Which processes first? A scoring matrix for mid-size firms

Impact is not an opinion but a calculation. You need two figures per workflow: how often it occurs and how long it takes. Both can be gathered within a few days without disrupting operations. For frequency, a look into the leading system is usually enough — the number of documents, orders, enquiries or reports in the last quarter. For duration, record five to ten real runs rather than asking for an estimate. Estimates from the people involved tend to come out too low, because searching, asking back and waiting between steps are not counted (project experience).

The two figures turn into an annual number. Runs per week times minutes per run times working weeks gives minutes per year; divided by sixty it gives hours. Those hours are the currency in which very different workflows become comparable — goods receipt in the warehouse and invoice checking in the office. Valuing the hours with an internal hourly rate produces a figure that also carries weight with management. Honesty matters in both directions here: in a mid-size company, saved hours rarely turn into saved headcount but into capacity for work that currently gets left undone.

The values in the example are chosen, not measured (project experience). What matters is the structure of the calculation: a workflow that looks small becomes large through repetition. Six minutes are nothing; six minutes a hundred and twenty times a week tie up more than a third of a full-time position. Conversely, a laborious workflow that occurs twelve times a year stays small, even when it takes up a lot of room in conversation.

The second value in the matrix is the effort of implementation. In practice it is regularly underestimated, because the visible part gets estimated — the new screen, the new report — while the invisible part is skipped: data quality, special cases, rollout. A rule of thumb: the more systems a workflow touches and the more exceptions it knows, the further the actual effort sits above the first estimate.

The matrix does not need a firm quotation, only a shared classification into five levels. It helps to anchor effort not in days but in its drivers. Four of them explain most of the spread between estimate and reality (project experience).

The most complex workflow is usually the one people complain about most loudly — production planning, quotation costing with many variants, route planning. The temptation is strongest there because the pain is visible. For a first step it is still the poorer choice, for three reasons: the requirements are not yet stable, the number of special cases is unknown, and the people involved have no shared experience of this kind of project work.

There is also a timing effect. A complex project often delivers its first visible result only after months. During that time nothing happens in the company that justifies the effort, while the people involved attend meetings and answer questions. Impact comes not from complexity but from repetition. The most complex process stays on the list, but further down and split into pieces that each carry a benefit on their own.

The matrix needs no software. It consists of one row per process and six columns, each column scored from one to five. Three columns describe impact, three describe feasibility. The sum gives the ranking; because both halves carry equal weight, a project cannot climb on impact alone if it is unrealistic in practice. Conversely, easy feasibility alone is not enough either.

More important than the exact score is that everyone uses the same scale. So before scoring, define what one point and what five points mean. The following assignment has proven itself in practice and can be adapted to your own wording and orders of magnitude in half an hour.

Do not fill in the matrix on your own. Management knows the frequency from the system; duration and errors are known to the people who carry out the workflow every day. When the two sides score the same row differently, that gap is itself a finding: the workflow is not documented the way it is lived. Resolving that difference is often already part of the improvement.

The sixth column of the matrix is the unusual one: visibility. It has little to do with economics and a great deal to do with whether the programme holds. A project that touches several areas every day creates conversation after the rollout. A project that runs in the background in accounting may save more hours but goes unnoticed in the rest of the company — and turns up at the next budget discussion without an advocate.

The first project is a proof, not a savings programme. It should show that a workflow really can get better in this company, and do so quickly enough that the link between effort and result stays visible. As a guideline, no more than eight weeks should pass between the decision and a tangible result (project experience). If it will foreseeably take longer, split the project into a first part that already takes work off people's hands and a second part that follows later.

A scoring matrix ranks, it does not decide. There are good reasons to deviate from the ranking, and they are legitimate as long as they are stated and recorded. A statutory deadline, a customer project with a fixed requirement, a system replacement that is due anyway — all of these can pull a workflow forward even though it does not lead on points.

Legal requirements deserve separate consideration. Retention obligations, the traceability of postings, data protection for staff data and works council codetermination for systems capable of recording behaviour or performance affect many digitisation projects directly. These points cannot be weighed with scores; they belong in a professional review before implementation, in individual cases by a tax adviser or a lawyer. This article sets out the context and does not replace such a review.

Record every deviation in one sentence. A process documentation that also holds the rejected candidates and the reason they were deferred saves half the discussion in the next round — and makes visible which precondition has since been met and which candidate therefore moves up.