Using intelligent sales forecasting to maximise revenue

We deployed recurrent neural networks to help CloudFM build a sales forecasting system to optimise pricing, identifying over £1.3m in lost revenue opportunities.

The problem

Maintenance costs can vary widely from building to building, depending on several factors from the property’s age and size to its use.

An accurate understanding of this fluctuation was essential for’s client, a national facilities maintenance company called CloudFM.  For a fixed annual fee, CloudFM offers business property maintenance services. If the fixed fee is shown to be higher than the actual property maintenance costs for that year, CloudFM will refund the shortfall to their client. To maximise revenue, CloudFM needed data insights to accurately forecast future maintenance costs

The solution

We partnered with CloudFM to scope, design and build a real-time statistical time-series forecast model. The deep learning model forecast the most likely costs and calculated the optimal fixed fee to charge the client. The system also included real time alerts to notify CloudFM if the fixed fee should be updated, or if it had been exceeded. 

£ 0
identified in lost revenue opportunities

The impact

  • Optimisation's system chose the optimal fixed fee based on the deep learning models

  • Real time alerts

    The system's real time alerts notified CloudFM if the fixed fee should be updated or if it had been exceeded

  • Increased revenue

    The solution identified lost revenue opportunities of over £1.3m per annum

If you’re interested to find out how AI technology could transform your business email or use the contact form below.