Decision right
The organization names the decision the data supports, the user, response, threshold, and accountable owner.
A data point is only useful when its meaning, quality, permitted use, owner, access path, decision role, retention, and exit are clear. Put those rules in the operating model and the contract.
For each promised use, name the operational decision, required source points, context and identifiers, acceptable quality, response owner, retention need, permissions, and exit format. Data with no decision or owner is an unmanaged cost and exposure.
Data path in view
Collection, decision rights, and a usable export path should be designed before the building starts producing data.
01
02
03Published 8 September 2026 · reviewed 9 September 2026. Editorial illustrations generated for Hamed Helped; they do not depict a named site, vendor, or client.
Decision criteria
The organization names the decision the data supports, the user, response, threshold, and accountable owner.
Points retain asset, location, unit, timestamp, status, units, lineage, and semantic relationships.
Users can see completeness, freshness, calibration, transformations, exceptions, and disputed values.
Authorized roles can retrieve required raw and normalized data through workable interfaces and formats.
Collection, sharing, model use, retention, deletion, and secondary use have named approval and review.
The owner can export data, context, configuration, and history in documented formats on a tested timeline.
Questions to ask
Identify the physical source, controller, gateway, transformations, quality checks, timestamps, and current status.
Preserve units, asset relationship, location, ontology or tagging, calculation method, and business context.
Separate access for operations, tenants, vendors, analysts, models, service providers, and downstream systems.
Name the person, trigger, action, exception path, response time, and record created.
Set retention by operating, legal, privacy, maintenance, analytical, and dispute needs—not unlimited default.
Test export of raw data, normalized data, tags, relationships, rules, configurations, logs, and history.
Evidence table
| Claim or decision | Evidence | Owner | When checked |
|---|---|---|---|
| Data purpose | Decision-to-data map with named action, owner, threshold, and outcome | Business process owner | Before collection |
| Point meaning | Data dictionary, identifiers, units, semantic model, lineage, quality status | Data steward + system owner | Integration and change |
| Permitted use | Role matrix, contract terms, privacy basis where applicable, sharing and model-use rules | Data owner + legal/privacy | Before access and at review |
| Decision reliability | Freshness, completeness, exception, calibration, and disputed-data process | Operations + data steward | During operation |
| Exit portability | Sample export, schema, context, configuration, transfer timing, deletion confirmation | Property owner + procurement | Before contract and before exit |
Failure modes
A CSV exists, but identifiers, units, relationships, calculations, and status cannot be reconstructed.
The owner can view a vendor interface but cannot reuse or transfer the underlying operational record.
Collection grows while no workflow, threshold, response, or accountable person changes.
Contract language promises control, but the team has never received a complete usable export.
Connect the decision
Evidence trail
Owner/operator considerations for direct data access and common-format export.
Research guidance on data platforms, semantic models, lineage, and governance capabilities.
Industry guidance connecting data management with smart-building architecture and lifecycle.