Decision guide

Construction-data migration checklist

Prepare organizations, projects, contracts, cost codes, budgets, commitments, changes, schedules, documents, workflows, field records, accounting, permissions, and cutover evidence.

1. Define the decision before the shortlist

Write one plain-language decision statement naming the buyer, affected population, workflow boundary, outcome, material consequence, authority or policy context, and the date by which a decision is required. Separate must-have obligations from preferred operating improvements and future options.

Record what is explicitly outside scope. This prevents an attractive demonstration from expanding the project into adjacent work that lacks an owner, budget, implementation path, or evidence need. Revisit the market boundary and category map when proposed organizations play different operating roles.

2. Turn the workflow into requirements

For every material stage, capture the trigger, required inputs, source of authority, responsible person or organization, decision or action, exception path, evidence retained, and downstream exchange. The following capability records are useful starting points for this guide:

A feature name is not a requirement. “Supports daily reports diaries and site-event records,” for example, should become a scenario with named inputs, a governed rule, an expected decision, an exception, and an exported audit record.

3. Classify the evidence

Evidence state	What it can establish	What it cannot establish alone
Official authority source	Status, text, dates, jurisdiction, and published scope.	Buyer-specific applicability or product conformity.
Official organization documentation	Current public positioning for a named offering.	Configured depth, package, performance, outcome, or customer fit.
Provider confirmation	A dated answer tied to proposal or product scope.	Independent observed behavior.
Independent observation	Behavior in a disclosed test scenario and environment.	Every configuration, population, jurisdiction, or production outcome.
Not established	The reviewed evidence does not support a conclusion.	It does not prove the capability is absent.

4. Run comparable demonstrations

Give every organization the same representative record and expected operating result.
Test a normal path, missing data, contradictory evidence, an exception, an override, and a material source change.
Ask who owns configuration, interpretation, review, approval, release, and error correction.
Inspect timestamps, versions, user actions, source links, reason codes, and downstream records.
Export the record and test whether an independent reviewer can reconstruct the decision.

5. Evaluate implementation and exit

Document migration, integrations, content or data licensing, validation or assurance responsibilities, policy configuration, services, support, release cadence, change control, training, security, records retention, portability, and the cost of replacing the component. An operating model can be attractive during implementation but fragile during a rule change, acquisition, service transition, or exit.

6. Write a conditional conclusion

State which situation favors each option, which requirements remain unresolved, what evidence supports the conclusion, and which assumptions would reverse it. Avoid a blended score that allows a strong low-consequence feature to hide a weakness in a critical workflow.

Approval gate

The buyer, population, workflow, authority context, and decision consequence are explicit.
Every material claim has an evidence class and source.
Official positioning is not described as independent testing.
Unknowns and conflicting evidence remain visible.
Commercial relationships cannot affect the conclusion.
The final artifact is understandable by an accountable operator who did not attend the demonstrations.