Flipkart, Myntra clean rooms show Indian brands' CRM data is the weak link

A CMO in Mumbai or Bengaluru can usually define a data clean room without being able to point to a live campaign running through one, and that gap frames where India's clean room market actually stands, according to agencyreporter.com. The most developed examples sit inside retail media: Flipkart and Myntra have spent two years building advertiser-facing measurement layers, often branded as 'audience insights' or 'closed-loop reporting,' that let brands bring in anonymised CRM segments and match them against exposure and purchase data without either side handing over raw records. Quick-commerce platforms Blinkit, Zepto and Swiggy Instamart have taken a similar approach into pilot territory, running measurement studies for a handful of FMCG advertisers who want to know whether a 10-minute delivery impression works alongside a television or connected TV ad seen the night before. JioAds is pursuing the same goal at greater scale, pairing telecom-grade identity resolution across JioHotstar, JioCinema and JioSaavn with Jio's subscriber base to give advertisers audience-overlap data once available mainly through Google and Meta. Amazon Marketing Cloud and Google's Ads Data Hub are also seeing uptake among larger Indian advertisers, typically accessed through agency data science teams.
Campaigns built on these collaborations run into two costs before they produce insight. The first is internal: agencies and brands feeding a clean room have found their own CRM records, riddled with duplicates and inconsistent identifiers across ecommerce, loyalty and call-centre systems, are the weakest part of the setup. One senior analytics lead at a large Indian ecommerce platform put it directly: 'the clean room did not disappoint them, their CRM did.' The second cost is time: a genuine clean room query needs legal sign-off from both parties and a bespoke build, which can take weeks against a quick-commerce or festive-season sales cycle measured in days, pushing some agencies to ask platforms for pre-built, templated query libraries. Brands should also expect the platform hosting the collaboration to learn more from the first cycle than they do, since the platform controls the exposure data and the analytical talent running it, with value improving mainly on the second or third engagement.
- Magnite and ITN bring AI agents to local linear TV buying, full rollout by Q4 2026
- Amazon Ads adds ChatGPT inventory to Amazon DSP through new OpenAI partnership
- FAST viewing surges 43% to 1.8 billion hours as platforms withhold audience data
- BidMatrix VP Mark Nedzelskii Says AI Agents Optimise Bids While Humans Set Strategy