What Survives After the Consultant Leaves — and Why Usually Nothing Does
ImplementationMid-Market

What Survives After the Consultant Leaves — and Why Usually Nothing Does

T. Krause

T. Krause

The symptom

There's a pattern familiar to anyone who has lived through a consulting project. At the end there's a solid analysis, a cleanly modelled target process and a presentation nobody argues with. The consultant leaves. Three months later things are still largely running the new way. After nine months the old route has crept back in at two or three points. After eighteen months the business remembers the project as "that time, with the workshops."

The usual explanation is that staff didn't support the change. In my experience that is almost never true. People carry change remarkably far when it works and when it has been thought through to the end. When a process falls back, it generally failed at a specific point — and nobody was responsible for noticing.

The mechanism

Three things are almost always missing, and none of them appear in a closing presentation.

The exception isn't described. A target process maps the normal case, because the normal case is 85 per cent of transactions. The remaining 15 per cent — the rush order, the special approval, the customer with the non-standard agreement — appear nowhere in the new flow. The employee has to improvise. Improvisation follows the old pattern, because that is the one they know. And since exceptions rapidly become routine in most businesses, the entire old process returns through the exception route.

There is no process owner. The department is responsible, which means nobody is. A process running across three areas has three managers and no owner. When it drifts, the first to notice is whoever bears the consequences — and they sit at the end of the chain with no authority over its beginning.

Nothing is being measured. Without a running number, regression is invisible. A process doesn't degrade in a day; it degrades by one per cent a month over a year. Someone looks only once it becomes uncomfortable — and by then, getting back is a new project.

There's a fourth, underrated point: knowledge of the why leaves the building with the consultant. Staff know the new rule but no longer its reasoning. A rule without a reason is classified as bureaucracy at the first friction and worked around.

The cost

An anonymised example, altered in detail, real in pattern. A mid-sized supplier had rebuilt its entire quotation process: clear responsibilities, a mandatory costing template, approval tiers by order value. Quotation lead time fell from nine days to three. The project cost around €45,000 and was regarded internally as a success.

Fourteen months later, lead time was back at seven days. Reconstructed, the regression traced to a single point: quotes above €50,000 required a second approval. The colleague responsible was absent for four months from the summer. No deputy had been named, so those quotes went "temporarily" back via the managing director. After his return the old route stayed, because it had settled in.

The damage wasn't the consulting fee. It was the thirteen months of restored quotation delay — with around 340 quotes a year and a demonstrable correlation between response time and win probability, an effect well into six figures of lost revenue.

The fix

The new process wasn't reinvented. It was completed at three points.

Every approval step got a named deputy — and a rule that when both are absent, the approval tier is waived rather than circumvented. A process that offers a legal shortcut doesn't get bypassed illegally.

The quotation process got an owner spanning all three departments involved, with one clear task: look at the actual lead time once a quarter and name any deviation. That's two hours every three months.

And the metric was made visible — one number, monthly, somewhere everyone already looks. Not as control, but as an early warning. Regression then announces itself instead of being discovered a year later.

This is why a process analysis with us doesn't end with a concept but with an implementation that still holds when nobody from outside is watching: named owners, described exceptions, a metric that keeps running. Fixed scope, fixed duration, fixed price.

The next step

Take a project completed one or two years ago. Ask someone who runs that process daily how they do it today. The gap between their answer and your recollection is the real number.

Cookie settings

We use technically necessary cookies to run this site. Optional anonymised analytics cookies (Google Analytics) help us improve it — we only set these with your consent.

More in the privacy policy