Story 3HF Accenture wins $102M contract to implement Australian Child Support system in SAP

Accenture wins $102M contract to implement Australian Child Support system in SAP

by
in code on (#3HF)
Accenture, a firm well known for its spectacular failures has won the AU$102m contract to replace the ageing Child Support system with a SAP system for the Australian Department of Human Services Child Support programme . Accenture will have five years to complete this herculean task. CUBA, the current CSA system for processing child support payments, dispenses $3.2 billion in payments in more than 1 billion transactions annually for 1.2 million children and 1.5 million parents per year.

Is this the deal of a lifetime or an impending disaster? And is SAP the right system to build? What would you use to support a transactional system of this magnitude?
Reply 3 comments

JIRA (Score: 1)

by fishybell@pipedot.org on 2014-04-04 19:03 (#ZK)

I'd keep it in-house, as in, in Australia. While JIRA itself is not necessarily built for this type of task, it's extremely flexible and Atlassian could, for a cool $102 million, definitely expand it or customize it to meet the expectations of the customers. Having the technology local would probably also reduce the costs of support later.

COBOL and DB2 (Score: 0)

by Anonymous Coward on 2014-04-05 02:08 (#ZQ)

List all solutions that can handle that handle millions of transactions per year.
Definitely not java tech. Billions of transactions per year. Definitely not java.
Not Microsoft SQLServer with .net. Compared to DB2 it is not even in the running.
Cobol could have died by now but no new batch processing technology has come out.
The SAP programming language devised in the 70s is similar to COBOL functionally.
Perhaps they are thinking that SAP can be the new COBOL for the next few decades.

Impending disaster (Score: 0)

by Anonymous Coward on 2014-04-05 02:55 (#ZR)

I predict that in the 4th year Accenture will declare that the project is 'too large' with 'great problems' that the 'scope was not well defined' with an outcome in the 5th year of a system being delivered that is worse than the old system that does not implement everything the old system had runs at a greater cost and required Accenture to stick around for two decades fixing and maintaining it. Did we learn nothing from the Medicare BI project where Accenture did not deliver? Did we learn nothing from the tax office where Accenture is in permanent residence?

History will repeat itself. From the linked article, here is what happened last time Accenture engaged in a project to moved a COBOL system to SAP...


The California State Controller's Office (SCO) is currently running on an old COBOL-based payroll system that dates back to the 1970s. The SCO began an initiative in 2006 to update this system, with initial estimates targeting full implementation by 2009. State Controller John Chiang said that the systems needed to carry out Schwarzenegger's minimum wage plan would not be available for six months. That was last summer.

Just this January, the SCO announced that it was cancelling its contract with the consulting company in charge of the project and had not estimated when it would hire another firm to carry on. That was $25 million into an estimated $69 million project.

Earlier this month, SAP co-CEO Leo Apotheker angrily denied there were problems with SAP's software, and blamed consulting firms like IBM (IBM) and Accenture (ACN) for sending people who knew nothing about the software to clients as experts on SAP. Leo also has said SAP's new cloud-like package, SAP Business Suite 7, should be easier to implement.

Plenty of blame to go around, we think. At least in the California bomb, the consulting firm involved was BearingPoint, which yesterday filed for bankruptcy. Accenture has already moved to acquire part of BearingPoint's operations.


The question here is: If the Australian government is in such dire straits that it can not afford a pay rise for its public servants then why is it risking $100m with a company well known for failing this type of project?