The Festive Season Attribution Puzzle — What Brands Learned Trying to Track Omnichannel ROI

Every year, somewhere between the first Navratri campaign and the last Christmas offer, Indian marketers run the most expensive experiment of their calendar. Budgets swell, creative output multiplies, and every channel on the media plan is asked to deliver at once. When the season ends, the dashboards light up, and almost every one of them claims the same sale.

The search team reports a surge in branded queries. The social team points to record engagement. The performance team shows a healthy return on ad spend. The retail team, meanwhile, notes that footfall in stores was the highest in years. Each report is technically accurate. Added together, they describe a festive season that sold far more than the business actually recorded.

This is the festive attribution puzzle. After several years of trying to solve it, brands are learning that the problem was never a shortage of data. It is a shortage of agreement about what the data means.

The Illusion of the Clean Funnel

The funnel is the most durable image in marketing, and the festive season is where it does the most damage. In theory, a consumer sees a film, grows curious, searches, compares, and buys. In practice, the festive shopper behaves nothing like this. She might see a television spot during a cricket broadcast, forward a product link to a cousin on WhatsApp, check prices on two marketplaces, walk into a store on a Saturday evening, and finally buy online three days later because a coupon landed in her inbox.

Which touchpoint earned the sale? Under most measurement setups, the answer is whichever one happened to be closest to the transaction. That is rarely the one that did the persuading.

The compressed nature of the season makes this worse. In a normal quarter, a brand might have weeks to observe how exposure turns into demand. During the festive run, dozens of messages land within days, from the brand and from every competitor. Cause and effect blur into a single stretch of high-intent noise, and conventional attribution models struggle to separate the signal from it.

Where Last-Click Breaks Down

Last-click attribution has been criticised for so long that criticising it has become a ritual. Yet it remains the default reporting logic in a surprising number of organisations, largely because it is simple, fast, and easy to defend in a review meeting. It rewards the channels that sit closest to the checkout, namely paid search, retargeting, and marketplace advertising, and quietly penalises everything that built the demand in the first place.

The festive season exposes this bias in a particularly costly way. Retargeting looks spectacular when intent is already at its annual peak. A shopper who has spent a week browsing televisions is going to buy one, and the retargeting banner that appears on the final day gets the credit. Brands that moved budget towards these high-performing lower-funnel tactics often found, a few quarters later, that their pool of new customers had thinned. They had been harvesting demand that someone else had planted, and eventually the field ran dry.

Multi-touch attribution was meant to fix this by distributing credit across the journey. It helped, but it introduced its own problems. The weighting rules are largely arbitrary, and the models depend on tracking that is increasingly unreliable. Privacy regulation, browser restrictions, and platform-level changes to tracking have all eroded the identifiers these systems need. A model that cannot see half the journey is not measuring the journey. It is estimating it, and presenting the estimate with unearned confidence.

The Offline Blind Spot

For all the talk of digital transformation, a large share of festive purchasing in India still happens in physical stores. Consumer durables, jewellery, apparel, and home goods all see customers who research online and complete the transaction on a shop floor. Digital dashboards, by design, record the research and miss the sale.

Brands have tried various bridges. Store-visit measurement from location data, unique offer codes redeemed at the counter, and QR-led campaigns have all helped. But each carries limitations. Location signals are directional rather than precise, coupon redemption captures only the customers who bother to use the code, and QR scans measure curiosity more reliably than intent.

Then there is the matter of dark social. A great deal of festive influence travels through private channels such as family WhatsApp groups, forwarded product links, and screenshots of offers shared among friends. This is arguably the most trusted form of recommendation in the country, and it is almost invisible to measurement. When a sale traces back to an aunt’s message in a group chat, most attribution systems record it as direct traffic and move on.

Quick commerce has added a further wrinkle. As ten-minute delivery platforms push into gifting, festive hampers, and last-minute purchases, a new layer of retail media now sits between brand and buyer. The platform owns the data, controls the shelf, and decides how much of the shopper journey the brand is allowed to see.

Walled Gardens and Competing Scorecards

Perhaps the most practical obstacle to clean attribution is that the major platforms grade their own homework. Each ad ecosystem measures conversions using its own rules, its own attribution windows, and its own definition of a view. Two platforms can legitimately claim credit for the same purchase, because each is counting a different type of interaction within a different timeframe.

The outcome is a familiar and uncomfortable exercise. At the end of the festive campaign, a marketing head lays the platform-reported conversions side by side and finds that the total exceeds actual sales by a wide margin. The gap is not fraud, and it is not carelessness. It is double counting, built into the architecture of how digital advertising reports on itself.

Marketers who recognised this early began to treat platform dashboards as a starting point for questions instead of a source of answers. The more useful comparison, many have learned, is between what the platforms claim and what the business ledger shows. When those two diverge sharply, the divergence itself is the insight.

What Brands Are Doing Differently

The most encouraging shift in recent seasons is a quiet change in the question being asked. Instead of asking which channel got the credit, sophisticated marketers are asking what would have happened without a given channel. That is the question of incrementality, and it requires a very different kind of measurement.

Incrementality testing works by deliberately withholding advertising from a portion of the audience or a set of regions, then comparing the outcomes against those exposed. Geographic holdout tests, in which a campaign runs in some cities and pauses in comparable ones, have become a favoured method because they do not rely on individual tracking. They measure the effect on sales directly, which sidesteps most of the identity problems that plague user-level models.

Marketing mix modelling, once considered the preserve of large FMCG advertisers with long data histories, is also finding a second life. Modern implementations are faster, more flexible, and less dependent on personal data. They estimate how each channel contributes to overall sales using aggregate patterns across time, which suits a landscape where granular tracking is shrinking. The trade-off is that mix models are slower to respond and demand clean, consistent inputs, so they work best as a strategic compass rather than a daily steering wheel.

The smartest teams are not choosing between these approaches. They are triangulating. Platform data offers speed, incrementality tests provide causal proof, and mix models give the long view. When the three broadly agree, confidence rises. When they disagree, the team has found something worth investigating before the next season’s budget is locked.

First-party data has become the connective tissue. Brands with strong loyalty programmes, app ecosystems, and direct customer relationships can stitch together journeys that third-party tools can no longer see. A customer who registers for a warranty, joins a rewards scheme, or buys through a brand’s own store creates a durable record. Over time, these records allow a brand to measure repeat behaviour and lifetime value, which matter far more than the vanity of a single festive conversion.

The Organisational Problem

It would be comforting to believe that better technology alone will resolve all this. It will not, because a good part of the attribution problem is political. Channel owners are rewarded for their own numbers, agencies are often compensated on the basis of reported performance, and a measurement framework that reduces one team’s credit rarely finds an enthusiastic sponsor.

Brands that have made real progress usually began by changing incentives rather than tools. They introduced shared business metrics, such as blended customer acquisition cost, incremental revenue, and contribution margin, that sit above individual channel targets. They brought finance into the room, because a finance team asks hard questions about whether reported return actually reached the bottom line. And they gave someone, usually a senior leader, the authority to adjudicate between conflicting numbers.

Agencies have a role to play here as well. Partners that are honest about what their dashboards can and cannot prove are becoming more valuable than those that present immaculate, unchallengeable charts. The relationship improves when both sides accept that measurement is an exercise in informed approximation, not certainty.

Lessons for the Next Festive Cycle

With another festive season approaching, the brands entering it in the strongest position are not the ones with the most elaborate technology stacks. They are the ones that have made peace with a few uncomfortable truths.

The first is that no single model will ever capture the full journey, and the search for the one perfect number is a distraction. The second is that lower-funnel metrics flatter performance during peak demand, so a strong festive return does not necessarily mean the underlying marketing is working harder. The third is that brand building and performance marketing are not rivals. The television spot, the influencer film, and the outdoor campaign create the demand that search and retargeting later collect, and a measurement system that cannot see this will eventually starve the thing it depends on.

The fourth lesson is practical. Plan the experiments before the season begins, not after. A geographic holdout or a controlled pause is far easier to design in September than to reconstruct in January. Brands that build testing into the festive plan, even on a modest scale, emerge with evidence rather than anecdotes.

Finally, there is the question of humility. The festive season will always be messy, emotional, and crowded with influences that no dashboard can fully capture. A family decides to buy a refrigerator because a relative recommended it, a festival offer tipped the timing, and an advertisement reminded them months ago that the old one was ageing. Marketing is one contributor among many, and the honest ambition of measurement is to understand its contribution well enough to invest wisely, not to claim every rupee.

The brands that will win the next attribution debate are unlikely to be those with the most confident charts. They will be those who ask better questions, test their assumptions, and treat every festive season as a chance to learn how their customers actually decide. The puzzle may never be fully solved, but the marketers who keep working on it are already making sharper decisions than those who insist it is settled.