettuce growth based on past years and provide the farmer with guidance on how early he/she can plant based on growth-risk factors and alert supply analysts that the lettuce season will produce early yields, weeks before the lettuce is ready.
We can model storage electricity consumption alongside temperature and humidity and provide site manager estimates of increases/decreases in electricity consumption. Incredibly detailed and accurate rainfall forecasts help us to warn transporters hours or even days ahead of a potential flood risk and sophisticated wind force models help us alert transporters on whether a bridge may close and whether their vehicles will be affected on the road.
We can model lettuce sales based on weather behaviour and provide a demand analyst with an objective demand forecast four weeks before the warm weather hits, helping them to more accurately plan where to move their stock and when to put it on the shelves.
These decisions ultimately reduce waste, increase profit and prevent reputation damage associated with stock-outs. Using the right data in smart ways, businesses can and do seen real change in their operations.
As with weather radar, what started as unwanted ‘noise’ is fast turning into an invaluable tool to help us understand the tightening balance between supply and demand.
FMCG News July 2017
Logistics and the supply chain
Manager of International Meteorological Consulting at MetraWeather
As a professional Meteorologist, Byron has spent the last six years helping companies and individuals make effective use of weather data, from supermarkets estimating FMCG variability to utilities estimating power demand, and even rugby teams determine match-day strategy.
Manager International Meteorological Consulting
T +44 118 3805063
M +44 7864 140 441