A Guide to Data Curation in Digital Advertising

Glossary of Key Terms

To navigate the evolving landscape of curation, it is important to establish a shared vocabulary. Here are some key definitions that help to better understand curation.

• Data Curation: The intelligent selection, enrichment, and packaging of digital media inventory using trusted datasets and contextual signals to create higher-quality, more transparent, and more effective advertising opportunities.
• Supply-Side Platform (SSP): Technology platforms that help publishers manage, optimise, and sell their digital advertising inventory across multiple demand sources, primarily focused on yield optimisation.
• Demand-Side Platform (DSP): Technology platforms that enable advertisers and agencies to purchase digital advertising inventory programmatically across multiple ad exchanges and supply sources.
• Private Marketplace (PMP): Invitation-only auctions where premium publishers offer their inventory to selected advertisers at negotiated terms, providing greater control and transparency than open auctions.
• Curated Marketplaces: Advanced programmatic environments where inventory is intelligently selected, filtered, and packaged using data signals, AI, or both, to align supply with specific buyer objectives.
• Deal ID: A unique identifier that enables programmatic access to specific inventory packages or curated marketplaces, allowing buyers to purchase pre-negotiated advertising opportunities.
• Bidstream: The real-time flow of bid requests and responses in programmatic advertising, containing information about available ad inventory and audience data.
• Semantic Targeting: Advanced contextual advertising that uses machine learning to understand the deeper meaning and sentiment of web page content, beyond simple keyword matching.
• Contextual Advertising: A targeting approach that serves advertisements based on the content and context of web pages rather than user behaviour or personal data.
• Identity Solutions: Technologies that enable audience targeting and measurement in digital advertising, including both deterministic (based on confirmed user attributes) and probabilistic (based on inferred attributes) approaches.
• ID-less Solutions: Privacy-compliant targeting methods that use anonymous, aggregated data rather than personal identifiers, often employing cohort-based audience segments.
• Supply Chain/Supply Path: The sequence of technology platforms and intermediaries through which digital advertising inventory flows from publisher to advertiser, affecting both cost and transparency.

The Basics: What is Data Curation?

Defining Data Curation in Digital Advertising

Data curation is the process of prioritising quality over quantity. Rather than casting the widest possible net and hoping for relevant impressions, curation involves the real-time selection and packaging of media inventory. This inventory is enriched with data signals to drive better outcomes for advertisers, ensuring that every impression serves a specific purpose.

"Data curation, in a nutshell, is the process of intelligently selecting, enriching, and packaging digital media supply using trusted datasets and contextual signals, therefore creating higher quality, more transparent, and more effective media opportunities."
— Rhys Denny, @curate

At its core, data curation is about optimisation across data signals to deliver more relevant impressions. In many cases, it leverages advanced data science, artificial intelligence, and machine learning to process and filter inventory at the supply source. This helps improve the quality of inventory send to demand-side platforms receive.

According to Onetag, this process is defined as "the real-time intelligent selection and packaging of media inventory, enriched with data signals to drive better outcomes for advertisers."

@curate’s Rhys Denny adds a crucial distinction regarding the industry's evolution: "Where programmatic once prioritised scale, curation prioritises transparency and relevance - ensuring every impression is bought with intent and insight, not waste and guesswork."

The Shift from Demand-Side to Supply-Side Curation

Historically, curation was primarily the domain of DSPs, operating on the demand side of the programmatic ecosystem. However, the industry has witnessed a significant shift towards supply-side curation.

According to Onetag, the adoption of sell-side data curation provides the demand side "with the assist it needs,"  delivering greater scale and efficiency.

This shift is driven by the need for "sell-side data decisioning" which delivers greater:

• Scale: Far higher addressability for targeting due to signal loss when applied on the demand side.
• Speed: Quicker, more intelligent impression selection and pricing.
• Precision: Incorporating the array of ad placement-level seller metadata—the opposite of traditional SSP "dumb pipes".
• Quality: Access to brand-safe performant inventory, eliminating programmatic waste in the bidstream.
• Transparency: Access to clean, direct supply paths.

Key Components of Data Curation

Effective data curation is not a single action but a continuous cycle. According to @curate, it encompasses several critical components:

1. Discovery: Identifying quality inventory, data partners, and contextual signals.
2. Packaging: Combining these into curated deals or marketplaces aligned with campaign goals.
3. Activation: Deploying those deals programmatically via DSPs.
4. Measurement and Optimisation: Analysing performance, refining data inputs, and maintaining quality control.

As Rhys Denny explains, "In short, it’s an always-on feedback loop between data, media, and performance."

What Good Data Curation Looks Like

Quality data curation is characterised by transparency, measurability, and an outcome-driven focus. It should simplify rather than complicate the media buying process. @curate, define curation using five pillars: transparency, efficiency, relevancy, privacy-safe, and sustainable. Who is involved, how and where the ad is delivered, and that the ad is delivered in a privacy-safe way, are all crucial parts of curating effectively.

Onetag adds that from a provider perspective, good curation involves "meaningful data"  (deterministic, probabilistic, contextual) and "direct access to real premium publisher inventory."  Crucially, it requires "the ability to optimise in real time based on the live, changing nature of programmatic inventory and buyer KPIs."

How Data Curation Works

The Curation Process

Data curation operates as a sophisticated, real-time process that transforms raw programmatic inventory into intelligent, targeted advertising opportunities. The process begins with the collection and analysis of multiple data signals from various sources, including publisher metadata, audience insights, contextual information, and quality indicators.

Modern curation platforms process these signals using advanced algorithms that evaluate each impression opportunity against specific criteria. This evaluation considers factors such as content quality, audience relevance, viewability likelihood, brand safety parameters, and predicted performance outcomes. This ensures that before a bid is ever placed, the inventory has been vetted for quality and relevance.

The Role of AI and Machine Learning

Artificial intelligence and machine learning form the backbone of modern curation technology. These systems continuously analyse vast amounts of historical and real-time data to identify patterns and optimise decision-making processes.

"For @curate, AI is the glue binding these layers together, turning fragmented datasets into meaningful, moment-level insight."
— Rhys Denny, @curate

Onetag reinforces this, noting that for curation platforms to achieve their operational potential, it requires "strong curation technology, including AI that has learnt from a wealth of historical performance data and makes real-time decisions on live data."

Real-Time Optimisation and Traffic Shaping

One of the most powerful uses of modern data curation is its ability to shape traffic in real-time. This process involves the selective filtering and prioritisation of advertising inventory based on specific campaign objectives and quality parameters.

Traffic shaping delivers tangible efficiency gains. Onetag reports that this approach can result in "50% fewer bids required to win impressions."  Crucially, for environmentally conscious brands, this efficiency reduces carbon emissions by "94% less... than the programmatic industry average."

Horizontal vs. Vertical Curation

The industry recognises two primary approaches to data curation, each serving different strategic needs:

• Horizontal Curation: These are managed curation services that work across multiple exchanges and supply sources. They provide a layer of intelligence and optimisation on top of existing programmatic infrastructure, offering broad market access and unified reporting.
• Vertical Curation: These platforms are integrated directly into exchange infrastructure. They offer direct access to curated inventory with tighter control over quality and performance parameters. Onetag highlights that "curation platforms are vertically integrated directly on top of leading Exchanges to give buyers and data providers direct access to create the IDs they need and activate immediately in their DSPs of choice."

Types of Data Used in Curation

Identity Solutions

As the industry moves away from cookies, identity solutions have become paramount.

• Deterministic Identity Solutions: Rely on confirmed consumer attributes and provide highly accurate audience targeting. However, scale can be limited in open web environments.
• Probabilistic Identity Solutions: Expand targeting scale by incorporating temporary attributes such as device characteristics and connection methods. These solutions use statistical modelling to infer user identity and interests.

Leading curation platforms enable a blended use of identity solutions.

ID-less Solutions and Cohort-Based Audiences

ID-less solutions represent the future of privacy-compliant advertising, using anonymous, aggregated data rather than personal identifiers. These approaches build audience cohorts over time based on interests, locations, behaviours, and actions.

Supply-side curation enables ID-less data activation at a significantly greater scale than demand-side approaches. Onetag points out that "sell-side curation already provides privacy-safe targeting at scale and is effective across markets operating with GDPR and local country-specific rules."

Contextual and Semantic Data

Contextual targeting has long been a cornerstone of digital advertising, but semantic technology is revolutionising its effectiveness. Whilst traditional contextual advertising relies on site classification and keyword matching, semantic technology uses machine learning to understand the deeper meaning and sentiment of web page content. This enhanced understanding enables more precise content categorisation, improved brand safety, and better alignment between advertising messages and content context.

Media Quality Signals

Media quality represents a critical component of effective curation. Curation platforms analyse metrics that indicate the likelihood of campaign success, such as attention metrics, viewability scores, and ad placement quality. Onetag notes that sell-side decisioning incorporates an array of cookieless data signals including "publisher, quality, contextual/semantic, attention, first-party and more."

Publisher First-Party Data

Publisher first-party data represents one of the most valuable assets in the curation ecosystem. This data includes audience insights and content performance metrics that publishers have collected directly from their audiences with proper consent. Curation platforms can activate this data securely and at scale, enabling publishers to monetise their audience insights whilst providing advertisers with high-quality, consented targeting options.

Sustainability Data

Environmental consciousness is increasingly important in digital advertising, with sustainability data becoming a key curation signal. Carbon scoring data enables advertisers to understand and optimise the environmental impact of their campaigns.

@curate highlights that "Carbon measurement APIs and emissions dashboards now make it possible to assess the environmental impact of every impression, allowing buyers to curate not only for performance, but for planet."

Practical Applications and Use Cases

Audience Refinement and Targeting

Data curation excels at audience refinement, moving beyond simple demographic targeting to incorporate behavioural, contextual, and quality signals. This multi-dimensional approach enables more precise audience identification and engagement.

• Onetag reports that this approach can deliver a "50%+ improvement in data targeting addressability."
• @curate describes this as "layering quality data (contextual, attention, or consented identity) to target users based on real intent, not assumed behaviour."

Supply Quality Control

One of the most immediate benefits of data curation is its ability to filter and improve supply quality. By identifying and removing problematic inventory sources, curation platforms ensure that advertiser budgets are invested in high-quality, brand-safe environments.

• @curate identifies this as "reducing duplication, fraud, and carbon waste by filtering out low-value inventory."
• Onetag notes that open exchange curation of seller metadata solves concerns that "the open exchange is unsafe or creates advertising waste."

Brand Safety and Fraud Prevention

Curation platforms integrate multiple brand safety and fraud prevention technologies. Onetag explains that for leading platforms, "inventory is analysed, filtered, scored and optimised in real time to ensure media quality and performance," often integrated with third-party measurement companies for verification.

Retail Media Networks

The growth of retail media networks has created new opportunities for data curation. Curation helps advertisers navigate increasingly crowded retail environments by enabling precise targeting and better product-content matching.

Benefits for All Stakeholders

• Advertisers and Brands: Gain "efficiency and outcome-based performance" (@curate) and achieve "faster and improved campaign outcomes" (Onetag).
• Agencies: Can "unlock scalable, privacy-safe differentiation for clients" (@curate) through operational efficiencies.
• Publishers: Gain significant value by "increasing yield through better matching of demand and context" (@curate) and "increasing demand by presenting their quality inventory more effectively" (Onetag).
• DSPs and Technology Platforms: Benefit from "simplified pipes and improved interoperability" (@curate), leading to improved system efficiency.

Industry Standards and Best Practices

Current Gaps in Consistency and Terminology

The rapid growth of data curation has created several industry-wide challenges. @curate identifies several gaps:

• Lack of a consistent definition: "Curation" is used loosely to describe anything from data packaging to private marketplaces.
• Variable transparency: Not all curated deals disclose who is involved or how value is distributed.
• No universal framework: There is no standard for measuring curated supply quality or sustainability impact.

Quality and Trustworthiness Measures

Establishing trust in curated data requires attention to three key dimensions. @curate suggests looking at:

1. Provenance: Is the data sourced transparently and consented?
2. Performance: Does it drive measurable uplift without inflating cost or carbon?
3. Partnership: Are the pipes clear, and do all parties see value?

Onetag suggests that the definition of good curation can serve as a "checklist for evaluating curation partners to sense check they offer a true service."

Common Myths and Misconceptions

The Future of Data Curation

Evolution in a Post-Cookie, Consent-Driven Ecosystem

As the digital advertising industry transitions away from third-party cookies, data curation becomes increasingly vital.

• @curate asserts: "Curation will become the new default for buying media... giving brands confidence in a world where personal identifiers are fading."
• Onetag supports this: The scope of curation will "continue to grow as demand side signal loss worsens and sell-side technology advances."

Technological Innovations

• Artificial Intelligence: AI is "transforming how signals are discovered, interpreted, and connected," allowing curation to move "beyond manual packaging into adaptive, impression-level decisioning" (@curate).
• Data Clean Rooms: When paired with curation, clean rooms "unlock the ability to apply first-party and contextual data safely and selectively."
• Interoperability Standards: As interoperability extends into broader business systems, "curated deals can move seamlessly across platforms, DSPs, and marketplaces."

Opportunities for Innovation

Onetag sees opportunities in "bringing business outcomes more into sell-side data decisioning and accelerating curation of ad creative alongside media." @curate highlights opportunities in:

• Automation: Using AI to curate dynamically.
• Sustainability: Measuring and optimising the carbon cost of each impression.
• Outcome-based trading: Moving from CPMs to cost-per-outcome or carbon-adjusted pricing models.
• Education: Helping the market understand that curation is an evolution, not just a layer.

Potential Risks and Barriers

Despite the clear benefits, several challenges could impede the full realisation of curation's potential. @curate identifies the following risks:

• Misuse of the term "curation" leading to market confusion.
• Over-reliance on legacy data or black-box AI models.
• Inertia—buyers still defaulting to "scale first" rather than "quality first."
• Lack of industry standardisation on how curated data should be valued and measured.

Conclusion

Data curation represents a step forward in open web advertising, moving the industry from quantity-focused to quality-driven approaches. By intelligently selecting, enriching, and packaging advertising inventory using multiple data signals, curation addresses many of the structural challenges facing modern programmatic advertising.

The benefits are clear and measurable: improved campaign performance, enhanced transparency, reduced waste, and better outcomes for all stakeholders. As privacy regulations tighten and environmental concerns grow, curation provides a sustainable path forward that balances performance with responsibility.

Success in implementing data curation requires understanding its core principles, selecting trustworthy partners, and maintaining a focus on transparency and measurable outcomes. The technology will continue evolving, but the fundamental value proposition-delivering the right advertisement to the right audience in the right context-remains constant.