Taylor Wimpey builds cloud forecasting process with Anaplan

UK homebuilder Taylor Wimpey has streamlined its annual budget forecasting process with the help of Anaplan.

Taylor Wimpey – winners of the Computer Weekly European User Awards for Enterprise Software – enrolled the help of Anaplan’s cloud-based modelling and planning platform after deciding its previous suite of complex Excel spreadsheets was inefficient and approaching its end of life.

Previously Taylor Wimpey relied on its Excel suite solution to run forecasts between its 24 UK regional offices. The process was undertaken every October and took six to eight weeks to complete. The company felt the solution was no longer flexible or fast enough to respond to business changes across Taylor Wimpey’s geographically disparate teams.

Taylor Wimpey started working with Anaplan in April 2012, after it turned down several on-premise financial planning solutions from the likes of IBM and Oracle. 

As a result, the company can now generate reports and forecasts with vast amounts of data, in real-time, with Anaplan’s software-as-a-service (SaaS) model. This can be added to or adjusted at any stage in the reporting process. The Anaplan system offers Excel-style functionality, but removes the need to manually collate disconnected spreadsheets across different locations and hierarchies in the company. 

Anaplan’s platform is based on its HyberBlock architecture, designed to unite memory data store and calculation. The platform automatically records updates at a granular level by amending only the affected cells. Taylor Wimpey carried out its first full budget on the Anaplan platform in October 2012.

David Nicholson, business systems manager at Taylor Wimpey, said that, with Anaplan, the company’s annual forecasting process is now better controlled and more efficient: “The use of a single cloud-based model removes many hours of effort involved in distributing, managing and consolidating Excel spreadsheets. Our aim was to simplify and modernise our estate, by reducing our application footprint and decommission a lot of our legacy systems,” he says.

“We moved to Anaplan as our Excel 2003 based forecast solution was under threat as we were moving to Windows 7 as part of a wider technology refresh project. We invited several technology vendors to demonstrate for us and we found Anaplan to be a fantastic toolkit.”

Nicholson says the company entered into a period of extremely detailed due diligence with Anaplan: “It was the first time we were working with someone in the cloud who wasn’t Microsoft and we wanted to ensure we were provisioning an environment capable of supporting the needs of a community of citizen developers in our business.”

Taylor Wimpey’s finance directors, in each of its teams, can now re-use their operational forecasting data from the firm’s ERP system. This data can be manipulated externally by a centralised team before being uploaded to Anaplan.

Taylor Wimpey has seen a 75% reduction in applications through Anaplan and an ERP system replacement. The company uses the ERP system Coins and is also currently undergoing a Microsoft 365 desktop refresh.

The company is looking to expand on its work with Anaplan by integrating an application performance interface, which will enable its finance directors to instantly upload data to the cloud. Each team can update its data throughout the year instead of only once, during the annual reporting period.

Andy Feldon, IT director, at Taylor Wimpey, says: “We found Anaplan quick to replicate functionally and quick to develop and operate, which was very useful, as the IT team were under extreme pressure working to an aggressive timescale.

“We have had ongoing support from Anaplan and have not had to increase our central overhead as a result. I would encourage other businesses to take the Anaplan Challenge, which got rid of all of the smoke and mirrors that we had been seeing in other pre-sales pitches.”

Feldon says that, previously, users had problems with the Excel suite failing to recalculate due to vast amounts of data and the complexity of the model, even after a 10-15 minute wait.