By Howard Roddie, FuturMaster Senior Supply Chain Consultant (@HowardRoddie)
This is about the weather. In the UK we all love to talk about the weather, so I know some of you, at least, will still be reading. You probably already know that I’m going to talk about
the influence of weather on demand planning. So far I’ve been predictable. If you have a stake in demand planning, this should give you a warm feeling.
Soup and salad is in the title. You already know what I’m going to say. Well, OK, here it is… we sell more soup when it’s cold and more salad when it’s hot. There, I’ve said it. I know this to be true, because I once found a forecast group called “seasonal” that included both with the result that the natural seasonality of the two groups was flattened. Ever so slightly. Once we separated them, a more realistic pattern emerged.
More recently I was talking to someone with a background in the Restaurant business who gave more insight into this phenomenon. Apparently the relationship isn’t quite what it seems. Within the obvious seasonal pattern more subtle factors lurk. You see, when the temperature drops below 50F (my friend is American), there is a switch from salad to soup. This is quite as we would expect. But even if the temperature remains low, after three days, people switch back to salads. There are several reasons for this. Soups aren’t always perceived as healthy and people can only have so much soup before they get bored. In the restaurant industry, this is a problem as soup has a much higher margin than salad. At a simple level, you just throw a load of ingredients into a cauldron, simmer it for a while and ladle out when the orders come in. Salads have to be washed, tossed, dressed, arranged and served individually from scratch. Salads have freshness issues too. So, to entice people into buying soup, we have a much greater range of products. Soup of the day anyone? Or would you like salad – Caesar, Green or Greek? Much more effort goes into creating new soups for this reason. The three salads I’ve mentioned are everywhere, but soup has many flavours.
This is also played out on the shelves of the retailers where new flavours and innovative packaging are the norm as suppliers rush to create premium products to compete with the salad in its sad transparent film covered box.
OK I’m generalising and I’m simplifying. There are many very good pre-packaged salads out there. But, by seeing the relationship between soup and salad as just the weekly seasonal curves, we could be over-simplifying the situation too. There is often a belief that a single weather factor (usually temperature) has a direct statistical impact on demand. This is not true. The soup and salad scenario is just one example where the relationship is not simple. Weather can be a big factor, but the impact of a rise in temperature is very much dependent on other factors too – is it sunny or cloudy? Have we had four weeks of solid sunshine? Is it the first chance in the year for a barbecue or Braai? Is it the weekend? What is the retailer expecting and planning for?
Pragmatic forecasting using such information is a collaborative decision-making process combining judgement and statistics to make the best estimate from a few occasions (how often does a bank holiday coincide with a sunny day?). Look again at your data for weather related food products on a day by day basis versus temperature and see if a pattern emerges. I’d be most interested to hear about your findings. One thing is for sure though… if it’s warm in Northern Europe, don’t try to sell soup!