Predictive analytics pays off for Australia Post

Government agency uses IBM’s SPSS software and TM1 software to improve agency payment forecasting.

Manual cash flow forecasts are a thing of the past for Australia Post with the implementation of two software packages which crunch data to produce accurate daily, monthly and annual forecasting reports.

Speaking at IBM’s Information On Demand conference in Las Vegas, Australia Post business analytics manager Armand Mizan told delegates that its manual forecasting took weeks to produce and was not always accurate.

In addition, Australia Post staff had little visibility around movement of agency cash flows. The company works as an agent on behalf of companies such as the Australian Tax Office (ATO), telecommunications and utility providers. It receives payments for these companies at 4500 retail outlets and online. These payments go through an Australia Post bank account and are re-distributed back to the agency.

Lack of data scientists could derail big data projects: IBM

In pictures: IBM Information On Demand day one

According to Mizan, Australia Post processes $50 billion worth of agency payments per year.

To help the department overcome the problem of manual cash flow forecasting, it decided to implement IBM’s SPSS predictive analytics software and Cognos TM1 enterprise planning software.

“We had a monthly cash flow model which was based on actual figures,” he said.

“This would then populate the daily cash flow through a translation. From that, the agency forecast model is produced through SPSS and the results sent back out into TM1.”

In order to check and determine the accuracy of the information in SPSS, Australia Post took historical data from August to October 2011 and put it through the software to produce a forecast.

“That forecast was compared to the actual data from 2011. We achieved an accuracy level of between 95 and 98 per cent using SPSS,” Mizan said.

Business outcomes

Australia Post now has fully integrated actuals and forecasted monthly cash flows for the next four years. The monthly and daily cash flow models are also aligned.

There is also greater visibility of agency financial data, something which Mizan said the company did not have before.

Technical difficulties

However, Mizan admitted that the SPSS and TM1 project was not without its teething problems.

For example, with agency settlement and transaction dates, the data that it took out of the general ledger reflected the transaction date. However, this date was different from when the cash settlement went through.

“When we are doing cash flow forecast, we needed that settlement date,” he said. “In order to overcome that we applied lags to transactions to arrive at a settlement date.”

The second challenge was the impact of state and territory based public holidays and special events.

The SPSS software had to take this into account so a calendar of holidays was entered into the forecasting model.

Future uses

The company is currently exploring the use of SPSS for daily revenue forecasting by individual retail outlet and fraud prevention/detection.

“We are the largest retailer in Australia so the risk of fraud is significant. Using TM1 and SPSS, we are looking at detecting credit card fraud in the 4,500 retail outlets,” he said.

In addition, Australia Post sales and marketing staff plan to use TM1 for customer analytics in order to understand who their customers are, their buying habits and to identity customer churn.

Hamish Barwick travelled to Information on Demand in Las Vegas as a guest of IBM

Follow Hamish Barwick on Twitter: @HamishBarwick Follow CIO Australia on Twitter: @CIO_Australia

Join the newsletter!

Or

Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

Tags business intelligencecognosaustralia postSPSSbusiness analyticsInformation on DemandforecastingTM1

Show Comments
[]