Imagine navigating routes with a GPS that becomes less accurate every month. It still offers directions, but you can no longer be certain whether it is leading you to the fastest route. That is the predicament the digital advertising industry finds itself in today. For years, every click could be tracked, every conversion attributed, and every rupee accounted for. But attribution (the system that connects marketing activity to business outcomes) now seems to be losing credibility as privacy regulations, third-party cookie depreciation, AI-generated traffic and platform silos obscure the digital customer journey.
The data still exists, but the picture it paints is becoming increasingly fragmented, making it difficult to answer a simple question: what actually worked?
Highlighting this shift, Vivek Das, Chief Digital Officer, Madison Media says, “The signals that traditional attribution depended on—third-party cookies, cross-device tracking, pixel-level data flows—are all under simultaneous structural pressure.”
Pointing out what this actually means for digital advertising, he says, “Industry data suggests that ad platform conversion reports and actual CRM outcomes can differ by 20–30 per cent or more, because fragmented consent, browser restrictions, and cross-device journeys mean the measurement system is only seeing part of the picture. Retargeting audiences are shrinking. Cross-device journeys are breaking attribution chains. A consumer who discovers a brand on their phone and converts on their laptop can look like two different people to a last-click model.”
Last-click attribution, for long, served as the default yardstick for digital advertising, assigning full credit for a conversion to the final touchpoint before a purchase or sign-up. This is becoming increasingly difficult to execute. Chetan Asher, Founder & CEO, Tonic Worldwide, believes that last-click was always a convenient fiction. “It credited whoever stood closest to the conversion while ignoring everything that actually built the case: upper-funnel media, brand activity, creator content and offline touchpoints. It didn't die of old age. It died because the tracking infrastructure that made it look credible fell apart. It may still have some operational value for campaign optimisation, but it is no longer a valid framework for measuring marketing effectiveness,” he remarks.
The fallout extends beyond attribution models to the numbers themselves. Marketers are seeing a growing gap between reported and verifiable conversions, which means that the real challenge is not just measuring performance, but understanding what percentage of that performance can be validated. Jyoti Chugh Bhatia, Group Director, Gozoop Creative says, “It is difficult to put an exact number on it, but I'd estimate that for most brands, somewhere between 50–70 per cent of conversions can be verified. But there is still a lot happening that marketers can't fully see, whether it is someone switching devices, organically interacting with content, or being influenced by channels that aren't directly trackable.”
Asher believes that only about one-third of conversions are deterministically verifiable, end-to-end. “The rest is modelled, inferred or platform-reported, which means trusting the same player who is selling you the media. That conflict of interest is the real problem,” he says, further clarifying that for brands with strong CRM ecosystems and robust first-party data strategies, verification rates are meaningfully higher.
The Privacy-Measurement Trade-Off
A major factor behind the attribution crisis in digital advertising is that the Internet itself is becoming less open. The biggest digital and social media platforms have evolved into walled gardens, where advertisers can see what happens inside each ecosystem but struggle to connect those interactions to the wider customer journey. Layer privacy-first regulations such as the DPDP Act with tighter consent requirements and browser-led restrictions on user tracking, and the gaps in attribution only widen.
Das emphasises, “Privacy regulations were meant to redistribute data power, but in practice, they have concentrated it further. When third-party cookies erode and cross-site tracking becomes legally constrained, the entities that already own rich first-party data at scale–Google, Meta, Amazon—become more valuable to advertisers, not less. The walls get higher precisely when outside visibility is reduced.”
He further points out that nearly 70 per cent of digital advertising spend already flows through a small number of closed platform ecosystems, where the platform is simultaneously the media owner, the data custodian, the measurement provider, and the optimisation engine.
However, in India, data localisation requirements and consent frameworks of DPDP Act are changing how platforms can collect and process user data, making it important for marketers to build independent measurement infrastructure that lets them validate what the platforms report. Yasin Hamidani, Director, Media Care Brand Solutions states, “The biggest shift is the growing importance of first-party data. Brands can no longer rely on unrestricted third-party tracking and audience sharing. As a result, performance marketing is becoming more trust-driven, with greater emphasis on owned data ecosystems and transparent consumer relationships.”
When AI Enters the Funnel
If privacy has reduced the amount of data marketers can access, AI is raising fresh questions about the quality of the data that remains. Automated agents are generating traffic, and AI-powered search is altering discovery journeys. These shifts are making it difficult for marketers to determine where a conversion came from, and whether the signals they are measuring reflect genuine consumer behaviour.
Bhatia states, “AI has made campaigns smarter, but it has also added another layer of complexity to measurement. Platforms are increasingly making decisions on our behalf, whether it is audience targeting, bidding, placements or creative optimisation. While that often improves performance, it can make it harder to understand exactly what contributed to the result.”
According to Das, the first issue is AI-generated and bot-driven invalid traffic. “Bots now account for over a third of all web traffic, according to Imperva's latest bot report. The WFA estimates that more than $50 billion was wasted globally on invalid traffic in 2025 alone. This inflates impressions, corrupts click-through data, pollutes conversion signals, and distorts the optimisation logic of automated bidding systems, because the platforms are training their algorithms on data that includes non-human behaviour despite all the safeguards.”
Asher backs the observation, adding, “AI has made modelling sharper and optimisation more effective, but it has also flooded the funnel with traffic that behaves like a person when it is not one. When part of your audience is non-human, your CTRs and funnels lie to you while looking perfectly healthy.”
The second challenge stems from AI-mediated discovery. Das explains that when a consumer asks an AI assistant for a product recommendation and purchases based on that suggestion, that chain of influence is largely invisible to conventional attribution models. As a result, upper-funnel brand-building efforts often go unrecognised, while the final click receives disproportionate credit.
The New Performance Playbook
The common thread running through privacy regulations, disappearance of cookies, walled gardens and AI is the gradual erosion of verifiable performance metrics. Marketers can no longer expect to follow every consumer interaction from impression to purchase, leading to a re-evaluation of key performance indicators. Bhatia says, “The conversation is becoming less about which channel got the last click and more about whether marketing is contributing to long-term business growth.”
Hamidani stresses that the industry is moving towards a blended measurement approach. “Incrementality, customer lifetime value (LTV), attention metrics, engagement quality, retention rates, brand search lift, and conversion efficiency are becoming increasingly important.”
Asher highlights a four-layer approach that marketers are increasingly leaning towards: business outcomes like revenue growth and customer acquisition; incrementality testing to determine whether marketing genuinely created demand; Media Mix Modelling, which evaluates channel contribution without relying on individual-level tracking; and first-party customer intelligence, which brings data you own rather than rent.
Das backs a similar multi-layer approach, presenting the ‘Hierarchy of KPIs’—business outcomes first, brand metrics second, behavioural signals third, media metrics last. “Goodhart's Law applies ruthlessly here—when a media metric becomes a target, it stops being a good measure of growth,” he concludes.
