How to choose one expensive workflow worth fixing

A workflow is not worth fixing because it is annoying. It is worth fixing when the recurring result matters and the business can act on the answer.

Use five tests before commissioning analysis or buying technology.

1. The problem recurs

A one-off failure may need correction. A workflow engagement needs a repeated path: invoices matched every week, exceptions reviewed every day, billing reconciled every month or customer evidence transferred for every case.

Frequency creates value and makes the current result observable.

2. The economic consequence is material

Name the consequence in the right class:

• cash or P&L cost;
• cost that can credibly be avoided;
• capacity tied up;
• revenue leakage;
• working-capital effect;
• service or control risk.

Do not turn every class into a savings claim. Two hundred hours released may be useful, but it is not cash unless a cost changes or the capacity is put to an agreed economic use.

As an initial qualification guide, TightShip looks for a plausible path to at least $100,000 in annual value or equivalent strategic importance. That is a threshold for investigation, not a promised result.

3. Evidence exists

A workflow needs a current-state observation. Useful evidence may include ledger entries, invoices, contracts, workflow events, system logs, queues, timestamps, exception records or an agreed sample.

If the only evidence is that people dislike the process, the case is not ready.

4. One person owns the result

A workflow may cross several functions. It still needs one accountable owner who can decide how it should operate and resolve conflicts.

A recommendation with no owner is unfinished work.

5. The business can implement

A good diagnosis is wasted when the sponsor cannot approve system access, process change, security review or adoption.

Test authority before detailed analysis. A paid diagnostic should lead to a real decision, even when that decision is to stop.

A simple disposition

Use four outcomes:

• Investigate when value, evidence and ownership justify a bounded diagnostic.
• Fix when the workflow and intervention are already well bounded.
• Hold when a specific missing condition can be closed.
• Kill when value, evidence or authority does not justify more work.

The discipline is choosing one workflow and one decision. A company-wide transformation mandate can wait.

Once a workflow qualifies, establish what a decision-grade workflow baseline actually contains. If exceptions carry the cost, use the exception-cost method to separate their source, handling and consequence.

Pressure-test one workflow.