How many processes should we track?

Three or four. Beyond that, in a company without a dedicated function, the review becomes reporting rather than decision-making.

Should this sit with IT?

No. IT is usually needed to extract data, but ownership belongs with the function that carries the commercial consequence, otherwise improvements stall at the point where a working practice has to change.

What if we have no baseline at all?

Spend the first month establishing one rather than launching an improvement. A change without a baseline cannot be defended later, and the baseline itself usually surfaces the first opportunity.

AI Strategy

Process intelligence for European mid-market companies

Process intelligence for companies without a transformation office: what to measure, who owns it, how to sequence improvements and how to keep the numbers honest.

Miguel Torres, Founder, Merjora · Published 4 September 2026 · Updated 4 September 2026

In short

Process intelligence is the practice of knowing, from evidence rather than opinion, how your core processes perform and what deviations cost. In a mid-market company it works when it is owned by one accountable operator rather than a central function, limited to three or four processes with a monetary meter, measured against a baseline agreed with finance, and reviewed on a fixed monthly cadence. The failure mode is not lack of analysis; it is analysis with no owner and no re-measurement date.

Key takeaways

Three or four processes measured properly beats twenty measured superficially.
Every process metric needs a euro translation or it will lose the budget argument.
One accountable owner per process, named, with time protected for it.
Re-measurement is the whole discipline: an improvement nobody re-measured did not happen.

Why enterprise process excellence does not port down

Large companies run process excellence through a permanent function: a transformation office, analysts, a methodology and a governance forum. A 200-person company has none of that, and creating a small version of it usually produces meetings rather than change.

What does port down is the discipline: an agreed baseline, a small number of tracked processes, a named owner and a review rhythm. That can live inside an existing operations or finance role.

The metric set worth maintaining

Process	Primary metric	Euro translation
Order-to-cash	Days sales outstanding; invoice dispute rate	Cash tied up in receivables; write-offs
Procure-to-pay	Invoice touchless rate; early-payment discounts captured	Processing cost per invoice; discounts lost
Service intake	First-response time; reopen rate	Deals lost to slow response; avoidable handling
Fulfilment	On-time delivery; rework rate	Credit notes; expedited shipping
Hiring or onboarding	Time to productive	Revenue or capacity lost per week of delay

A minimal process metric set for a mid-market company

Ownership: the part that decides whether this survives

One named owner per process, from the function that carries the consequence — not from IT
A monthly review of no more than 45 minutes, with the same three questions each time
A written baseline agreed with finance before any improvement starts
An explicit stop rule: an improvement that has not moved its metric in a defined period is closed, not extended

Sequencing improvements when capacity is the constraint

Mid-market companies rarely lack ideas; they lack the attention to run more than one or two changes at a time. Sequence by the ratio of value to organisational disruption, and deliberately front-load one improvement that finishes quickly, so the cadence earns credibility before it takes on something hard.

Anything that requires changing how three departments work simultaneously belongs later in the sequence, not first, however large the modelled value.

Keeping the numbers honest

The fastest way to discredit process work internally is to claim savings that never appear in a budget. Separate three things in every business case: hours released, cost avoided, and cash actually saved or collected. Only the third is a euro finance will recognise without argument, and only if someone has agreed in advance what will change in the budget.

Baseline agreed and dated before the change
Value expressed as a range, with the assumptions written down
Re-measurement date set at the same time as the go-live date
Realized value reported against the original claim, including the misses

Frequently asked questions

How many processes should we track?: Three or four. Beyond that, in a company without a dedicated function, the review becomes reporting rather than decision-making.
Should this sit with IT?: No. IT is usually needed to extract data, but ownership belongs with the function that carries the commercial consequence, otherwise improvements stall at the point where a working practice has to change.
What if we have no baseline at all?: Spend the first month establishing one rather than launching an improvement. A change without a baseline cannot be defended later, and the baseline itself usually surfaces the first opportunity.

Find the process worth fixing first

Merjora maps how your business actually operates, scores each opportunity and quantifies the value range before you commit budget.

Discover your AI opportunities