How ACIF Gives a CTV Creative a Durable Identity
The IAB Tech Lab’s Ad Creative ID Framework (ACIF) is a way to give a CTV creative an identity that can survive movement between companies and platforms. The important question is not simply how an ID is generated, but how registration, registry context, validation, and day-to-day handoffs work together. Tracing that lifecycle shows why a durable identity is more than a platform upload number—and why it can still disappear when the surrounding workflow does not carry it forward.
Registration Creates the Identity’s Meaning
A creative file does not become cross-platform identifiable merely because an ad server has stored it. An ad server’s local ID usually answers a local question: which object did this system receive or upload? That answer may be perfectly useful inside the system that created it, but it does not necessarily tell another platform that the object is the same advertiser’s creative. A second platform can assign a different value without either system being wrong on its own terms.
ACIF starts earlier, at ad registration. The framework describes registration as filing the ad creative with a regional ad authority under the advertiser’s name, together with the brand and product represented. Additional context may include an ad title or campaign ID. The result is a registered creative identity that can be referenced as the asset moves through the video supply chain.
The labels matter because an ID without useful meaning is difficult to apply in an automated decision. A buyer or publisher may need to know not only that two records refer to the same file, but also which advertiser, brand, and product the file represents. Those labels can support rules such as competitive separation, where a system tries to avoid placing one product beside a competing product. They can also make a later report intelligible to people who did not create the original ad tag.
This is the first state in the lifecycle: an unregistered creative becomes a registered creative with an identity and associated metadata. The registry is not just a storage location for a number. It provides the authority that gives the number a place in a defined naming system. ACIF identifies regional registries such as AD-ID in the United States, Clearcast in the United Kingdom, and CIM in Belgium as examples of organizations that maintain records and assign unique IDs. The point of naming regional registries is not to imply that one registry covers every market. It is to recognize that a globally distributed supply chain may encounter several authorities and must know how to interpret the identity it receives.
That regional structure is an important design constraint. If a receiving system sees only an opaque string, it cannot safely assume that the string uses the same namespace as another string with a similar shape. A registry-aware identity gives the receiving party a way to ask, “Which authority owns this namespace, and how can its metadata be checked?” Without that context, an identifier may be unique inside its originating system but still ambiguous at the boundary between systems.
ACIF therefore describes several components that operate together. Registries maintain the creative records. Standardized use of registered IDs gives supply-chain formats a place to carry them. A directory points technology providers toward participating registries and the information needed to validate an ID. Validation gives a system a way to check whether the identifier exists and retrieve associated metadata, subject to what the relevant registry provides.
The directory is useful because the receiving party may not know the registry in advance. ACIF’s proposed directory model includes a registry URL or code as a unique reference and lookup information describing how IDs can be validated. The framework also proposes a Registry Validation API as a common request-and-response pattern. In the intended flow, an ad-tech service encounters a registered ID, finds the relevant registry through the directory, and asks the registry to confirm the ID and return basic metadata about the advertiser or brand.
That flow turns a local-looking string into a checkable reference. It does not turn every string into a verified truth. A registry authority URL can tell a system where to look; it does not prove that the ID is current, that the correct creative carries it, or that every registry exposes the same interface. ACIF describes the validation API as a proposal and allows registries to provide additional metadata according to their own services. Identity context improves the conditions for validation, but validation remains an active step.
The difference becomes clearer if the lifecycle is expressed as state changes rather than as a list of components:
The sequence has a beginning, but it does not have a single owner. The brand or agency may coordinate registration. A campaign system must store the value. A VAST generator or another transport mechanism must place it into the ad response. Ad tech and publishers must know where to find it and how to use it. A measurement service may need the same identity to attach observations to the right creative. The registry can validate the record, but it cannot force every downstream system to preserve it.
That shared ownership separates three questions that are easy to blur. First, does a registered record exist? The registry and its lookup path answer that question. Second, did the identity and its authority context survive the handoff into the campaign and ad response? That is a transport and workflow question. Third, did the receiving system actually use the identity in decisioning or logging? That is an adoption question. A positive answer to the first does not imply positive answers to the other two.
Those distinctions produce different failure states. A creative may never have been registered, so a platform-local ID is the only identifier available. It may have been registered correctly, but the value may not have reached the campaign system or tag generator. Or the value may have reached the response while the receiving platform ignored it and continued to organize records by its own label. The symptoms can look similar in a report—unmatched rows, inconsistent advertiser names, or apparently separate versions of one spot—but the remedy is different in each case. Registry participation cannot repair a dropped field, and a field cannot create a validation path that the receiving system does not know how to use.
This is also why a local identifier should not be treated as useless. It can be the right key for a platform’s own storage and operations. The problem arises when that local key is mistaken for a shared identity at the next boundary. A durable CTV workflow can preserve local references for local work while carrying the registered identity as the common reference for cross-company decisions. The two values answer different questions; replacing one with the other hides that difference.
Validation has a similar boundary. When a receiving system finds the registry and requests metadata, the lookup can give meaning to an otherwise opaque value and support a decision about the advertiser, brand, or product. It does not by itself prove that every downstream copy of the video is identical, that the ad rendered, or that the resulting exposure should be counted. It is a check on identity context. The later measurement system still has to establish what happened at playback and how confidently that event can be reported.
Persistence Depends on Workflow and Handoffs
The framework calls workflow the most common barrier to ad registration. That observation changes how “support” should be interpreted. A company can agree that persistent creative identity is valuable and still lose the value if no one decides who registers the ad, when registration occurs, where the ID is stored, or how it reaches the next interface.
Consider a campaign in which a creative is approved by an agency and then sent to a platform partner. Several transitions must line up. Someone needs to register the creative at the chosen point in campaign development. The registered value must be entered into the campaign-management system. The system or its operator must pass it to the software that generates the ad response. The partner receiving the response must retain it rather than replacing it with a local upload ID. If a later system needs to validate the identity, it must also retain the registry context or have a reliable way to recover it.
None of these steps is conceptually difficult, but each is a place where the identity can be lost. A creative may be registered after the campaign record has already been created. An account representative may know the campaign by a familiar internal label and omit the registered value. A tag generator may have no field for it. A publisher’s player may ignore the relevant field. An SSAI provider may preserve the video stream while failing to preserve the metadata path. The failure can look like a measurement discrepancy later, even though the first break occurred during campaign setup or transport.
That is why ACIF’s recommended workflow is organizational as well as technical. The framework suggests assigning a lead role, deciding when registration happens, documenting naming conventions, specifying how the ID enters campaign software, and automating transfer into a transport mechanism such as VAST where possible. Staff who handle creative need to know where the ID is, how to enter it, and how to pass it along. Partners should receive onboarding information about how the ID is supplied and used. These are controls for continuity, not administrative decoration.
The responsibilities are distributed across the supply chain. Regional registries need a way to participate and explain validation. Publishers, networks, and distributors need to recognize registered IDs in ad responses and update decisioning logic to use them. Ad-tech providers need to look up registries, validate inbound identities, and return or retain the metadata that supports decisions and reconciliation. Brands and agencies need to coordinate with creative and demand partners about who registers each asset. Creative developers may be the designated registrant and must deliver the ID with the creative when the asset changes hands.
The framework’s role-specific guidance also exposes a hidden coupling: registration has no downstream effect if the systems that make decisions never use the identity. Conversely, a publisher or measurement vendor cannot reliably use the identity if the brand and agency never register the asset or if the tag generator drops it. Persistence is an end-to-end property produced by several local responsibilities.
When those responsibilities align, the immediate benefits can appear before the campaign report. A decisioning system can recognize that an ad arriving from different sources is the same ad. It can apply frequency-capping logic to the creative rather than to a set of unrelated local IDs. It can use the registered advertiser, brand, and product labels to support competitive separation. Later, the same identity can serve as a common hook for reconciling delivery logs and associating performance with the creative.
Those benefits should be stated conditionally. ACIF identifies them as use cases and intended improvements; the framework does not provide evidence that every deployment achieves them or that a particular campaign will see a measured performance lift. Registration also does not answer whether an ad rendered, whether a viewer initiated the stream, or whether the traffic was valid. It makes the identity question easier to ask consistently. The measurement question remains separate.
There is another boundary worth preserving. A registered identity usually describes a creative asset, not every possible version of a campaign experience. If an advertiser changes the material creative, changes the product represented, or creates a materially distinct asset, the parties need a clear registration and naming policy rather than assuming that an old identity covers the new one. The supplied framework discusses registering the ad creative and associated campaign metadata, but it does not establish a universal rule for every versioning scenario. That is a workflow decision that must be made deliberately.
A useful operational test follows from the lifecycle. Ask a partner to demonstrate not only that they accept a creative ID, but also where the value is stored, which system writes it into the ad response, what the receiving player or stitcher does with it, and how a validator finds the corresponding registry. Ask which metadata is returned and how an invalid or unrecognized ID is handled. The answers expose whether “support” means a field exists, a field is transported, or the identity is actually used in decisioning and reporting.
This lifecycle also explains why registry context is not an optional footnote. The ID has to remain meaningful when the creative leaves the organization that registered it. The authority information lets the next participant locate the right validation path instead of relying on local assumptions. Once that context and identity enter the ad response, the next challenge is keeping them intact across the versions and handoffs of the transport itself. That is where ACIF’s lifecycle meets VAST.