By rights, Steve McGivern, should be a worried man. As the National Forecasting, Roster & Planning Manager at British Gas Services, he has an impossible job: accurately predicting how many service engineers British Gas needs to service and repair millions of domestic boilers during any given week, even while the weather is becoming more extreme.

With more than 7000 engineers looking after 4 million customers, British Gas Service & Repair has used forecasting for some time to help plan how many employees they need, and where and when they need them. But cold winter months, like December 2010 when Britain suffered the coldest December in 300 years, can still be a challenge.

In cold weather, we turn on our boilers and we keep them on. Any faults that were just waiting to happen, happen, and we call an engineer so that we can continue to live in warm comfortable homes. And for some vulnerable customers, keeping their homes warm can literally be a matter of life and death.

When British Gas upgraded its demand forecasting, it turned to SAS to:

• Use more data from more sources, including:  geography, parts availability, vulnerable customers, staff working patterns and rostering constraints (such as the Working Time Directive).
• Provide a longer-range resource forecast (seven days instead of the current two). 
• Improve service to customers and better deploy its workforce. 
• Optimize delivery of the service demands of the forecast. 

In my view, the results – including a 98 percent accuracy rate – are spectacular, which brings me to this post's title:

Happy customers – With a 98 percent forecasting accuracy for boiler breakdowns, the number of customers who could be serviced on the day of their call went from 65 percent to 85 percent.

Happy staff – Because forecasting was so good, fewer staff needed to be on call during "anti-social" hours: They got more of their weekends back and could plan their personal lives better.They also had more trust that scheduling decisions were scientific and, therefore, less prone to human error.

Happy shareholders – Because of the results of SAS optimization, staff were able to spend more time each day servicing and repairing boilers than, for example, travelling between customers (down by 30 percent). The net result? Greatly improved productivity for no extra cost.

The list goes on.

So, whilst McGivern may have other things on his mind, planning how to get the best out of the workforce is one less thing to worry about. And other organizations with variable workloads and a large workforce (consider healthcare providers, postal and delivery service providers, and so on) could also reap the benefits. Take a look at SAS Forecasting and Econometrics software for help identifying previously unseen trends and anticipating fluctuations.