Mumbai: The festive calendar for 2026 has been torn up. Independence Day, which was never a big marketing moment, became the pre-season test.
Onam goes national, Independence Day becomes a litmus test This season started in August, not October. Independence Day sales were up 25% in value, and brands used the day to test creatives, offers and audiences before the real festive burst. Onam then saw 40 to 70% of festive budgets move to digital in many categories. The creative shift was hyper-local.
Instead of generic Happy Onam posts, brands told rooted stories - Godrej talking about washing Kasavu sarees, Mathrubhumi and Birla Opus building on Onam traditions. That specificity travelled beyond Kerala because vernacular creators, localised assortments and commerce-first content took it to dispersed audiences across metros. As Saket Dandotia, Co-founder, CEO, Onetab.ai puts it, "The festive season starts earlier now. Independence Day has become a warm-up - brands use it to test what works before Onam and Diwali." For influencer, Onam was the first real spike.
FMCG, home appliances and kitchen appliances were the most active. Brands paired celebrities for scale and recall with mid-tier regional creators for relatability. "By the time brands approach influencers, many are already booked with shoots and commitments. With festivals like Ganesh Chaturthi, Navratri and Diwali coming in quick succession, the window to plan is very limited," says Shivashish Tarkas, Founder, CEO, The InterMentalist.Quick commerce is the new prime time The second big move is media money shifting to quick commerce and retail media, even though Blinkit and Zepto slots cost 30-40% more during Diwali.
Brands are paying the premium because the last-minute, high-intent purchase is closing there. Rajiv Dingra, Founder, CEO, ReBid, frames it as a funnel collapse. "The consumer journey from seeing an ad to purchasing a product is collapsing from days into minutes." Retail media is no longer just a conversion push in Diwali week. It is being switched on from August to build familiarity before consumers start comparing options.
Vishal Shrivastava, Head of Business Strategy, AnyMind Group India, says that discovery and purchase now happen in the same environment, and brands are using retail media much earlier in the cycle. That changes what creative works. Shrivastava notes that a grand festive ad film alone may not cut it now. The brands winning are the ones connecting creator content, retail media and performance in the exact order consumers actually buy - discover through creators, validate on marketplaces, buy where it is fastest.
The classic mistake Everyone flagged the same error - treating festive as a fixed calendar where X goes to Independence Day, Y to Onam and Z to Diwali. Consumer demand does not follow that calendar. It moves by city, product, inventory, price and even hour of the day. "The biggest mistake is treating festive marketing as a calendar rather than a demand curve.
A Rs 10 crore festive budget shouldn't necessarily be divided on September 1; increasingly, the algorithm should help determine where the next Rs 10 lakh goes," says Dingra. Dandotia adds the other version of the same mistake - treating the whole season as one big Diwali week and throwing a discount on top. When everyone does that, ad prices shoot up and all ads start looking the same. Tarkas says that there is also a structural miss in how regional is handled. Many national agencies try to do regional without genuine market expertise, and brands rely on familiar agency relationships rather than partners who actually understand the local creator ecosystem.
Shrivastava says that brands still treat cultural relevance as the end goal, when it is actually the entry point. A beautiful narrative built with AI in minutes will still have little impact if the brand is not equally visible through the decision points that follow. When Diwali budgets get locked Large brands are not deciding in September. Broad allocations are shaped in Q2, and by the third quarter premium media, creator partnerships and commerce activations are already committed.
What is new this year is that smart marketers are deliberately keeping a flexible pot unallocated. That pot moves in real time based on creator momentum, marketplace trends and early demand signals. The ability to course-correct in-season is now more valuable than a perfect media plan on day one. AI: backend workhorse, front-end hype The consensus on AI is clear - as a gimmick inside the ad film, it is mostly headlines.
As an engine behind the media plan, it is already table stakes. Behind the scenes, AI is deciding when to show an ad, to whom and in which language. The next leap is from generative to agentic - a system that can spot demand accelerating in Pune, identify which audiences and creatives are driving incremental conversion, predict where marginal ROAS will be higher tomorrow, and then recommend the budget shift. "We're not handing campaigns over to AI; we're using it to shorten the gap between what's happening in the market and how quickly brands can respond," says Shrivastava. But human faces still matter.
Consumers are already sceptical of AI characters replacing real people. "Real influencers bring credibility, relatability and an existing relationship with their audience - something AI cannot replicate yet," says Tarkas. He adds a caveat on predictive - a superstar may rank lower on digital metrics but still deliver more real-world recall and cultural relevance, so analytics should be a supporting tool, not the final decision-maker. From media plan to trading desk Diwali 2026 will be managed like a trading desk. Instead of spending evenly for three weeks, brands will front-load spends in the 2-3 weeks before Diwali when browsing and add-to-cart peak, and pull back in the last 2-3 days when bidding gets expensive.
Budgets will also be split by region and product type, not just by channel. The signals being watched are not just revenue, which tells you what happened yesterday. Teams are tracking leading indicators - search velocity, product-page views, add-to-cart rates, save and share on creator content, repeat visits, city-wise demand shifts, CRM behaviour, CPC/CPM movement, creative fatigue, inventory and marginal ROAS. Dingra's test is to identify when additional spending has stopped producing incremental returns and move the next rupee where it will generate the highest incremental return.
Dandotia's thumb rule for pulling back is blunt. "Spending money to sell something you don't have is the biggest waste." Push when brand searches are rising and conversions are healthy, pull back when cost per sale keeps climbing, frequency is too high, or delivery slots are running out. Gold is a good example of how creative is adapting to this reality. With prices so high, jewellers have stopped shouting about grams and discounts and are talking about design, lightweight collections and gifting instead.
Does this save money? Mostly it moves money to better places. Total budgets are not shrinking. What gets saved is waste - ads pushing sold-out products or spraying money in September for no reason.
If AI stops Rs 1-2 crore from going into inefficient inventory out of a Rs 10 crore pot, the economics of the whole pot improve. There is a herding risk though. If every brand's predictive tool sees the same signals, they all jump into the same window and push prices up together. The small brand window This is where challenger brands get a real shot.
The tools inside Meta, Google and quick commerce are available to everyone, so a small brand can time its ads as well as a big one. Big brands still win on data depth and the ability to run campaigns for two to three months before Diwali. As Dingra notes, "They may not be able to outspend a market leader, but they can outmanoeuvre them. A challenger brand that spots a demand pocket six hours earlier can temporarily outperform a much larger competitor." The playbook for smaller players is to not fight in Diwali week at all, but pick one festival or one category and own it completely.
How to know if it worked, and why Christmas is different Don't trust platform ROAS in October - everything looks good in festive. The real test is to run predictive in some cities and the old way in others, then compare total revenue divided by total marketing spend versus last year, adjusted for higher ad prices. Check if the tool was actually right about peak days, and more importantly, look after the season to see if those customers came back. Dingra's filter is sharper.
You cannot claim AI improved performance just because ROAS went up during Diwali - demand would have gone up anyway. The right question is, "What would have happened without the prediction?" Measure predicted versus actual demand, marginal ROAS, holdouts and incrementality, and how much inefficient spend was avoided. And one nuance that brands miss - Diwali and Christmas are not the same wallet. During Diwali, discounts are expected and compared across marketplaces. By Christmas, urgency, availability and presentation start competing with price. "A gift that arrives on time and feels considered can outperform a deeper discount, which is why packaging, bundles, and fulfillment become much bigger levers," said Shrivastava.