Programmatic advertising doesn't have a data shortage. It has a language problem.
A campaign can begin with a very simple business objective: launch a new car, acquire players for a game, introduce a new streaming service or drive subscriptions. Once that campaign enters the advertising ecosystem, however, the simplicity disappears remarkably quickly.
The advertiser may work with several agencies or technology partners. Those partners may activate through different DSPs, publishers and supply paths. The same campaign may simultaneously run across CTV, social, mobile, digital out-of-home (DOOH) and online video. Each system creates its own campaign structures, placement names, IDs and reporting conventions.
Six weeks later, when everyone sits down to understand what actually happened, a surprising amount of time can be spent simply establishing whether everyone is talking about the same campaign.
This is why I think IAB's Campaign Data Standards 1.0 is worth paying attention to.
Released for public comment in May 2026 as the first major deliverable from Project Eidos, the initiative attempts to create a common structure for campaign and placement-level data across the digital advertising ecosystem.
It isn't another measurement platform and it isn't another attribution methodology. In many ways, the idea is much more basic: can advertisers, agencies, publishers and technology platforms agree on a more consistent language for describing the advertising they are buying, selling and measuring?
That sounds simple. In practice, it addresses a problem that has existed in programmatic for years.
Consider a gaming company launching a new title.
The advertiser may not want to give the entire acquisition budget to one technology partner. It might allocate £500,000 to a proven vendor that has consistently delivered performance, £200,000 to another established partner and £50,000 each to two newer vendors it wants to evaluate.
This is perfectly rational. Advertisers want competition, price efficiency and the ability to test new partners rather than becoming dependent on a single source of performance.
But what happens to the data?
What began inside the gaming company as one campaign can become multiple campaigns as soon as it enters the buying ecosystem. Each vendor may create its own campaign IDs, insertion orders, line items, audiences and placement names. Those vendors may then activate through different DSPs, SSPs, exchanges and publisher relationships, and in some cases there may be additional intermediaries between the advertiser and the ultimate inventory.
Commercially, everyone may still be executing the same advertiser brief. Technically, the campaign has fragmented.
When the results come back, the advertiser wants to answer a relatively straightforward question: what did each partner actually contribute?
Before answering that question, however, teams may first need to map the different vendors' reporting structures back to the original campaign. Even after doing that, they may discover that partners classified similar inventory or placements differently.
There is another complication. Vendor B may appear cheaper than Vendor A, but perhaps both accessed some of the same audiences or inventory through different supply paths. Vendor C might appear to have delivered incremental performance, but establishing whether it genuinely added something new requires considerably more than looking at five different dashboards.
Campaign Data Standards won't solve attribution or prove incrementality. What it can potentially do is preserve a more consistent campaign structure as the campaign moves through different participants.
If the original campaign, objectives, formats and placement classifications can be represented more consistently across those systems, the advertiser starts with a cleaner dataset when the results return.
Instead of spending so much time asking "How do I map these five vendors' reports back to my original campaign?", more time can be spent answering the question that actually matters: "What did each partner contribute, and was that contribution incremental?"
This becomes especially important when advertisers deliberately use smaller budgets to test new partners. Experimentation is only useful when the results can be compared on a reasonably consistent foundation.
The same fragmentation appears in another form when a campaign crosses media channels.
Imagine a major automotive company launching a new model across the UK. The marketing strategy might include CTV, social, DOOH, online video and mobile, with all of those channels contributing to the same overall launch.
From the brand's perspective, this is one campaign. Inside the advertising ecosystem, it can rapidly become five different worlds.
CTV might be responsible for broad reach and storytelling, introducing the design, technology and positioning of the new vehicle. DOOH might build high-impact awareness around major cities, commuter routes and dealerships. Social could generate engagement and consideration, while mobile and display might bring consumers towards the manufacturer's website, vehicle configurator or test-drive booking page.
The problem starts when somebody asks: "Which channel performed best?" It's a perfectly understandable question, but it can also be the wrong question.
A CTV impression isn't doing the same job as a mobile click. Someone seeing a large DOOH execution on their commute isn't generating the same immediate signal as someone interacting with an Instagram ad. A consumer might see the car on CTV, encounter it again on DOOH, watch a review on social and only several days later search for the model and visit the manufacturer's website.
Which channel deserves the credit? More importantly, should we even expect them to produce similar results?
This is where I think the conversation around cross-channel measurement sometimes becomes confused. Consistency shouldn't mean expecting every channel to produce the same result. It should mean having a consistent framework for understanding the different contribution each channel makes towards the same business objective.
A common campaign-data structure can help create that foundation. If CTV, social, DOOH, mobile and video executions retain consistent information about the underlying campaign, objectives, formats and placements, it becomes easier to bring those datasets together and understand them as components of the same marketing strategy.
That doesn't suddenly make the channels equivalent. Nor should it. CTV and DOOH might primarily be evaluated around reach, frequency and brand outcomes. Social may contribute engagement and consideration signals. Mobile might provide stronger behavioural signals such as site visits, configurator activity or test-drive intent.
The value of standardisation is not forcing all of those signals into one KPI. It is helping preserve the relationship between the different activities and the common campaign they are serving.
At its simplest, Campaign Data Standards 1.0 attempts to establish a baseline structure for describing campaign and placement-level information consistently across the advertising ecosystem.
The proposed framework introduces greater consistency around areas such as ad types, placement classifications and campaign data fields, with the intention of creating something that can operate across platforms and media environments.
Importantly, it is designed as an interoperable layer rather than a replacement advertising platform. Existing DSPs, SSPs, publisher systems, agency platforms and measurement providers can potentially adopt the standard while continuing to operate their own technology.
That distinction matters because the industry doesn't need another database containing campaign information. It needs the systems already holding that information to understand each other better.
Think again about how a typical campaign travels: advertiser → agency → DSP → exchange/SSP → publisher → measurement provider.
At every handoff, campaign metadata can be mapped, renamed or transformed. Once multiple vendors and multiple channels are involved, the number of possible variations grows quickly.
Individually, these differences can appear trivial. Across thousands of campaigns, placements, creatives and partners, they become a significant operational cost.
The cost appears in engineering resources, reporting teams, custom integrations, spreadsheets, mapping tables and delayed analysis. More importantly, it creates uncertainty around the information being used to make media investment decisions.
There is something slightly ironic about an industry capable of moving enormous volumes of bid requests between systems in milliseconds while still sometimes requiring days to reconcile the reporting afterwards.
The timing is particularly interesting because advertising measurement is moving in exactly the opposite direction: it is becoming more sophisticated.
Advertisers are investing in marketing mix modelling, incrementality testing, clean rooms, attention measurement, cross-platform attribution and increasingly AI-driven optimisation. These are important developments, but sophisticated models still depend on the quality of the information entering them.
If a measurement system first has to determine that "CTV_Premium_01", "StreamingVideo", "OTT-TV" and another partner's proprietary classification are describing broadly comparable campaign activity, we are asking the measurement layer to solve a data-normalisation problem before it can solve the marketing problem.
This becomes even more relevant with AI. Machine learning systems are very good at identifying patterns across large datasets, but inconsistent classifications create unnecessary noise. If essentially the same campaign element is represented differently across multiple platforms, part of the model's job becomes interpreting those inconsistencies.
Consistent campaign taxonomies therefore shouldn't be viewed simply as reporting infrastructure. As more optimisation and analysis becomes automated, they could increasingly become part of the infrastructure on which AI-driven advertising operates.
It is important not to turn Campaign Data Standards into something it isn't. The standard won't solve identity loss or cross-device deduplication. It won't decide which attribution methodology is correct, determine whether a consumer exposed to CTV and DOOH should be counted twice, or establish whether the £50,000 gaming vendor genuinely delivered incremental users rather than reaching consumers who would have converted anyway.
Those remain difficult measurement questions.
It also won't make every media channel directly comparable. A DOOH exposure, CTV impression, social interaction and mobile click are fundamentally different consumer experiences and should not suddenly be treated as equivalent because they share a taxonomy.
What standardisation can do is give the industry a cleaner starting point from which to tackle those harder questions. Before arguing about which measurement methodology produces the right answer, it helps enormously if everyone has greater consistency in describing the underlying campaign.
There is a broader strategic implication here. Programmatic has spent much of the last decade adding capability: more channels, more signals, more audience segments, more optimisation and more measurement. Complexity has accumulated faster than standardisation.
Campaign Data Standards represents something of a return to fundamentals. Before we add another optimisation layer, perhaps we should first make sure we're consistently describing what we're buying and measuring.
If the standard gains adoption, interoperability could also become a more meaningful product advantage. Platforms that make it easy for customers to move campaign information between systems without complicated mapping exercises should benefit from faster integrations and lower operational overhead.
It could also change where measurement companies differentiate. A surprising amount of measurement work today still involves data plumbing: cleaning datasets, mapping fields, reconciling campaign structures and normalising outputs. If standards reduce even some of that work, measurement providers can spend more of their energy on methodology and insight.
The conversation can move from "Do these datasets actually line up?" towards the questions advertisers really want answered: "What drove incremental performance? How confident are we? And where should the next pound of budget go?"
The public-comment period for Campaign Data Standards 1.0 closed in June 2026, with IAB indicating that the final version is expected later in 2026. Publishing a standard, however, is the easy part. The real test is whether the ecosystem adopts it.
The value increases significantly when campaign information can retain a common structure across advertisers, agencies, buying platforms, supply platforms, publishers and measurement companies. That will take time. Legacy systems won't disappear, platforms will implement standards at different speeds, and mappings between old and new structures will remain necessary.
But if adoption reaches sufficient scale, I would expect support for common campaign-data standards to increasingly become part of technology evaluations and integration discussions. That's the point at which a taxonomy stops being documentation and starts becoming infrastructure.
The reason I find Campaign Data Standards 1.0 interesting isn't the taxonomy itself. It's the problem underneath it.
Whether it's a gaming advertiser splitting one campaign between five vendors or an automotive brand launching across CTV, social, DOOH and mobile, the pattern is similar. One business objective enters the advertising ecosystem and becomes fragmented across platforms, partners and channels.
Some of that fragmentation is necessary. Different vendors should compete. Different channels should do different jobs. Different measurement companies should have different methodologies. The problem is unnecessary fragmentation — the part created simply because our systems don't describe the same advertising activity consistently.
Campaign Data Standards won't solve measurement, attribution or incrementality on its own. But it could remove some of the structural noise that makes those problems harder than they already are.
What we do need is a common enough language that when those systems disagree, we know they are disagreeing about the measurement — not simply describing the underlying campaign differently.