Before renewing an agency dashboard, ask a colleague to pick up one recommendation from last month's review without the original presenter explaining it. Can they find the exact ad, understand the evidence behind the recommendation and decide whether it still applies?
For a brand commissioning content from a China-based production team, this is a practical test of continuity. A folder of video files and a folder of performance screenshots do not necessarily connect. What needs to survive a change of agency or software is the relationship between the creative, the observation and the decision—not a copy of somebody else's interface.
A service relationship deserves more than a report date
On September 22, 2026, IAB released its Measurement Services Addendum v1.0 for public comment, addressing relationships involving advertising measurement, verification, attribution and analytics. The announced comment period ends October 22. As checked on October 16, this remains a consultation draft, not a universally binding rule.
We read the announcement, not the gated draft documents. Nothing here interprets its clauses or establishes a customer's legal ownership of data. Agreements and applicable law need their own review.
The commissioning inference is ours: if reporting is a recurring service, specify how its supporting work can be handed over, not just when the next presentation arrives. This is not an IAB requirement. It is a way to prevent a continuing content programme from losing its memory when personnel or tools change.
Motion's current product page describes reports combining creative assets and metrics, with grouping and filtering features. Checked October 16, that is a useful example of visual creative analysis—not independent proof of predictive accuracy, nor evidence of what a departing customer can export. Handoff arrangements still need to be examined separately.
Try ONCE's offline handoff rehearsal
Choose one recommendation that will influence spending, such as keeping a product-action opening in the next shoot. Give an authorised colleague who did not attend the original review only the materials agreed for handover. Ask them to recover the creative, reconstruct the observation and state what should happen next.
“Offline” means without relying on the former provider's live presentation or borrowing another person's credentials. Do not disconnect accounts, delete records or disrupt a working system for this exercise. This is a proposed acceptance method, not a test ONCE claims to have run for a client. A clearly labelled synthetic example can test the process before launch; it cannot prove a real integration works.
| Handoff item | What the receiving team needs to establish |
|---|---|
| Asset-to-ad map | Which exact edit, language and aspect ratio corresponds to each advertising record? |
| Dated observation | Which account, market and period was reviewed, when was it extracted, and what was incomplete? |
| Metric notes | What counts as a purchase or qualified lead? Which currency, timezone, attribution window and filters apply? |
| Calculation and grouping | Which export, query or manual selection produced the result? How were several ads combined into one creative group? |
| Decision memo | Why was the recommendation made, what else changed, what remains uncertain, and who reviews it next? |
A spreadsheet and a clear note may be sufficient. SQL is not a compulsory deliverable; unexplained SQL is not a handoff either. The required level of detail depends on the decision the next team must make.
Recover the edit before trying to repeat its success
Imagine an overseas lighting brand asking a new editor in China to repeat a supposedly successful opening from a desk-lamp campaign. This is a fictional example. The archive contains similar thumbnails, while several language edits share a filename. The report groups advertising records, but nobody has documented which delivered files they represent.
The immediate job is not to make another variation. Establish a stable identifier for each delivered edit and map the advertising records to those files. Leave uncertain matches unresolved rather than inferring them from thumbnail similarity. A later subtitle, price or model change needs a version record, not a silent replacement of the old asset.
That mapping establishes what people are discussing. It does not establish that a particular shot caused a purchase.
Matching the old number is not always the right acceptance test
Google's BigQuery Export documentation explains that exported event and user-level data excludes additions made to standard Analytics reports and explorations, so figures can differ from the interface. Its comparison guide identifies reporting identity, timezones and excluded streams or events as settings to check. Both were checked on October 16, 2026.
For a handoff, preserve the source surface, extraction time, calculation or selection method, relevant settings and known reasons for differences. A data specialist should establish the comparable scope before the team decides something is broken. A video editor should not have to diagnose analytics discrepancies from two unmatched columns.
Equally, keep platform-specific ROAS definitions attached to their observations before combining or ranking them. Record missing observations as missing. Preserve amendments when later data changes the picture. If an old result cannot be reconstructed, retain its snapshot and limitation rather than manufacturing an exact migration match.
Give each conclusion an honest handoff status
Reproducible: the receiving team can rebuild the result from permitted inputs and the documented method, with differences explained. Whether that method supports the business decision is a separate question.
Explainable: platform or agreed-scope constraints prevent full reconstruction, but the team can recover the original observation, conditions and reasoning. Recheck it before reuse; do not turn it into a permanent creative rule.
Archive only: a screenshot, summary or unlinked file remains. Keep it as history, not as reliable direction for a new production order.
These are operational labels, not supplier scores. They tell a buyer whether to commission asset reconciliation, analytical support or a fresh observation. More edits do not repair a missing explanation.
Price the continuing work—and the handoff
Separate connection and refresh work, creative interpretation, and archive preparation in the proposed scope. A running connector does not mean someone is explaining changes. A monthly meeting does not mean its reasoning has been preserved for another team.
The production partner should maintain accurate versions, asset mappings and creative-change records. The media or analysis partner should document metrics, calculations and limitations. The brand should name the custodian and the person making the decision. For ONCE, a managed review service would need these responsibilities to work before becoming a subscription promise. This is a service-design proposal, not an announced ONCE package.
Portability is not unrestricted transfer. Agree what can be retained, for how long, what depends on third-party material or methods, and what happens when the engagement ends. Use authorised aggregate samples for the rehearsal. Do not circulate passwords, tokens or unapproved user-level records. Assign implementation, privacy and contractual questions to qualified people rather than treating a handover checklist as clearance.
A smaller brand without steady media activity need not buy a warehouse to begin. A controlled project directory containing the asset map, dated aggregate observations, metric notes and decision memos may be enough. Consider automated infrastructure when update frequency, volume and actual decisions justify it.
When scoping video production and deliverable versions with ONCE, bring one prior recommendation, not just a preferred dashboard. If the next team cannot find the edit, fix the map. If it cannot explain the number, fix the documentation. If conditions have changed, revisit the judgment. Make one decision transferable before trying to transfer an entire reporting system.
