Telstra's finance division tackles bad data

Telstra has embarked on a fresh data quality initiative in a bid to cut the number of erroneous site visits and returned bills caused by bad data.

The program, which kicked off last year, saw finance take over some responsibility for Telstra’s data governance while IT set up a Data Management Centre of Excellence to look after critical master data sets.

According to Telstra’s chief architect of information and corporate architecture Mark Kortink, it was the latest of several data governance attempts by the telco.

“Data governance in Telstra is run by finance,” Kortink told the CIO Strategy Summit in Melbourne last week, adding that CIO Patrick Eltridge was an “active participant” in the program.

“IT plays a role in data quality fix projects and some of the data stewardship stuff around key data.”

Addressing chief information officers from Australian banks, airlines and government agencies, Kortink described IT as the “dams, pipes, pumps and tanks” of data.

Data and IT were two different resources, he said, with the latter classified as a fixed asset on company balance sheets and the former more difficult to value.

Yet data was the third most important resource in a company, he said, after money and people.

"Companies are pretty good at IT; we’re not quite as good at managing data,” Kortink said.

“A lot of companies have had a shot at proper data management but they never seem to really take the way that accountants have taken control of financial management or HR has taken control of people management.”

Echoing the comments made by members of the Institute of Actuaries Australia last September, Kortink said poor data quality was to blame for about 20 percent of errors across the industry.

For Telstra, errors may manifest as "ineffective truck rolls and returned bills". Kortink declined to disclose the cost and frequency of those errors, but said there was a business case for projects that addressed them.

“How do you justify a data quality project? Normally you find a process that’s falling over and quite often, data quality will be a factor,” he said.

“Over the years, we’ve looked at ineffective truck rolls, returned bills, things like that ... There’s often two or three data quality effects causing them.”

Prior to Telstra’s finance-led data governance program, data was managed by a number of Chief Data Officers, the last of whom was appointed in 2004.

Kortink managed aspects of data governance during Telstra’s IT transformation project, which was unveiled in 2005 and expected to rationalise legacy systems and deliver a new enterprise data warehouse for storing customer information.

He said the new enterprise data warehouse could absorb three of Telstra’s seven existing data warehouses within the next two years.

The telco was also two years into a “progressive exercise” to consolidate legacy databases into a master data management (MDM) system and integrate it with relevant systems.

Kortink told the conference that Telstra had about 240 systems that contained address data, and “about the same number of systems” with product data.

Although “traditional data warehouse architecture” called for a single warehouse and single business intelligence (BI) tool that all employees used to analyse data, Kortink said that model could be unrealistic for large organisations.

“The reality is if you try that approach, you’ll get people using other things [business intelligence tools] anyhow,” he said.

“If you’ve got a legacy environment like Telstra where we’ve got seven big data warehouses … we’re always going to have fairly chunky data warehouses and at some point it becomes a not very good business case to get rid of them.

“We’re working on the ideal template of one data warehouse and one BI tool. In reality, [we’ll have] multiple data warehouses and multiple BI tools.”

Kortink said the new, finance-led data governance program was “very effective” but noted that data governance needed to become a more permanent fixture within the organisation.

“It’s not a project, it’s a process,” he said. “You don’t say let’s run an HR project to manage our people or let’s run a finance project to manage our money.”

He urged conference attendees to take control of data governance, noting that if CIOs did not step up to the task, “no one else will, in most organisations”.