work

British Gas
Winter is Coming

Cold weather directly affects demand for British Gas HomeCare products and the fleet of engineers who service boilers.

For every 1-degree that the temperature drops below seasonal norm, sales of HomeCare and Home Installs increase by 3%. Consumers search for help when in Boiler distress so beating the competition at high demand moments was vital.

At the same time, we needed to manage demand for engineers. There are a limited number of engineers in each area, and wait times increase during peak periods. If no engineers were available, we didn’t want to drive too many people through and have service suffer.

Our target? A 7% increase in conversions year on year.

 

So what was our big idea?

A search strategy that modified PPC bids based on both temperature and engineer availability.

We needed to be reactive to both factors. Getting either wrong would mean missed sales or demand outstripping supply.

We created a script that connected a series of data sets (actual temperature, seasonal norm temperature, city, postcode and engineer availability) and used these factors to amend bids and copy on a daily and weekly basis.

 

Making it Happen

At the centre of our hyper-reactive campaign was a Weather Bids sheet, providing all the bid modifier inputs to the script. The sheet pulled in real-time temperature by city and postcode via the OpenWeatherMap API, comparing it to the weekly norm by location.

Each week British Gas provided engineer availability by location while we updated the bid modifiers by location based on norm temperature and engineer availability. Whilst the parameters were set weekly, the script itself iterated by location by hour based on real-time temperature.

Sound simple? It wasn’t:

  • If the norm temperature for the upcoming week was 6C or below, and engineer availability was low, we added an extra modifier that reduced bids by 50% per degree below norm
  • If availability was medium, the modifier would be reduced to 10%25%
  • If availability was high, no modifier was applied as there were enough engineers to cope with demand A second script was created to activate bespoke ad copy whenever the temperature in each location reached 7C or below. The API inserted the real time temperature into the ads to create urgency.

Results

The approach increased click rate by 5% which in turn improved Quality Score. Our CPC fell significantly, meaning that Cost Per Acquisition fell by 16%. However the best news was that sales increased 45% year on year.

“Our data has always shown the importance of how weather influences demand, up until now we’d never been able to use this information effectively in search. Implementing these scripts allowed us to use this data to exceed sales targets and lower CPAs. We drove more value for the business but, importantly, we improved customer experience by promoting products when we could provide the right level of service.” Patrick Smith and Bhawna Garg Digital Marketing, British Gas

Previous
Game Changer – FIFA Women’s World Cup 2019