Winemaker turns to ERP, predictive modelling

Chief information officer of the Adelaide-based company, Ryan Klose, told ZDNet Australia the business had gone live within the last two to three weeks with the first stage of an JD Edwards EnterpriseOne financials software deployment.

Klose said the deployment covered packaging, work-order management, scheduling, costing and variance accounting at the McGuigan packaging plant in north-western Victoria.

It had allowed the winemaker to create a separate packaging business unit in its own right to complement its existing export sales, domestic sales, wineries and vineyards divisions. Klose said the specialised software being replaced by EnterpriseOne -- EzySystems' EnterpriseEzy accounting module -- was not as flexible in allowing McGuigan to break out the types of information required to run the operation as a separate unit.

The stage two deployment -- scheduled in July or August this year -- would see accounting and accounts payable modules implemented across the entire group.

As part of the project, McGuigan is shifting from an SQL Server database to an Oracle 10g database which will be embedded under the EnterpriseOne software. The new database will, according to Klose, support the mission critical environment for the group as a whole. "Oracle lends itself to a lot better order tracking, growth and scalability than the SQL database [did]," Klose said.

McGuigan expects to see a return on investment on the EnterpriseOne implementations within 12 months, Klose said.

McGuigan is also developing in-house a solution designed to fulfil the winemakers' need for better visibility and a real-time view of the AU$200 million worth of stock it has on the ground at any one time. This stock is, Klose said, moved constantly around between McGuigan's facilities and customers, sometimes outstripping the current systems' ability to keep track.

Stage one of the project is scheduled to go live next month, while stage two, which drills down into detailed issues such as traceability of wine by vintage, is expected to go live by March next year.

Klose said the new systems were required to support a shift in winemaking whereby smaller volumes of wine are held by the wineries, meaning less of a margin for error should there be a problem with an order.

McGuigan is also looking at enhancing its predictive capabilities during harvest in a project designed to maximise the harvest of grapes through the winery crushing and winemaking process.

The company is collaborating with Adelaide University's Professor Zbigniew Michalewicz -- a specialist in heuristics -- to build a software model which takes into account harvest methods, maturity analysis and wine facility capabilities to allow McGuigan to eliminate unnecessary costs and boosts return.

Some simulations would be run on the 2007 harvest, Klose said, with the aim being to use a live model to control the harvest in 2008.