Senthil Info

automating business processes, honestly

What a Payback Period Contains

Vendor material in this field commonly quotes payback in six to nine months. The figure is consistent enough across suppliers to be a convention rather than a finding, and it is worth taking apart because the arithmetic is simple and the assumptions are not.

Reviewed August 9, 2026. Six-to-nine months and 100–200% first-year ROI are the vendor-published baseline; treat them as marketing rather than as measurement. A product-side example of how workforce software approaches this topic is available in this resource.

The arithmetic

Payback is total cost divided by monthly benefit. Both terms are softer than they look. For broader background and an independent point of comparison, see Red Hat Automation.

Total cost should contain: licences, the build, integration work, your own people's time during the project, testing, training, and the exception-handling design. Vendor models usually contain the first three.

Monthly benefit should be the cash you stop spending, plus a separately stated capacity gain. Vendor models usually blend them.

And the clock starts at go-live, not at signature. Everything before is cost with no benefit.

Four costs outside the usual calculation

Your own people's time. Process discovery, testing, sign-off, the meetings. For a mid-sized automation this is frequently the largest internal cost and it appears in no proposal, because the vendor is not spending it.

The exception-handling design and staffing. Automation concentrates the exceptions, and somebody handles the residual. If the answer is "the existing team," their capacity gain is smaller than the model says.

Maintenance from year two. Rules change, systems update, the automation breaks. A model showing nine-month payback and no maintenance line is modelling eight months of reality.

And the parallel run. Most sensible implementations run manual and automated together for a period. That period costs more than either alone and it is rarely in the plan.

What has to be true for six to nine months

Working backwards from the convention.

A high straight-through rate, so the benefit applies to most of the volume.

Implementation inside the quoted duration. Slip removes benefit months while adding cost months, so a third of slippage moves nine months toward twelve.

Immediate full adoption, with no ramp and no volume routing around the automation.

Cash savings rather than capacity. If nobody stops being paid, payback in the cash sense may never arrive, whatever the hours-saved figure says.

And no maintenance in the payback window, which is true and is why the window was chosen.

Where all five hold, six to nine months is achievable. Where two fail, the honest figure is eighteen to twenty-four, and that is still a reasonable investment — it just is not the one that was approved.

The number worth calculating instead

Cost per transaction, fully loaded, before and after.

It is harder to game than payback because both sides include everything: licence amortisation, maintenance, exception handling, and the residual manual work.

It also survives the second year, which payback does not — payback is a one-time claim, and cost per transaction is a running measure that tells you whether the automation is still worth having. Which is the question nobody asks in year three.

What to ask for

A payback calculation with maintenance included from month one, even at a placeholder rate.

The internal effort estimate, in your people's days, not theirs.

A cash-versus-capacity split in the benefit line.

And the same calculation at 70% of the assumed straight-through rate, which takes them ten minutes and tells you how sensitive the case is.

A vendor who produces all four is showing you something real. One who resists the fourth is telling you the case does not survive it.

The short version