The sun is out, so it’s likely the barbecues will be too. Which got us thinking about one of our early projects for the UK’s biggest retailer Tesco.
It was 2015 and the UK’s BBQ market was worth over £7 billion. Statistics suggested that there are only five or six opportunities to have a barbecue in the UK every year. This presented a massive challenge for retailers. In the summer months, over-stocking results in wasted inventory, while under-stocking means a loss in potential sales.
Being barbecue ready
It’s much more complicated than simply understanding the weather. There’salso the idea of people being ready to have their first big barbecue of the year, and retailers have no idea when this will happen.
Back then our main offering was an AI forecasting approach we called Demand Prediction. Based on this, we used a combination of external and client data to create a predictive supply-chain model for Tesco that would be more accurate than their existing approach.
Essentially, we looked at people’s propensity to want to have a barbecue. Scanning Social media gave us a good indicator. For example, we found that people on Twitter were saying they weren’t going to work on Monday because they planned to have a big one on the weekend. The weather was obviously a huge indicator. Looking at Google search and past shopping patterns were also big insights.
As well as this, we sourced five years of sales data from some retailers and overlaid them with this noise from the internet. As our co-founder Steve King explains:
“We began to see a groundswell that would happen on the Tuesday before the weekend for people who were ready. And when you overlaid that sunny weather with the trigger for people being ready to barbecue, you can guess within 96 % accuracy how many burgers you’re going to sell on a Saturday the Tuesday before.”
10% increase in forecast accuracy
Our powerful Demand Prediction model was a customised version of what statisticians call ‘random forest regression’. Through this we were able to forecast a barbecue weekend that was 83% accurate – 10% more than Tesco’s forecast – up to three days ahead of the event. This enabled Tesco to increase their stock availability, reduce waste and boost revenues on what became a bonanza BBQ weekend.
Our Consulting Director Markus Frise was a Senior Project Manager at Tesco at the time and was impressed with the findings. He says:
“The work Black Swan Data did for me at Tesco helped us enhance models and processes that we believed were already best-in-class”.
Combining localised weather data and Social media conversations led us to develop more hyper-local seasonal predictions. For example, we were able to locally predict when cold and flu would hit four days before any other local authority in the world.
Prediction has evolved
Today our technology has moved on considerably. Now we offer big brands the most comprehensive view of their category and the consumers and trends that shape it. Our prediction analysis gives foresight up to 12 months ahead to help brands innovate quickly and smartly in an ever-changing market.
For example, we can predict the hot food trends that will shape this year’s BBQ season and why these matter to your brand.
So perhaps now the question is not if you’re having a BBQ this coming bank holiday, but what ‘flavour’ BBQ you’ll be having?