Build
Build only when the differentiating business rule is stable and poorly served by the market.
Practical mapping · Decision in ten days
One critical flow. One quantified cost. One clear decision in ten days.
Useful process mapping does not describe the whole company. It isolates one flow blocking an outcome, makes its decisions and rework visible, then turns observations into testable hypotheses.
The method
Map enough to decide, not to decorate.
Each step produces actionable information. Observed facts remain separate from cost hypotheses and solution choices.
1. Set the flow boundaries
Name one trigger, one delivered outcome, a start, an end and excluded cases. The scope fits in one sentence and follows a case, order or request end to end.
Expected output
A measurable flow, one owner and one outcome indicator.
2. Place actors and decisions
Identify who receives, enriches, arbitrates or approves, using which tools and rules. Thresholds, exceptions and loops matter more than the organization chart.
Expected output
Handoffs, decision rules and critical dependencies.
3. Measure waiting, rework and re-entry
Separate processing time from waiting time. Count reopened cases, re-entered information, repeated approvals and queues with no clear owner.
Expected output
A short baseline: volume, lead time, effort, rework and quality.
4. Frame the annual cost hypothesis
Make frequency, unit time, loaded cost and any delay impact explicit. Use a range when data is missing and document every assumption.
Expected output
A refutable order of magnitude, never a savings promise.
5. Decide: build, buy, integrate or no-go
Compare rule stability, available data, market solutions, integration debt and the cost of inaction before choosing software or custom development.
Expected output
A reasoned decision, its conditions and the next reversible test.
Simple annual direct-cost hypothesis
Annual occurrences × rework time per occurrence × loaded hourly cost
Lead-time, quality or revenue impacts are added separately only when documented. This prevents a correlation from becoming a guaranteed gain.
Composite example — fictional figures, no client data
From received case to triggered invoice
A multi-site services company wants to understand why complete cases wait before invoicing. This teaching example combines common situations and does not describe a real company.
- Volume assumption: 120 cases per month
- Rework assumption: 35 minutes per case
- Loaded-cost assumption: €45 per hour
- Observation to test: information is copied before centralized approval
Illustrative calculation
120 × 12 × 35/60 × €45 = €37,800 per year in hypothetical direct cost
This fictional amount only demonstrates the calculation. It excludes lead-time, quality and revenue impacts and must be replaced with an observed baseline before a decision.
Illustrative decision
Integrate first: share the required data and instrument lead time. Buy if a standard tool then covers the stabilized rules. Do not build custom software while exceptions remain unclear.
See the shape of a complete deliverableThe map does not choose technology for you.
It makes criteria comparable and can also show that no investment is justified yet.
Buy
Buy when the need is standard and adoption costs less than customization.
Integrate
Connect existing sources when value mainly comes from data continuity.
No-go
Stop or defer when problem cost, sponsorship or baseline does not justify action.
Frequently asked questions
How is process mapping different from an organization chart?
An organization chart shows formal responsibilities. A flow map follows an outcome end to end and also shows the decisions, waits, rework, tools and exceptions encountered in practice.
Must every company process be mapped?
No. For an operational decision, isolating one critical flow with a start, end, volume and measurable outcome is more useful than documenting the entire company without a priority.
How can flow cost be quantified without inventing ROI?
Separate observations from assumptions, document volumes, time and unit costs, then use a range. Lead-time or revenue impacts stay excluded until they are demonstrated.
Does mapping necessarily lead to automation?
No. The decision may be to simplify a rule, clarify a role, buy a tool, integrate systems, build a limited component or launch nothing until the required conditions are met.
Is one flow slowing an economic outcome?
Describe the symptom, volume and blocked decision. Robinswood replies with an initial flow hypothesis before recommending any tool.