Three years ago, the Alberta Pensions Services Corp. was facing a bitter fact. Though one of our key corporate goals is to begin paying benefits within 30 days of plan members completing their retirement application, we were reaching this goal for barely one of every four retirees.
This was a significant issue. APS administers the pension contributions and benefits of more than 365,000 current or former public sector employees in the province. Thousands of Alberta’s public employees retire each year and depend on the timely payment of their benefits for post-work income.
There was just one problem. We didn’t know why this was happening.
“We couldn’t reliably tell you how many files were even in the building, let alone the best order in which to process them,” says Waleem Alausa, the organization’s manager of business intelligence and analytics.
APS established its analytics team in 2016, the same year it undertook a transformational technological upgrade to the pension administration system. The new system facilitated better data collection, allowing Alausa’s team to identify and analyze problems while it processed pension transactions.
Using the capabilities of this system, the team developed a new tool called Retirement Radar, which helped visualize the number of retirement files pending at any given time, as well as everything needed to get them paid within 30 days. Armed with the new tool and other improvements, APS was able to turn its 27 per cent pension inception rate into a successful 95 per cent by 2017.
This is probably the most dramatic example of how we’re using data analytics at APS to improve processes and to predict trends. But as we expand this approach into other aspects of our business, it’s driving service improvements in other areas and the full potential is just starting to be tapped.
“APS has been reporting for quite a while with the available data, but I wouldn’t call it business intelligence prior to 2016. We were just using data at its crudest level for troubleshooting purposes,” says Alausa. “As we’ve become more data-aware, our role has become even more transformative.”
This July, our president and chief executive officer Darwin Bozek announced the next step for data analytics at APS. The challenge was to incorporate disparate sources and types of data to measure performance across the organization, from call centre service levels to buyback transactions.
Enter APS Dash, a real-time dashboard-style display of the organization’s performance against key measures. It aggregates data in much the same way as the Retirement Radar, except on a larger, corporate-wide scale. The APS Dash is displayed on monitors on all three floors of the APS building and operational managers have access to it on their desktops.
The value comes in flagging where APS is doing well against its performance indicators and where work is needed. Managers are now able to see immediately what areas need attention and can investigate whether it’s simply a matter of temporarily higher service volumes, or whether there’s an ongoing, systemic concern that needs closer scrutiny.
“It’s a cliché, but there’s a lot of truth to the saying, ‘What gets measured, gets managed,’” says Bozek.
It took careful planning to roll out the monitors throughout APS. Bozek says he was wary of creating competition or blame between different business areas. Alausa’s team held orientation sessions for managers before the tool was launched, and organized lunch-and-learn sessions for employees who wanted to better understand the numbers on the dashboard.
“I compare it to the dashboard display on your car,” says Bozek. “If the gas gauge is low, you don’t blame the car, you find a station and fill up.”
Bozek was also intentional in explaining to employees that they all have a stake in APS Dash and a responsibility for the results. He encouraged business teams to hold regular discussions about the indicators and how they can contribute to improvements.
“For example, our information technology employees don’t interact with pension plan members or process retirement files,” says Bozek. “But they keep all our other tools running so we can do those things without having to break out an adding machine or abacus. They also create new tools, like our online retirement application PensionEase.”
While APS Dash is live, work on it continues. “We’re developing a three-year analytics strategy,” says Alausa. “It will shape us into a mature analytics centre of excellence that powers the organization forward.”
As well, Alausa and his team are already planning their next act: predictive modelling and artificial intelligence.
“We process about 25,000 terminations a year. Some result in payouts, while many members leave their funds in the pension plan. But we currently treat all these transactions the same,” he says. “In the future, we may be able to predict which files are going to go which way, and process them accordingly at those preliminary stages to provide more efficient and timely customer service.”
Using data and artificial intelligence sounds like science fiction, and it’s certainly still in its infancy for the pension industry. But the results are starting to speak for themselves and APS is just getting started.
Mike Berezowsky is executive director of corporate communications at the Alberta Pensions Services Corp.