The Observer


APS Predicts a Bright Future for Predictive Data Analytics

by Max Maudie, Alberta Pensions Services Corporation

Q: Why did the chicken cross the road?
A: The answer is trivial and is left as an exercise for the reader.

Unless you’re a data scientist or mathematician, you’re not chuckling at this joke, but the tactical advantage offered by data analytics is no laughing matter. Analytics plays a central role at Alberta Pensions Services Corporation (APS), past, present and future.

In 2016, for instance, APS’ new analytics team used data to drive the organization’s pension inception rate—that’s the rate of new retirees who collect their first payment within 30 days—from 27 per cent to 95 per cent in a year.

 “That was when APS really harnessed the power of data analysis, of business intelligence,” says Waleem Alausa, APS’ manager of Business Intelligence and Analytics (BIA). “It set us on our way.”

That process gave rise to APS Dash, a real-time display of corporate-wide performance against key measures. APS Dash is featured on monitors on all three floors of the APS building, and operational managers have desktop access to the dashboard.

Currently, Alausa and his team are targeting termination files with an eye to helping improve turnaround times and client service. APS processes about 23,000 terminations a year; that’s 37 per cent of all files. While some terminating members retire, leave their funds with their plan, or transfer to a different plan, other members want their money—and that’s a priority.

“They have a fair expectation to get that money soon for whatever they need it for,” says Alausa, who holds a PhD in economics from the University of Alberta.

Once a termination package is received, the member has one year to make a decision. Traditionally, half of the members will take a payout; this is where predictive data analytics enters the scene. How to predict which members will take a payout, thereby giving the Terminations Team a head start on those files?

Alausa and his team developed a predictive algorithm through the automated analysis of 8,000 completed termination files. Patterns were analyzed by considering factors such as:

  • Demographics
  • Salary
  • Employer
  • Pension plan
  • Years of service
  • Vesting status
  • Number of times a member has previously left a plan

The algorithm was shown to be 96 per cent accurate when tested on another 4,000 completed files. With predictions in hand, the Terminations Team may soon be able to start work on the likely payout files with great confidence. And every night, the algorithm searches the database to make new termination predictions. The goal is 99% accuracy. This terminations process enhancement may also offer significant cost savings.

You’d be forgiven for thinking that setting up this process demanded a lot of computer horsepower, but it didn’t. In fact, Alausa trained the predictive algorithm with his rather unimpressive desktop computer.

BIA is also working to predict seasonal demands in the Member Services Centre (MSC). Ordinarily, temporary staff are hired to tackle seasonal spikes. It takes a few months to train the new staff, though, and once the seasonal increase has been handled, most of the temp staff are let go. It can be a time consuming and costly undertaking.

Sure, December and January is the retirement season, and July and August are all about terminations, but the subtleties of scheduling call for more precision than that. Instead, an Integrated Queue Management System will anticipate spikes and provide predictive analytics. A dedicated mobile team can spring into action to help take care of what may well have otherwise become a backlog—staff can be proactive, not reactive.

Brian Leeson, a manager in the MSC, says integrated queue management is a key piece of their strategic plan. “We’ll be certain that we have enough appropriately skilled people available to serve our clients at the right times, no matter what,” he said.

Proper staffing levels and quick, accurate payments are critical for APS, says the organization’s president and CEO, Darwin Bozek. “That delivers to the pension plan boards and the plan members the service they deserve, with the quality they deserve.”

The work of Alausa and his team continues, with an eye to becoming an “analytics centre of excellence.” The goal is to make it easier for staff to access and understand relevant data, and develop insights of their own. That engages and empowers a workforce, notes Alausa, who tags his emails with a quote widely attributed to renowned statistician W. Edwards Deming.

“In God we trust, all others must bring data.”


 Max Maudie
 Content Strategist, Corporate Communications
 Alberta Pensions Services Corporation