Process optimization fails at the handover, not at the tool
7 min read
Process optimization in mid-sized companies fails in most projects not because of the wrong method or a missing tool. It fails at the handover between departments that are neatly separated in the org chart but overlap in reality. Those who don’t address this will spend six weeks optimizing something that falls apart again just beyond the next door.
Key Takeaways
- Handovers are the real bottleneck. Tools don’t heal a process whose departmental interfaces remain undefined. Introducing software first documents the break, not the fix.
- Three sub-processes account for most hours. In procurement, sales and back-office lie the levers that never appear on the strategy slide yet consume weekly working hours.
- Measure before you manage. Without before-and-after numbers, every optimization project is just an assertion. Three simple metrics are enough-set them before you start.
Related:Process Optimization Without the Permanent Project / If You Need Three Days, You’ve Already Lost the Lead
Why process optimization breaks at the interface, not the tool
What does process optimization look like in practice? Process optimization is the systematic improvement of a workflow by cutting wait times, duplicate work and handover losses between functions. In mid-sized companies it rarely means a new ERP module and almost always means clarified responsibility at a concrete interface between two departments.
Three real-world snapshots from actual projects: A purchase request bounces back and forth between the specialist department and procurement for 14 days because neither side knows who secures supplier approval. An offer is finalized in sales and sits in accounting review for three days because the employee there is also processing invoice runs. A complaint circulates endlessly between service, quality and accounting because the last handover was never properly defined.
In all three cases a new workflow tool would not reduce hours; it would merely document the process faster, not execute it faster. Spending money on software first buys visualization, not impact.
Three quiet levers in purchasing, sales, and back office
These three interventions share one thing: they require no project budget. All they need is a management decision to clarify responsibility and a Monday morning meeting to announce the new rules. Teams that implement this within two weeks will see measurable effects within three weeks.
What process mining realistically delivers in mid-sized companies
Process mining has been the consulting world’s darling for a decade. In mid-sized companies, however, it rarely delivers what the pitch promises. The tool reads event logs from ERP, CRM, and workflow systems and reveals the actual process paths. That’s useful when logs are complete. In mid-sized firms, logs are often patchy because transactions bypass systems via Excel or email.
A process-mining report on a process that runs half in Outlook doesn’t show the process-it shows the system skeleton beneath the real workflow.
That doesn’t mean process mining is useless. It means it’s valuable only once processes truly live inside the system. Before that, it’s smarter to bring the processes into the system than to hire an external analyst who misses half the picture.
How AI Really Saves Time in Back-Office Reality
AI promises for administration are grand and have been reheated for two years now. What actually works measurably by 2026 is far smaller than the pitch decks suggest. Invoice recognition with OCR plus LLM validation runs stably in many ERP add-ons and saves a finance team roughly two minutes per invoice. At 1,000 invoices a month, that’s one full day freed up each month-not a quarter. Measurable? Yes. Transformational? Not quite.
What doesn’t work: AI assistants hunting receipts in Outlook, auto-renegotiating contract clauses in Word, or auto-generating quarterly reports from ERP data. Pilot projects exist, but they rarely deliver productive time savings and collapse under the variability of incoming data.
The pragmatic path for back-office AI: one or two tightly scoped use cases with high volume and low variance. Invoice workflows, supplier master-data entry, travel-expense classification. Three use cases per mid-market company are plenty for 2026. Anything more and you’re likely to lose focus.
Four Steps Every Process Project Must Clarify Before It Starts
Before any tool is selected or workshop booked, four decisions are worth making. Together they take a single morning and decide whether the project still has substance six weeks later.
First: Which interface is being optimized. Not “the procurement process,” but: the handoff between the business unit and purchasing for purchase requests over €5,000. Specific, bounded, and documentable.
Second: Which baseline metric will be measured. Cycle time in working days, number of handovers, number of queries, processing errors. One metric is enough-it must be captured before any changes.
Third: Who owns it. One person, not a committee. If the optimization fails, that person knows why. If it succeeds, they made it happen.
Fourth: When will it be reviewed. Not “in a year,” but: in four weeks the baseline metric will be compared to the new value. Four weeks is long enough for a measurable effect and short enough that everyone remembers what they changed.
Frequently Asked Questions
Is Six Sigma or Lean worth it for SMEs?
Six Sigma as a complete methodological package is almost always overkill for SMEs. Individual tools such as value stream mapping or cause-and-effect diagrams deliver solid results in short workshops. Lean concepts like Kanban work very well in production processes, and in office environments with some adjustments. Hiring a Six Sigma consultant locks in methodology you rarely need in full depth.
When should you introduce a workflow tool?
Only once the process itself is organisationally clear and the bottleneck is simply the speed of execution. That scenario is rarer than vendor demos suggest. Plugging a workflow tool into an unclear process freezes the break in software. Clarify the hand-offs first, then pick the tool that supports those transitions.
How do you convince leadership to accept small, incremental steps instead of a big-bang project?
Show the numbers from the first intervention. Reducing a sub-process by 30 percent in four weeks gives you the ammunition for the next step. Large projects often launch after a failed small intervention because the mandate evaporated. If you want the optics of a big project, you must first rack up the small wins.
What if a sub-process cuts across five departments?
Break the process into at least three sub-processes, each with its own owner and defined hand-off. Cross-functional optimisations routinely fail because they lack clear accountability. Three smaller, owner-led interventions move faster and stick better than one end-to-end project with no single point of responsibility.
Editor’s Reading Picks
More from the MBF Media Network
Cover image: AI-generated (May 2026)
Cover image source: Pexels
