Why the adtech ecosystem continues to struggle with consumer data

Last month, Yahoo became the first DSP to adopt IAB Tech Lab's Data Transparency Label. Created after consultation with numerous industry bodies, the label promises a solution to a decades-old problem for marketers—the lack of visibility into what’s inside an audience segment.

However, what is notable is not just Yahoo’s push for greater data transparency, but the fact that IAB Tech Lab first came up with the Data Transparency Standard (DTS) in 2019. It has taken over half a decade for it to be adopted by a DSP.

Despite the opacity of audience segments being a longstanding issue, attempts at creating a universally applicable solution remain elusive or difficult to implement.

The trouble with audience data

While agencies and marketers have bought data since the dawn of online advertising, difficulties assessing its quality have led to billions in wasted ad dollars. An estimate on ad spend lost to invalid traffic was pegged at $71.37 billion through 2024 according to the Wasted Ad Spend Report from Lunio.

Last year, former UM privacy chief and current COO of Check My Ads, Arielle Garcia, made news for dismissing data from adtech vendors as “useless garbage” after accessing her own profile to find herself in 500 extremely contradictory audience segments.

While less harsh in his assessment, Sunil Naryani, president partnerships and product solutions, Dentsu APAC told Campaign Asia-Pacific, “Data segments are a lot like an iceberg. What media buyers see is just the tip, while beneath the surface lies a wealth of rich attributes that remain hidden. When it comes to data—especially transparency, reliability, and solutions that are available or heavily adopted—it is riddled with gaps; there are more holes than a solid foundation. It feels more like emmental than mozzarella.”  

Even as data providers offer multiple, seemingly disparate audience segments, there is a frustrating lack of definition when it comes to who or what is included. The trouble, according to Jayesh Easwaramony, founder of Spectra Global, begins with the identifiers used to define a segment.

He said, "You can define attributes in many different ways and call something a segment. But once it reaches the programmatic ecosystem, it has already been classified, and it cannot be verified anymore." The absence of granular insights means that a media agency cannot decide which of two superficially similar segments is most appropriate.

AI has compounded the problem—particularly its use in modelling techniques to bridge gaps in audience data. The result, in industry parlance, is synthetic data. What began as a process to enrich baseline data sets using small amounts of original data, is in danger of spiralling out of control, according to Hari Shankar senior VP at YMT Ads. There is little or no insight into the approaches used for modelling or the algorithms powering the process.

Making this opacity more concerning is a report from Gartner which states that by 2030, synthetic data will overshadow real data in AI models. Research from Emporia compared answers to the same set of questions from human respondents pitted against synthetic AI-generated users and found the latter group to be overly optimistic, with reactions that fell in a narrower range than those of their human counterparts.

Shankar cautioned, “Taken too far—of which there is a high probability in today’s competitive and cluttered industry—AI data sets could be embraced more than original data sets, because the former are more easily accessible and cheaper. The uncontrolled tampering of data from original sources via so-called AI modelling will one day undermine the value of original data.”

Even in the case of original data, questions linger about provenance, traceability and validity, recency, how often it was refreshed, and whether it was collected in compliance with prevailing norms surrounding privacy.

While the better data companies collect ethically sourced granular information, a lot of it could get lost in the transition to a data segment. Most programmatic buying platforms are typically only made aware of two basic data points: the data segment’s name, description, size, and CPM.

Naryani said, “Million- and billion-dollar advertising decisions are being made on these data points. It is crucial that we flip this ‘iceberg’ to reveal the rich metadata and context associated with the data segments, to instil confidence in data integrity and campaign decision making.”

Adding to the complexity in APAC are diverse markets with barriers on interoperability. Nishanth Raju, MD APAC of Lotame, a data and identity firm that was recently acquired by Publicis Groupe, said, “Localisation laws require data collected within the borders of a particular nation to be stored and processed locally, adding significant costs for international businesses.”

Moreover, companies are grappling with issues that have been endemic to less developed markets within the region: poor data quality and outdated information. The models have traditionally been demographic-based and have not kept pace with innovations around media and commerce. “Being able to come to grips with these spaces requires more sophisticated—but also more challenging—psychographic and transactional analysis”, said Raju.

Mitch Waters, senior VP client services APAC at The Trade Desk, added, “It’s a constant struggle to find data segments that are not only accurate but broad enough to drive meaningful reach and engagement. Walled gardens further complicate the landscape, acting as closed ecosystems that offer little insight into ad investment, audience targeting, or campaign performance. Without a media-buying platform that prioritises transparency, brands are often left flying blind.”

The tough path to data solutions

The data transparency standard from IAB Tech Labs was launched to address such issues by giving buyers essential information about audience segments in a standardised template, helping them make more informed choices.

Shailley Singh,​ EVP of product and COO, IAB Tech Lab said, “The goal was to create a ‘nutritional label’ for audience data—something that clearly spells out where the data comes from, how it’s collected, how it’s segmented, how recent it is, and how it’s cleaned.”

Giovanni Gardelli, VP of DSP data products at Yahoo, said, “Without this standard, buyers are left with very little information to discover and select audiences for their campaigns (usually name, owner, description, and user counts only).”

However, the slow pace of adoption points to the complexity inherent in implementing such a model. Singh acknowledged, “It isn’t as simple as flipping a switch—especially for large DSPs like Yahoo that work with a wide range of data providers. It requires time, resources, and alignment across multiple stakeholders.”

Timing too has impacted the widespread acceptance of DTS. Naryani says that it was a very good first step. However shortly after the launch of the DTS, Google made its first announcement on impending cookie deprecation in January 2020 resulting in a seismic shift in priorities. Naryani says, “Over the last five years, the entire adtech ecosystem was trying to find ways to evolve as third-party cookies came under threat of extinction. We saw almost 15 to 20 different identity or cookieless solutions coming in, moving from cookies to hashed email and other signifiers.”

Lotame’s Raju added, “Many platforms already have their own data transparency frameworks aligned with regulations and do not see an immediate need for DTS adoption, along with the significant system changes and employee training that it might require. The high bar set by global privacy laws like GDPR and CCPA has led many businesses to consider regulatory compliance ‘good enough’, with the IAB’s own Transparency and Consent Framework (TCF) filling in any leftover gaps.” 

For instance, Lotame built its proprietary Panorama ID to ensure data connectivity and addressability. Raju said, “Panorama ID does not rely on third-party cookies, which makes us an attractive option for marketers who want to get ahead of the inevitable cookieless future.” YMT Ads relies on a mix of channels including first party app usage data, telco data segments, offline shopper segments from a proprietary O2O platform, and finally its own first party interest/affinity data collected from many campaigns across multiple categories.

For overcoming the current limitations, Naryani advocates a multi-layered approach. “It’s not just about having data—it’s about having the right data with strong governance in place,” he said. “At Dentsu, we blend our proprietary strategies with datasets from trusted partners, including data providers, publishers, SSPs, DSPs, and DMPs. It’s a matter of process, compliance, and carefully curating our partners while setting robust technological guardrails.”

Why fixing data transparency matters

However, the survival of the open internet hinges on data transparency being resolved. Shankar said, “Unless there is an industry standard ID there is no hope for a reliable open web data play.”

Describing such an industry standard, he added, “It ought to encapsulate persistent signals like device ID  (instead of only cookies) and be used mandatorily by all open web publishers, managed by a single industry body—an official data DMP—which is then used by all the open web SSPs. All SSPs and vendors would then pipe themselves into that single DMP, and any further data enhancements would be done by individual vendors depending on the type of service / business model they are engaged in.” He adds ruefully, “In this industry where walled gardens call the shots, the threat of the open web dying out is very real if checks and balances aren’t established urgently by some governing body, as it is being done in the EU and as it was done originally in Australia.”

The process also involves marketers being far more engaged with the data solutions on offer and asking probing uncomfortable questions off their agencies and data partners*.

Creating a data transparency standard that delivers

In today's AI-driven landscape, DTS should require clear disclosure of whether the data segments come from actual user data or are synthetically generated, including the underlying methodologies, to enable informed and trustworthy decision-making. Naryani wishes to see retail media data incorporated into frameworks like DTS for a more complete audience assessment. He said, “Previous purchases, GMV, frequently purchased categories are attributes that retail data can really highlight. It says a lot more about purchase history, product lines, and where you can upsell or cross-sell. It could be a great evolution from an already very well thought-through process from the IAB Tech Lab.”

Yahoo expects its adoption of the IAB Tech Lab’s data transparency standard to be particularly impactful in APAC. Dan Richardson, director of data & insights, AUSEA at Yahoo, said that 34% of advertisers in Singapore ramped up ad spend on the open web, according to the Yahoo Singapore Marketers Digital Pulse study.

“Initiatives like the DTS help the industry navigate this minefield and make the open internet more attractive. Underscoring transparency, choice and control, ultimately, is performance. It’s the perennial top-of-mind objective—65% of marketers in Singapore said campaign performance and efficiency are their main focus. The cheapest click is not always best, especially if the data source reveals a mistargeted or irrelevant audience,” he added.

Singh remains optimistic that the IAB Tech Lab data transparency standard will become a pivotal solution. IAB Tech Lab and industry leaders are working to encourage more DSPs to adopt the DTS, but as Singh points out, adoption ultimately depends on market demand.

He said, “We provide the framework, tools, and education and do not enforce compliance. What drives adoption is regulatory pressure and advertiser expectations. Brands and agencies are demanding more transparency in audience data, making it harder for DSPs and other providers to ignore DTS if they want to stay competitive. As privacy laws tighten and advertisers prioritise data quality, DSPs that don’t adopt DTS may find themselves losing trust—and business.”