Submitted by Jason Clevenger
Summary
On October 15th, 2020 the university’s new financial aid application, PowerFAIDS, went live for the Danforth campus. The planning for this had started barely a year before and went forward despite the shrinking IT resources due to MyDay efforts. Phase 1 of the project came in on time overcoming the disruptions from Covid-19, allowing applicants to the university to apply for financial aid. This was made possible by a close collaboration between the business and IT and the efforts from many different IT groups. The project also provided an opportunity for Enterprise Applications to deploy their new integration technology – DRA (Data Relay Applications) — into a full blow production environment.
In the Fall of 2019 Student Financial Services (SFS) came to Enterprise Applications (EA) with a request to consider replacing STAR — their legacy financial aid system — with PowerFAIDS (PF), the industry leading solution. STAR had been developed a decade before and was showing its age. It was taking north of 1,500 hours a year of programming to comply with changes in federal financial aid regulations. There had been a number of small but embarrassing outages in recent years, raising questions about reliability. SFS was looking to change how it awarded financial aid and didn’t think STAR could handle that without a substantial rebuild. The kicker was that if we decided to go forward with implementing PF, we would have to be ready by October of 2020. Financial aid is a critical part of the undergraduate admission process. The new system would have to be able to accept and act on requests from tens of thousands of high-schoolers applying to the university in the fall. We could not be late.
But by Fall 2019 the trickle of MyDay implementation work streams that started in January had swollen into a flood, swallowing up more and more IT resources. We were still operating on the assumption of a July 2020 go-live (the formal decision to delay for one year did not come until the end of January 2020). The easy thing to do would be to punt on the request for PowerFAIDs and ask SFS to come back next year. We have a good working relationship with them and they would have accepted that answer. But they were not making the request lightly. We owed them more than just a brush off. So, we set out to answer SFS’s question – what would it take to implement PowerFAIDS by October 2020? In doing so, we began to find a way to say yes, we can do that.
Working with leadership we put together a three-phase project outline. Phase 1 would run from March to October 2020 with a focus on getting the infrastructure built out, the initial integrations developed, and the configurations that are needed to begin accepting applications for financial aid for the following academic year (2021-2022). Phase 2 and 3 would follow on in 6 month increments to finish the build out of the integrations and functionality plus retire the legacy systems. STAR would continue to run in 2020-2021 to support the one more year of financial aid for current students.
Resourcing the project required a departure from past EA practice, where a small team often did the vast majority of the work. Getting PF Phase 1 up and running by October would need contributions from many parts of the university. The business (Student Financial Services and Undergraduate Admissions) working with consultants would do the configurations in PF and the changes needed in Slate (our Undergraduate Admissions system). PowerFAIDS is a throwback – a classic on-premise client-server architecture. So, Enterprise Engineering and End User Services got involved in setting up servers and creating a process to install the application onto user machines. VAM (Vendor Application Management) took on the central software installations and the network and file configurations. A technical team from EA developed the integrations (seven of them in Phase 1). We adopted the new Data Relay Application (DRA) technology being spun-up for MyDay to create the PF integration pipelines. The EA Technical Architecture team was a key partner in pushing that technology forward and getting it ready for real-world production ahead of the MyDay schedule. The EA Prod-Ops team came on board towards the end to help build out the day-to-day technical management pieces. Along the way were engagements with SMEs on such esoteric topics as Dummy SSN Assignment. PowerFAIDS also became the first piece of the Workday Student program that will deliver a new SIS (Student Information System) ecosystem around 2024/2025. Project management, business analysis, and program oversight came out of that organization.
The PF ecosystem footprint is substantial. The core of the application is a central database and applications installed on about 50 user machines. PF also has web components for read-only admin users and for a student financial aid portal, so a pair of web-servers were needed. Integrations and batch processes are run through an application known as PF Scheduler, which required its own server. File storage is needed to hold student and parent documents. Phase 1 integrations were from SIS (student info), Slate (applicant info), the Department of Education, and the College Board. A new applicant financial aid portal was constructed in Slate which required integrations from PowerFAIDS to Slate.
PowerFAIDS highlighted a number of approaches and methodologies that we have been moving towards in EA projects:
- Preference for vendor products over built-in-house
- Close partnership with the business (configurations) and IT (infrastructure and integrations) in delivering the implementation.
- Use of the DRA technology for integrations along with help from Technical Architecture.
- Multiple technical teams implementing major slices of the project
- The role of the Technical Lead as a conduit from business and architecture needs to the technical teams.
- Active participation of Prod-ops for go-live and transition to operations.
The project also had its share of lessons learned:
- Complex systems can require complex data – the SIS integration to PF alone consists of 120+ data elements and required the construction of a new staging environment to properly marshal it.
- “Greenfield” development is very different than building onto an established ecosystem:
- You need to track changes created in the test environment and then make sure that they are created in the new prod environment – redoing permissions was an ongoing issue.
- Use a smoke test process to confirm the prod ecosystem is properly configured.
- Complete end-to-end system/integration functional testing before you start user acceptance (business) testing. This avoids wasting end user time and avoid embarrassing mistakes that would be easy to fix.
- Distributed work teams can mean a loss of control over implementation details
It was just under a year from the first discussion to the PowerFAIDS Phase 1 go-live. It was a year of distraction with the restructurings of MyDay and the disruptions thrown at us by Covid. But on October 15, 2020 PF started up with about 2,000 applicants (which will grow to almost 30,000 by January 2021), who began applying for financial aid through a new portal in Slate. The first financial aid awards began flowing out in December for Early Decision applicants. About 17,000 current students are also in PF and are applying for aid for next year through their new portal. All told over $170 million dollars will be awarded, disbursed, and reported on for this group by the end of the 2021-2022 academic year. The cycle starts all over again next fall.
“It takes a village” has become a bit of a stale trope. In the highly decentralized world that is WashU, it makes more sense to say it took several villages to make PowerFAIDS happen. The close collaboration between the technical integration team and the business configuration team produced a remarkable set of complex financial aid products in about five months. As other teams rotated into the mix the first question was “What’s PowerFAIDs?” and the second was “How can we help?” The project that started out with everyone asking “Can we do this later?” came in on time thanks to the efforts of the many villages of the university.