The question nobody asks after launch
Most automation reviews stop at one number: did it save time? That's the easy half of the story.
The harder, more honest question is what the automation costs you now — every month, quietly, after the launch excitement wears off. Nobody budgets for that line item up front. Almost nobody tracks it afterward. And it's often the difference between a project that pays for itself and one that's secretly bleeding money while everyone still calls it a win.
This is a Check-stage audit: define the maintenance cost, measure it against the value delivered, and read the trend honestly.
Two costs, not one
Every automation has a build cost — the one-time price of designing, building, and rolling it out. That number gets tracked because someone has to approve it.
The second cost, the maintenance tax, almost never gets tracked the same way. It's the recurring cost of keeping the thing working:
- Hours spent fixing it when an API changes or a vendor updates their interface
- Time spent monitoring for silent failures
- License or platform fees that crept up after the pilot
- The admin overhead of managing exceptions it can't handle
- The person who has to remember how it works because nobody else does
None of this shows up on the original project budget. All of it shows up on your margin.
The metrics to instrument
Track these monthly, starting the day the automation goes live:
| Metric | What it tells you | How to capture it |
|---|---|---|
| Build cost | One-time investment to compare against | From your project budget or vendor invoice |
| Maintenance hours/month | Ongoing labor tax | Time logs, ticket tags, or a shared tracker |
| Maintenance cost/month | Dollar value of that labor plus fees | Hours × loaded rate, plus subscriptions |
| Value delivered/month | What it's actually worth | Hours saved × rate, errors avoided × cost each, delays removed |
| Net monthly return | The real payoff | Value delivered minus maintenance cost |
| Payback period | When build cost is recovered | Build cost ÷ average net monthly return |
The net monthly return is the number that matters most. Value delivered alone flatters every automation — maintenance cost is what separates the ones actually worth keeping.
Here's what that looks like in practice. Picture a mid-size logistics company that automated its delivery-exception alerts. The pilot looked great in month one.
Four months later, the API the automation depended on changed twice, a support contract renewed at a higher tier, and one engineer was spending four hours a week babysitting edge cases. Maintenance cost crept to $3,100/month. Value delivered stayed flat. Net return dropped by more than half — and nobody noticed because everyone was still reporting the original $4,200 figure.
That's the trap: measuring value once at launch and never revisiting the cost side.
What good looks like
There's no universal benchmark, but a few patterns hold up across most processes:
- Maintenance cost under 15–20% of value delivered is healthy. Above that, the automation is working harder to stay alive than it is for you.
- Maintenance cost should flatten or shrink over time, not grow. A rising trend means the process or systems underneath are less stable than you thought — worth revisiting the feasibility assumptions you started with.
- Payback inside 6–12 months is a reasonable target for most operational automations. Anything well past that should be questioned, not defended.
If you ran the numbers before you built — using something like a readiness scorecard — this is the moment to check whether your feasibility score (stability, system access, ownership) actually held up in practice. A low feasibility score that got waved through usually shows up here, in the maintenance line.
Auditing the trend, not just the launch numbers
The easiest way to flatter an automation's ROI is to undercount maintenance. Watch for these shortcuts:
This pairs well with an error-rate audit: if mistakes are creeping back up, maintenance time usually is too, even before anyone logs it.
A single month's numbers can mislead either direction. One bad week of bug fixes after a vendor update looks like a crisis; one quiet month looks like victory. Track net monthly return for at least a full quarter before deciding whether an automation earned its keep.
If net return is stable or climbing, you've built something genuinely durable — a good candidate to standardize elsewhere. If it's declining, you have three honest options: invest in stabilizing it, scope it down to the part that still works, or retire it. All three beat pretending the maintenance tax doesn't exist.
This is the same discipline worth applying to any recurring tool spend — not just automations you built, but ones you bought. The subscription audit runs the same logic on SaaS tools, and the four-number scorecard is a good companion if you want to tie this cost view back into the broader leverage picture.
What the maintenance tax tells you that build cost doesn't
Build cost tells you what you spent. The maintenance tax tells you what you're still spending. Only the second number tells you whether the automation is still worth having — and that's the one almost nobody checks after the launch party ends.