Flip flops, sunglasses and sun hats across the country will have been dusted off during the first few weeks of April. The mini heat wave saw temperatures reach 25°C, way above the long term monthly average of nearer 13°C. With the May Day bank holiday here many of us will be planning the opening of the Great British Barbeque season. So isn’t it just typical to see the temperature plunge this week and find showers, snow and frost hitting parts of the UK?
This has left many of us wondering if we’ll be outside flipping burgers on bank holiday Monday or tucked up inside on the sofa. Many retailers will be asking the same question.
Seasonal changes have a major impact on grocers, supermarkets and DIY stores. The Met Office estimates that a change in temperature from 20°C to 24°C results in a 42% increase in burger sales. Sales of strawberries rise 20% on the first hot weekend of the year. Predicting and being prepared for these changes can boost retailers’ turnover and profitability. For example the combination of warm weather and the four day weekend for the Royal Wedding contributed to a 5.2% increase in like for like sales in April 2011 according to the British Retail Consortium.
But retailers need to strike the right balance to avoid marking down or writing off excess products, or running out of stock completely. Data and analytics play a major role in achieving this balance. So what kind of techniques can they use to predict the arrival of summer?
British weather is too varied to rely only on historic trends as those affected by the summer 2012 hosepipe ban, which was lifted after abnormally heavy rainfall, will know. Many retailers make significant use of the improving accuracy of shorter term weather forecasts. A three-day weather forecast today is more accurate than a one-day forecast in 1980. Supermarkets use this data to make swift changes to production lines. For example Morrisons divert from producing beef mince and casseroles to make burgers within minutes of receiving a forecast for good weather. As longer term forecasts and geographical variations become more reliable manufacturers and retailers can plan for these events with greater certainty.
Look who’s talking
However the relationship between barbeques and temperature isn’t quite as straight forward as it at first seems. We are deeply affected by the ‘first signs’ of changes to the weather and this is when seasonal products sell in the greatest quantities. Tesco analysed five years of data which showed that sales of barbeque related food peak in the spring rather than the summer despite the usually higher temperatures in June to August. This makes it increasingly important for retailers and suppliers to predict when the early surge in market interest will occur.
A similar pattern is found when we compare our web activity to the average temperature over the last 12 months. Using data from Google Trends we see the UK’s barbeque related web searching is highest as temperatures were rising in the spring and early summer of 2014. Temperatures and web searches both peaked in late July to coincide with the start of many school holidays. The number of barbeque related searches then fell steeply during late summer and autumn while the average temperatures dropped at a shallower rate. The recent rise in temperature in April 2015 has seen many of us excitedly reach for the internet to search for barbeque related items again in anticipation of the summer ahead.
Weekly internet searches for “BBQ” in the UK and average temperature (May 2014 to April 2015)
Sources: Google Trends and AccuWeather.com
Social media and other web activity can provide insight into interest in seasonal products or leisure events. By applying this with weather forecasts retailers and manufacturers are better able to estimate demand levels.
Location, location, location
Of course the British weather has distinct geographical variations, and so do customer preferences. Sainsburys found that barbeque sales will triple in Scotland when the temperature hits 20°C, whereas it needs to be 4°C warmer to achieve the same result in London. To know which products will sell well in different locations supermarkets can turn to their greatest data asset: customer loyalty and sales data. Tesco combine customer behaviour modelling for sales of each product and store with weather data and key events (like major sporting events or bank holidays) to automatically order stock based on live sales. Analytic systems such as this allow supermarkets to work with suppliers and distributers to plan demand for products and reduce waste.
Data and analytics aren’t just helping retailers and suppliers match our demand for seasonal products; they use it to make us buy them in the first place. Those of us with loyalty cards and a penchant for gardening will probably receive offers for garden furniture or tools in the next few weeks. Our internet search history and social media profiles will be used to advertise specific products to us. Retailers will also use more subtle techniques to raise our interest such as sharing their favourite summer drinks and barbeque recipe ideas.
Advertisers are making increasing use of weather dependent digital marketing as well. Digital out-of-home (OOH) panels at bus stops, train stations or on taxis can detect local weather conditions and display relevant marketing messages. Stella Artois reported a 65.6% increase in Cidre sales with a digital advertising campaign which was automatically triggered by a 2°C rise above average normal temperature.
So if you still don’t know if you’ll spend this weekend indoors or outdoors it may be worth looking at what’s in the shops or the adverts around you. Retailers and marketers may have a better idea than you do.