Skip to content
Blog
Measurement · 7 min read

A baseline is the only business case that survives procurement

Automation proposals fail in review for one reason: nobody can say what the process costs today. Here is how we take a baseline that holds up when finance asks.

Enterprises

Every automation proposal eventually meets the same question: what does this process cost us today? If the answer is an estimate, the proposal is an opinion. A baseline is what turns it into a number somebody can sign.

Why the business case usually fails

In our experience the technical case for automation is rarely the problem. The failure is almost always evidential: the team knows the process is slow and expensive, but the last time anyone measured it was during a project that has since been replaced.

Procurement does not need certainty. It needs a number with a stated method, a capture window and a named owner, so that the post-implementation figure can be compared against it like for like. That is all a baseline is.

The five numbers worth capturing

A useful baseline is smaller than people expect. Five measures cover almost every back-office process we have automated, and each one can be captured without buying a tool.

MeasureHow it is capturedWhy it matters
VolumeCount of items per weekSets the size of any saving
Cycle timeTimestamp at intake and at completionThe claim most likely to be challenged
Touch timeSampled observation of hands-on minutesSeparates waiting from working
Rework rateItems returned for correctionThe hidden cost of poor quality
Exception shareItems needing manual judgementSets the realistic ceiling

Capture them for four consecutive weeks. Anything shorter catches a month-end or a holiday and produces a baseline nobody trusts.

Separate waiting from working

Most of the time a process takes is queueing, not work. If a document sits in an inbox for two days and is processed in eight minutes, automating the eight minutes changes very little; routing the document is the actual opportunity.

This distinction is why touch time and cycle time are measured separately. Reporting only cycle time invites a proposal that automates the wrong half of the process, and the resulting saving fails to appear in the numbers the business actually watches.

  • Cycle time is what the customer experiences.
  • Touch time is what the team costs.
  • The gap between them is what automation usually removes.

Decide the exception policy before the build

The exception share determines whether automation removes work or just relocates it. A process with a thirty percent exception rate needs a human decision path designed up front; discovering that mid-build is how projects acquire a second phase nobody budgeted for.

Write the policy down in the baseline document: what happens to an item that fails validation, who owns it, and how long it may wait. When the build starts, that page becomes the acceptance criteria rather than a debate.

Agree the post-implementation method now

The most common reason a measured saving gets rejected is that the after-numbers were collected differently from the before-numbers. Same source, same window length, same counting rules, stated in advance.

We put this in the baseline as a short protocol: which system the data comes from, how an item is counted, and who signs off. It takes an hour to write and it is the difference between a saving the business books and a saving the business argues about.

We put this in the baseline as a short protocol: which system the data comes from, how an item is counted, and who signs off. It takes an hour to write and it is the difference between a saving the business books and a saving the business argues about. The savings calculator on this site is built on the same assumption: the number is only as good as the baseline underneath it.

What the baseline is not

It is not a business case on its own, and it is not a promise about what automation will deliver. It is the denominator. The proposal is what you put on top of it, and the honest version states the range of outcomes rather than a single number.

An organisation that can produce a defensible baseline is, in our experience, ready to automate. One that cannot will spend the first month of any engagement building it, which is why we do it before the scoping conversation rather than during the build.

Keep reading

More from the blog