e4m Report Card 2025: Why digital marketers stopped trusting the dashboard

By any topline measure, 2025 should have been a reassuring year for Indian marketers. According to MAGNA, India’s digital advertising market grew 7.8 per cent in 2025 to approximately $16.01 billion, with digital formats accounting for roughly half of total advertising spend. 

Growth was broad-based, driven by resilient consumer demand and the continued migration of budgets toward ecommerce, quick commerce, social and video. On paper, the system looked healthy.

And yet, this may have been the year marketers stopped trusting the dashboard.

AI in Indian marketing and what actually mattered

Not because dashboards disappeared or failed outright, but because the confidence with which they reported performance began to feel increasingly detached from what businesses could actually verify. The numbers kept improving. The explanations kept thinning. And the gap between reported efficiency and perceived reality became harder to ignore.

One reason was the rapid escalation of AI-led optimisation without a corresponding improvement in explainability. By 2025, AI-driven bidding, creative rotation and targeting were no longer optional layers. They were default settings. 

Google Ads’ AI Mode and similar systems across platforms shifted optimisation away from static metrics like clicks toward continuously recalculated signals based on predicted outcomes. Performance often improved, but causality became harder to trace.

This opacity shows up clearly in industry research. Surveys conducted in 2025 found that while 68 per cent of marketers said AI-driven insights added value to their decision-making, only 13 per cent said they fully trusted those insights when allocating budgets. More than a third explicitly cited black-box decisioning as their primary concern. 

Marketers were using AI because it worked often enough to be indispensable, but they did not believe the dashboard narrative explaining why it worked.

Attribution fatigue turned that unease into active scepticism. Privacy-driven signal loss, probabilistic modelling and vendor-specific methodologies made attribution increasingly fragile. Internal audits across brands showed that minor changes in attribution logic could dramatically reassign credit across channels without any observable change in revenue. 

Dashboards showed growth. Finance teams struggled to reconcile it with cash flow. Sales teams saw little improvement in demand quality.

Even among marketers who invested heavily in advanced attribution frameworks, confidence remained uneven. B2B research in 2025 showed that 61 per cent of marketers who described themselves as “confident” in attribution used advanced modelling approaches, compared to 32 per cent among others. 

Yet even in this group, triangulation gaps persisted. 

Attribution models told one story. Business performance told another. The dashboard did not collapse. It contradicted reality politely.

Fraud sharpened this credibility crisis in ways dashboards were poorly equipped to surface. By 2025, AI-driven ad fraud had evolved well beyond obvious bot traffic. It mimicked human behaviour with high fidelity, including scroll depth, session duration and conversion-like actions. 

Read e4m deep dive into how ad fraud is distorting marketing decisions

Pixalate’s Q3 2025 APAC benchmarks documented elevated invalid traffic rates across desktop web, mobile web, mobile apps and connected TV environments in India, placing it among the more exposed markets in the region.

The scale of the problem was no longer speculative. At Goafest 2025, Hindustan Unilever’s media leadership publicly warned that 25 to 30 percent of digital advertising spend could be wasted on bots or fake traffic. This was not framed as a measurement inconvenience but as a material business risk. When invalid traffic behaves like a user, optimisation systems do not filter it out. They optimise into it. The dashboard reports success while training the machine on noise.

Global projections reinforce how structural this problem has become. Juniper Research estimates that global ad fraud losses are on track to approach $172 billion annually by 2028. In markets with rapid digital growth and fragmented supply chains, leakage rates of 10 to 12 percent are widely cited by industry bodies. 

When spend leaks at that scale, the dashboard’s authority weakens not because it is inaccurate, but because it cannot distinguish signal from contamination.

Retail media, often positioned as the solution to attribution ambiguity, complicated trust further in 2025. India’s retail media market stood at roughly $1.5 billion in 2024 and is projected by industry bodies to grow at a CAGR of over 16 percent to reach around $3.5 billion by 2030. In 2025, it became one of the fastest-growing components of digital media plans, driven by promises of deterministic data and closed-loop reporting.

Retail media

Retail media dashboards could clearly show exposure followed by purchase on the same platform. What they could not reliably demonstrate was incrementality. As retail media budgets scaled, this limitation became impossible to ignore. Brands began asking whether they were driving new demand or paying to capture demand that already existed. 

How retail media took over Indian advertising plans

Platform-level dashboards delivered platform-level truth, but not a holistic view of business impact. As the fastest-growing channel in the mix, retail media dragged dashboard scepticism into boardrooms.

Marketing Mix Modelling returned as a counterweight, but not a cure. In ecommerce and D2C contexts, MMM studies in 2025 showed that brands could lift revenues by 2 to 3 percent at the same budget by reallocating spend more intelligently. 

Is one-screen economy changing the face of advertising?

That mattered. But MMM remained slow, assumption-heavy and backward-looking. 

Even with AI enhancements, it struggled to inform live decisions in performance-heavy environments. Marketers found themselves trapped between fast dashboards they did not fully trust and slower models that arrived after decisions had already been made.

Consumer perception widened the gap further. Research published in 2025 showed that around 70 per cent of consumers could identify AI-generated elements in ads or marketing communications, but only 25 percent viewed them positively. Nearly 60 per cent expressed doubts about the accuracy or intent of AI-driven messaging. 

Dashboards continued to measure engagement and conversion. They did not measure suspicion or erosion of trust. AI-generated creative scaled faster than sentiment could keep up.

By the end of 2025, the dashboard problem was no longer about data availability. It was about credibility. Marketers were not rejecting dashboards. They were demoting them. Performance metrics became directional rather than definitive. Reported ROAS was cross-checked against revenue lift, cohort behaviour, margin pressure and repeat purchase rates.

Old questions resurfaced with urgency. 

Are we driving incremental growth or just reallocating credit? Are we optimising for what the machine can see or what the business actually needs?

The irony is that dashboards did not become worse in 2025. They became more sophisticated. What changed was the ecosystem they were trying to describe. 

Advertising became more automated, more fragmented and more polluted at the same time. Dashboards continued to speak with confidence in a system that increasingly resists certainty. When the confidence of a number exceeds the confidence of the people using it, belief erodes.

That is why 2025 stands out. Not as the year dashboards failed, but as the year marketers stopped pretending they could explain reality on their own.