← All articles

Operations

Compare the Documented Process With the Process That Actually Happened

The documented process explains how work is supposed to move. Captured evidence shows how distinct attempts moved in practice. Compare both without treating either one as the whole truth.

Text-free Shadow evidence-map composition in ink black, warm cream, mustard gold, and muted sage, showing connected process paths and evidence points.

Start with the documented process as a reference point

Begin with the approved procedure, checklist, playbook, or policy. Treat it as the intended path: the required steps, decision points, handoffs, and expected outputs. It gives you a consistent reference for comparison, even when the real work does not follow it exactly.

Do not assume that a difference is automatically a failure. A documented process may leave out local judgment, exceptional cases, or steps that were added after the document was written. The purpose of comparison is to make those differences visible and understandable.

Reconstruct each attempt from captured evidence

Review evidence in batches that belong to a distinct business attempt, such as a request moving from intake to approval or a case moving from review to resolution. Reconstruct the sequence of observed actions, artifacts, transitions, and outcomes for that attempt. Keep the boundary of each attempt clear so unrelated work is not blended together.

Record what the evidence shows before explaining why it happened. A citation establishes where a detail came from; it does not prove that the detail is complete, correct, or representative. Mark gaps, conflicting records, and ambiguous transitions instead of filling them with assumptions.

  • Recorded evidence: what appears in the captured material, including its source and position in the sequence.
  • Interpretation: a reasoned explanation of what a step or variation may mean.
  • Planning assumption: a proposed condition for testing or redesign, not a fact about past work.

Compare patterns across distinct attempts

Place several reconstructed attempts beside the documented process. Look for recurring paths rather than isolated deviations. Useful patterns include the same undocumented handoff, repeated requests for missing information, a decision that occurs earlier than expected, or a system change that consistently appears between two documented steps.

Keep exceptional paths visible, but do not let one unusual attempt define the process. A recurring variation is a stronger candidate for process improvement than a single deviation. At the same time, frequency alone does not establish that a variation is desirable, compliant, or caused by a particular constraint.

  • Where does the observed sequence consistently diverge from the documented sequence?
  • Which steps are repeated, skipped, combined, or moved to a different point?
  • Which variations appear across multiple attempts, and which appear only once?
  • What evidence is missing before a proposed explanation can be treated as more than a hypothesis?

Use app and site sequences as discovery signals

Sequences across applications and sites can help reveal where work may move between tools, where information may be re-entered, or where an undocumented handoff may exist. Treat these sequences as signals for investigation, not as a complete account of the work. An application sequence does not by itself establish intent, completion, time spent, or business impact.

Validate a suspected workflow pattern against business records, artifacts, and interviews or walkthroughs where appropriate. Keep the sequence separate from the reconstructed business attempt until the evidence supports a connection. This avoids turning navigation traces into claims about people or performance.

Turn the comparison into a focused improvement plan

Choose one specific mismatch that matters to the outcome and can be investigated further. State the observed pattern, the evidence supporting it, the uncertainty that remains, and the change you want to test. For example, you might test whether adding an intake field reduces a recurring clarification loop, rather than claiming that the current loop has already caused a measured loss.

Define what would count as a useful result before changing the process. Identify the records to review, the cases in scope, the owner of the experiment, and the conditions that could explain a different result. Keep the plan grounded in evidence and label every forecast as an assumption until it is tested.

  • Observed pattern: the recurring difference between documented and reconstructed sequences.
  • Open question: what still needs confirmation or additional evidence.
  • Proposed test: the limited process change and the cases where it will be examined.
  • Review point: when the evidence will be reassessed and the process documentation updated.