Challenges
Manual data entry
Disconnected systems, including Slate
Uncertain data accuracy
Solution
- Parchment Receive Premium With Slate Integration
- Parchment Data Automation
Results
- An efficient solution for processing 50,000 transcripts
- Fully integrated workflows with Slate
- Quality OCR data extraction, with 95%+ accuracy
The situation
At the height of the COVID pandemic, the University of Texas at Austin went test-optional for first-year students. But without test scores, the admissions team needed another way to evaluate applicants. They settled on a predictive GPA built from transcript data.
That put the full weight of the admissions decision on the transcript. A temporary team of over 100 people entered every transcript by hand.
For Matthew Barsalou, senior information technology manager, the bigger problem was where all those hours went. “Before, we were manually pulling up the transcript and had to do quite a bit of training and quality assurance. It was taxing and time-consuming.” The team reviewed every transcript that arrived, including ones from students who might never finish applying. “We needed to focus our time and attention on students who actually completed their application in Slate,” Barsalou says.
This wasn’t sustainable, so the admissions team looked for an OCR solution that could take over manual transcript data entry.
The old challenges
Staff keyed in every transcript by hand, which took a temporary team of more than 100 people.
Transcript data had to reach Slate and other admissions systems with no automated connection.
Manual entry required constant training and quality checks, and the team still worried about accuracy.
Insights
Higher ed has no shortage of OCR vendors.
“The first vendor wasn’t successful,” explains Allison Radke, assistant director of admissions. “The communication and follow-through weren’t what we needed or expected. We decided to continue to pursue OCR with another partner. We had a previous relationship with Parchment, and we knew we could trust them. We were impressed by the demo and liked that their solution could integrate with our other Admissions systems, including Slate.”
The second time around, the team picked a partner it already knew it could count on.
Data Automation has been a game-changer for us. If we didn’t have a tool like this, we would be in a lot of pain.
Allison Radke
Assistant Director of Admissions, The University of Texas at Austin
Building the Slate connection
To scale, UT Austin needed Data Automation to work inside Slate, its application platform. Tom Picklesimer, Slate solutions manager, laid out the workflow:
Slate exports a query with zipped transcript PDFs and an indexed CSV to the Slate SFTP.
Parchment Data Automation picks up the files and extracts the course data.
Parchment sends the results back as a CSV over SFTP.
The team drops the data into the Profile Courses tab in Slate.
The team expected a long implementation. “Within six weeks, we were working with test data in the system before rolling out fully,” says Radke.
The support stood out as much as the software. “Everyone has been phenomenal,” says Picklesimer. “If there were any challenges, they were resolved in very short order.”
Radke agrees. “Customer service has been exceptional. The team has been extremely responsive and patient. They made sure everything was working according to our desires. It was night and day compared to our previous vendor. They took the time to explain what was possible and what wasn’t.”
Accuracy across the board
Transcript processing costs dropped by half. But for most OCR buyers, the first question is accuracy. If staff have to double-check every data set, they lose the time they saved.
“It’s been extremely accurate. We can really rely on this data. It’s been about 95% accurate,” says Picklesimer. “If there are errors, they’re readily visible. We’re receiving around 50,000 transcripts, so when you’re talking about that kind of scale, it’s nothing short of remarkable.”
The need for a huge temporary team disappeared. “We’ve been able to move from a team of 100 temporary staff to a team of around 30,” says Radke. “If we didn’t have a tool like this, we would be in a lot of pain. We have a consistent process that can scale.”
We’re receiving around 50,000 transcripts, so when you’re talking about that kind of scale, it’s nothing short of remarkable.
Tom Picklesimer
Slate Solutions Manager
Big gains with Parchment
Transcript processing costs dropped 50%.
Staff need 70% less time to process transcripts.
The data entry team went from more than 100 temporary staff to about 30.
Data Automation extracts transcript data at about 95% accuracy across roughly 50,000 transcripts.
Transcript data flows straight into the Profile Courses tab in Slate.
Leave behind manual data entry for automation that works
