How we measure
Measured or projected · never mixed
This page is the arithmetic.
Every automation vendor claims ROI. Almost none will show you the arithmetic. This page is the arithmetic: how Ace Work counts time and money saved, what we count conservatively, and what we refuse to count at all.
The rule everything follows
A number is either measured or it's projected, and the two never get mixed. Projected means our estimate before a workflow runs, discounted on purpose. Measured means counted from real runs against the baseline. When a workflow starts running, the measured number replaces the projection. It doesn't get averaged with it.
When a workflow starts running, the measured number replaces the projection.
Step 01 · The baseline
Priced as it is today, before anything gets fixed.
Before anything gets fixed, the workflow gets priced as it is today. Saor interviews the people who actually do the work: who does it, how often, how long a cycle takes, and where the time goes inside it. The baseline comes from the people doing the work, not from a manager's guess or an industry benchmark, and it's written down before the fix is built. A baseline captured after the fact is a story, not a measurement.
The method's limit, stated
We'll be straight about the method's limit too: the baseline is what your team reports, structured and cross-examined, not a stopwatch. That's why everything downstream of it is discounted, floored, and capped, which is the rest of this page.
Step 02 · Projections, discounted on purpose
Deliberately pessimistic, before a single run.
Before a workflow runs, the roadmap shows what we project it will save. That projection is deliberately pessimistic:
Payback math adds a 20% buffer on implementation cost on top.
85%
Automation cap
No workflow is assumed fully automatable. Automation potential is capped at 85% of the baseline, no matter how clean the process looks. Some are rated as low as 20%.
72%
Conservative mode
Conservative mode cuts deeper. The conservative projection applies an 80% realization rate and a 90% ramp discount, so it credits 72% of the optimistic hours, and payback math adds a 20% buffer on implementation cost.
0
Missing data
Missing data counts as zero. If we don't know a workflow's volume or frequency, the arithmetic doesn't invent it. Unknown means uncounted.
Step 03 · Counting the runs
Baseline, minus what the run actually took.
Once a workflow ships, every run is logged, and the measured saving per run is simple: the baseline time for a cycle, minus how long the run actually took, including any human time reported inside the run. Floored at zero: a run that saved nothing reports nothing, never a negative massaged away and never a positive invented.
Run 04 saved nothing, so it reports nothing. Never a negative massaged away.
No runs, no savings
No runs, no savings. A workflow that ships and never gets used reports zero, which is exactly what it saved. Adoption isn't assumed in our numbers. It's counted.
Step 04 · Time into money
Your numbers, at your rates.
Hours convert to dollars at a blended rate you control. Set it per division or per workflow; where you haven't, a $100/hour default applies until you change it. Your numbers, at your rates. We'd rather report a smaller figure at your rate than a bigger one at a rate we picked.
Blended rate · yours to set
What we refuse to count
The fastest way to inflate an ROI number is to count things that can't be measured.
So we don't:
"Happier team"
"Happier team" doesn't get a dollar figure. It's real, clients tell us about it, and we'll quote them. It's not in the arithmetic.
Errors avoided
Errors avoided and risk reduced stay out of the dollars. A process that stops producing mistakes is worth a lot, and we won't pretend we can price disasters that didn't happen. Error-prone workflows rank higher on the roadmap. They don't pad the savings.
Revenue attribution
Revenue attribution stays out. When a client grows 4× on 2× headcount, the operations work contributed, and we won't pretend we can isolate the percentage. We report the capacity created and let you draw the line.
If a saving can't survive this page's arithmetic, it doesn't go in the number. It goes in a quote or a case study, clearly, as the softer thing it is.
Check the numbers yourself
If a number can't show its runs, we shouldn't be reporting it.
Every measured figure in your dashboard opens up: per-workflow run counts, minutes saved, last-run dates, and a run-by-run log showing every run's date, status, minutes saved, and what happened at each step. Self-serve, not a walkthrough we schedule.
Why we hold the line here
Ace Work's pitch is that we find the invisible tax and prove we removed it. That second half only works if the proof is boring: counted, discounted, and checkable. The moment a savings number can't survive a skeptical CFO, everything upstream of it, the map, the roadmap, the pricing of every project, stops being trustworthy too.
So the numbers stay dull, and hold up.