We recently made a major-version jump in ProPresenter, moving from ProPresenter 21.4 to ProPresenter 22. Most of the transition went exactly the way you hope a software upgrade will go. Our presentations opened, the software was stable, and we started making our normal changes for the upcoming Sunday without much drama.
Then we found one very specific part of our pre-service workflow that no longer behaved the way it had before. It involved something incredibly simple: a slide that was supposed to sit on screen for several minutes and then automatically advance. That little timer sent us much farther down the troubleshooting rabbit hole than I expected.
The workflow wasn’t complicated
Our pre-service sequence is pretty straightforward. An operator triggers the first slide, that content stays up for roughly four minutes, and then ProPresenter should automatically advance to the next slide, which contains about a one-minute video. Once the video finishes, a third slide handles the cleanup before we move into the first song of the service.
The point of the automation is simply to remove a few predictable tasks from the operator during one of the busiest parts of Sunday morning. They start the sequence, and ProPresenter takes care of the timing from there.
After upgrading to ProPresenter 22, that automation stopped behaving correctly. The strange part was that Go To Next wasn’t completely broken. A five-second timer worked, but longer timers did not.
That made troubleshooting much more interesting because the feature itself appeared to function. The problem only showed up when we tried to use it the way we actually use it on Sunday.
A five-second test didn’t tell us much
One of the first lessons from the whole experience was how easy it is to accidentally test the wrong thing. When something breaks, the natural response is to simplify the test until you can prove the feature works. Set Go To Next to five seconds, watch the slide advance, and assume everything is fine.
But that only proves that a five-second Go To Next works. It doesn’t prove that the actual four-minute workflow works.
We started changing durations and testing different combinations. At one point we set the timer to around 12 seconds just to shorten the troubleshooting cycle. The timer indicator would visually progress as expected, reach the end, and then the slide still wouldn’t advance.
That was when it became clear this wasn’t simply a missed setting. Something else in the workflow was interfering with it.
Then the announcement loop got involved
Our pre-service content also interacts with an announcement loop, which added another variable to the problem. We eventually found that when the announcement loop advanced, it appeared to interfere with the Go To Next timer on the slide. That started explaining why a very short test might work while the real-world version failed.
We tried restructuring the content, converted media into a slide object, changed timing, and removed variables one at a time. The timer could still visibly count down without producing the transition we expected.
At that point, the question changed from “What setting did we miss?” to “Do we actually want to keep building this workflow around this particular feature?”
That turned out to be a much more useful question.
We considered rolling back
The obvious answer when a new software version breaks part of your workflow is to go back to the version that worked. So we tested that.
We opened ProPresenter 21.4 again, thinking we might simply stay there until the issue was resolved in version 22. The problem was that by then we had already been working in the new version. Our current playlist, new presentations, theme changes, and other work had been created or modified after the upgrade.
Going back to the older version did not feel like simply returning to yesterday’s software. It also meant returning to an earlier state of our workspace.
That’s where “just roll back” stopped sounding quite as simple.
There are absolutely better ways to plan for migrations, exports, backups, and version changes before a major upgrade. But in the moment, we were facing a practical Sunday decision: do we unwind work we’ve already completed because one automation is broken, or do we stay on the new version and find another way to accomplish the same thing?
We chose to stay on ProPresenter 22.
One broken workflow wasn’t enough reason to undo the whole upgrade
That decision is important because I don’t want this to sound like ProPresenter 22 was unusable. It wasn’t. Most of what we needed was working normally.
We had one very specific workflow that wasn’t behaving correctly, and the workaround was significantly easier than rebuilding our current Sunday environment in the older version. So instead of treating rollback as the automatic answer, we started looking at the workflow itself.
That led us toward ProPresenter’s Timeline. Rather than depending on one long Go To Next timer, the idea became to build the entire pre-service sequence around a known timeline. The operator can trigger the beginning of the sequence, the announcement content can run for the intended amount of time, the video can fire at the correct point, and the final cleanup can happen before transitioning toward song one.
We’re essentially trying to solve the same problem with a tool that is better suited for coordinating a sequence of timed events. And honestly, that may ultimately give us a better workflow than the one we had before the upgrade.
Sometimes a bug exposes a fragile workflow
The Go To Next issue was frustrating, but it also exposed how much our pre-service sequence depended on one timer behaving exactly the way we expected. When that timer stopped working, the entire automation stopped working.
That doesn’t necessarily mean the old workflow was bad. It worked for us for a long time. But failures have a way of showing you where the fragile parts of a system are.
If one timer, macro, scene, plugin, network connection, or piece of software can completely derail part of Sunday, the question isn’t only how to fix that component. It’s also worth asking what happens the next time it fails.
That’s especially important with automation. I love automating predictable tasks because if the computer already knows that a video should start four minutes after a sequence begins, I don’t necessarily need a volunteer watching a stopwatch.
At the same time, automation should reduce the operator’s workload without making the entire service dependent on an invisible chain of events that nobody understands. The best automated workflow still needs an obvious manual fallback. If something doesn’t fire, the operator should be able to recognize what happened and move the service forward without needing to understand everything happening behind the scenes.
Major upgrades deserve real-world testing
There’s another lesson here that sounds obvious until you get caught by it: major software upgrades need to be tested against your actual Sunday workflows.
Opening the program is not testing. Loading a presentation is not testing. Clicking through a few slides is not testing.
If you use ProPresenter for timers, announcement loops, macros, stage displays, video playback, MIDI, lighting triggers, audio routing, calendar events, or any other automation, those things need to be tested the way they will actually run.
Our five-second Go To Next test worked. Our real workflow didn’t.
That difference matters.
A proper test should be intentionally boring. Run the whole pre-service sequence. Let the four-minute timer actually run for four minutes. Let the announcement loop advance. Play the full video. Confirm the cleanup happens. Confirm the next presentation loads correctly.
It takes longer, but that is the behavior you are actually trusting on Sunday.
“Just roll back” isn’t always the safest option
I also came away from this with a different perspective on rollbacks. Keeping a known-good previous version available is smart. Having backups is smart. Knowing how to restore a workspace is smart.
But rolling back software is not automatically safer simply because the older version worked last week. By the time you discover a problem, your environment may already have changed. New presentations have been created, themes have been modified, playlists have been built, and volunteers may already be working inside the new version.
At that point, the rollback itself introduces operational risk.
That means the decision should be based on impact. If a major upgrade breaks critical video output, crashes repeatedly, corrupts presentations, or makes the system unreliable, that may absolutely justify abandoning it immediately.
If one automation stops working but the rest of the system is stable, the better decision may be to work around the problem while you investigate it.
That’s what made sense for us this time.
Build Sunday so it can fail gracefully
We stayed on ProPresenter 22, not because the problem didn’t matter and not because I think production teams should simply accept broken workflows after an upgrade. We stayed because the rest of the system was working, and one broken automation was easier to redesign than unwinding everything we had already built in the new version.
More importantly, the experience forced us to look at that automation differently. Maybe the goal isn’t to build a Sunday where every transition happens automatically and nothing ever requires human intervention. Maybe the better goal is to build a Sunday where automation handles the repetitive work, volunteers understand what the system is supposed to do, and there is an obvious path forward when something doesn’t happen.
Software will fail. Timers will misbehave. Updates will change things you didn’t expect them to change. The best production system isn’t the one where nothing can ever break.
It’s the one where a broken timer doesn’t break Sunday.


