Back in March 2025, we rolled out Vertical Face Imagery (VFI) to everyone. Adoption grew faster than we expected, which was exciting, but it also put some strain on our processing pipeline. That meant a few of our customers experienced slower turnaround times. We knew this wasn’t ideal, and we took it seriously.
Over the past six months, our engineering team has been hard at work making Propeller even faster, more reliable, and easier to count on. The goal was simple: deliver the data you need, when you need it, without ever compromising accuracy.
Faster turnaround times, across the board
Since August 2025, every fully-automated survey has hit our Service Level Agreement (SLA).
Turnaround times have improved across all survey sizes for these surveys:
- Small surveys (0–20 GCPs): now 2h 45m, 30% faster
- Medium surveys (20–60 GCPs): now 6h 12m, 31% faster
- Large surveys (60+ GCPs): now 11h, 31% faster
VFI data turnaround times (the feature that kicked off this whole journey) saw even bigger improvements:
- Small VFI surveys: 73% faster
- Medium VFI surveys: 70% faster
- Large VFI surveys: 79% faster
These gains hold even as the volume of VFI datasets has more than doubled since last year.
How we made it happen
Our team focused on three key areas:
- Reworking the processing pipeline to handle higher data volumes without slowing down
- Adding more dedicated compute for large and VFI-heavy surveys
- Streamlining automated quality checks so accuracy and reliability remain high while processing speeds increase
These upgrades weren’t just about faster numbers. They make the Propeller experience more consistent and dependable for everyone.
