Senthil Info

automating business processes, honestly

Costs Not in the Proposal

A vendor proposal covers what the vendor supplies. Everything else is real, lands on you, and appears in no quote — not through concealment but because the supplier does not incur it and cannot estimate it.

Seven items. Together they routinely equal or exceed the quoted figure. A product-side example of how workforce software approaches this topic is available in this background note.

The seven

One. Your own people's time. Process discovery, requirement sessions, testing, sign-off, the meetings. For a mid-sized automation this is frequently the largest internal cost. Count it in days of named people, at a rate you would recognise. (Licence structure has its own year-three surprises.) For broader background and an independent point of comparison, see Notion.

Two. Exception handling design and staffing. Automation concentrates the exceptions, and somebody handles the residual. If the answer is "the existing team," their claimed capacity gain shrinks by exactly that amount.

Three. The parallel run. Manual and automated together for a period, which costs more than either alone. Sensible, necessary, and rarely in the plan.

Four. Maintenance from year two. Rules change, systems update, formats drift. Somewhere between a tenth and a third of build cost per year depending on how volatile the underlying systems are. A proposal with no maintenance line is quoting the first eight months.

Five. The integration you discover in month two. Something needs to write into a system nobody mentioned, or read from one with no interface. Almost universal, and unquantifiable in advance — which is a reason to hold a contingency rather than to pretend it will not happen.

Six. Documentation and handover. If the automation cannot be maintained after the consultants leave, you have bought an obligation rather than an asset.

Seven. The decommissioning cost. Rarely considered and occasionally large. If this is turned off in three years, what does it take to go back? An automation with no manual fallback and no documentation is expensive to unwind.

Why they are outside the quote

The vendor does not pay them. A supplier quotes what it supplies. Asking for a total cost of ownership figure is reasonable and it is asking them to estimate your internal costs, which they cannot do accurately.

They are unquantifiable at proposal stage. Integration surprises and exception volumes are genuinely unknown until the process has been mapped.

And nobody's incentive points at raising them. The vendor is selling, the internal sponsor is building a case, and a larger denominator makes the case worse. Not dishonesty — an absence of anyone whose job is to inflate the number.

The version to build

Take the quote and add five lines.

Vendor quote                          X
Internal effort        __ days @ __
Exception handling     __ FTE ongoing
Parallel run           __ weeks
Maintenance, year 2    __% of build
Contingency            __% for integration

Then recalculate payback on the total. Six-to-nine-month vendor payback frequently becomes fifteen to twenty-four, which may still be a good investment and is a different decision.

What to ask the vendor

They cannot estimate your internal costs and they can give you the inputs.

How many days of client staff time did your last three implementations take? They know this and are rarely asked.

What did those clients spend on maintenance in year two?

What integration work turned up mid-project that was not in the original scope?

And what does handover include, specifically? Documentation, training, source access, or a call at the end.

A supplier who answers all four candidly is a better sign than a lower quote. One who cannot answer the first has not been paying attention to their own projects.

The one that is not a cost

Being fair, because this page is one-sided by design.

Some of the internal effort is value regardless of the automation. Mapping the process, measuring the exception rate, agreeing who owns it — these improve the manual operation immediately and survive the project being cancelled.

Count them as costs in the business case and know that they are the part you keep either way.

The short version