Small Companies, No Base Rate
Nearly all serious data in this field is enterprise. RAND's 80.3%, Gartner's 28%, the 350-executive survey at companies above a billion in revenue.
There is no published base rate for a company of forty people, and the enterprise figures are quoted at them constantly. Some of what those figures describe transfers and some of it inverts. A product-side example of how workforce software approaches this topic is available in this resource.
What is harder at small scale
No volume. The fixed build cost does not scale down, so the arithmetic that works at 10,000 instances a month fails at 300. For broader background and an independent point of comparison, see ProcessMaker.
No specialist. Nobody whose job is process improvement, nobody who has done this before, and the person evaluating the proposal is also running the department.
No leverage with vendors. Enterprise clients get attention, custom work and a named contact. A small client gets the standard package and a support queue.
And no redundancy. The named owner who has actually run it is one person who also does four other things, and their departure is a genuine crisis rather than an inconvenience.
What is easier
Genuinely, and it is rarely said.
The process is knowable. Mapping it takes a day, not a quarter, and one person can hold the whole thing in their head. That is a real advantage over an enterprise process spanning four departments and two countries.
The decision is fast. No steering committee, no eighteen-month procurement. The unowned-process failure mode is much rarer when there is one person who can decide.
Change management is a conversation. The four people affected can be in a room. At enterprise scale that is a programme; here it is an afternoon.
And the ceiling is lower, which helps. A small automation that saves six hours a week is a good outcome and needs no board approval. Enterprise projects fail partly because they are large, and small ones can afford to be small.
What actually suits small scale
Off-the-shelf over custom, almost always. The build cost is the barrier, and a configured tool at a subscription price changes the arithmetic entirely.
Integration over automation. Much of what small companies want to automate is two systems not talking. A connector is cheaper, more stable and more maintainable than a robot re-keying between screens.
One process, thoroughly, rather than a programme. The highest-volume, lowest-exception thing you do.
And boring targets. Scheduled reports, data transfers, notifications, reconciliations. Unglamorous, low exception rates, and they work.
What to do about the missing base rate
Accept that published figures are orientation only. They tell you the causes of failure, which transfer, and not the probability, which does not.
Measure your own. Two weeks of tallying gives you the number that matters, and at small scale that is a genuinely small amount of effort.
And ask vendors for references your size. A supplier whose case studies are all enterprise has not implemented at forty people, and their estimates for internal effort will be wrong in a direction that costs you.
The one enterprise finding that transfers cleanly
The causes of failure are organisational rather than technical. Process selection, ownership, expectations, change management.
That transfers completely, and it is good news at small scale, because every one of those is cheaper to get right with forty people than with four thousand. The advantage is real and it is available to anyone willing to spend a day mapping before spending anything else.
The short version
- All serious data is enterprise; there is no published base rate for small implementations
- Harder at small scale: no volume, no specialist, no vendor leverage, no redundancy in the owner
- Easier: the process is knowable in a day, decisions are fast, change management is a conversation, and small projects can afford to be small
- What suits: off-the-shelf over custom, integration over automation, one process thoroughly, and boring targets
- Published figures give you causes, which transfer, not probabilities, which do not
- Ask vendors for references your own size — enterprise-only case studies mean their internal-effort estimates will be wrong