Like other publishers around the world, HT Media is facing the challenge of declining referral traffic thanks to such things as zero‑click searches, AI summaries, and changes to Google Discover.

During the recent INMA South Asia Webinar, Tushar Garg, vice president of marketing and user growth at HT Digital Streams, shared how the company has reinvented its advertising and data infrastructure for a post‑cookie world.

The three-year transformation, Garg said, began with the realisation that publishers can no longer rely on platforms for traffic or on third‑party cookies for monetisation. Today, they must know their users and find a way to activate that relationship across the full funnel.

Programmatic revenue volatility added to the challenge. Traffic from the United States, United Kingdom, and other international markets had dropped, affecting programmatic revenue, and at the same time, advertisers wanted more: “Brands now demand more from each and every ad spend … they want measurement, they want a focus on ROI rather than just a banner inventory.”

To meet these demands, HT needed a data foundation that could support personalisation, segmentation, and full‑funnel advertising.

This became the starting point for a transformation built on three pillars: collecting first‑party data, converting signals into salable cohorts, and full-funnel activation.

Building the first‑party data foundation

Garg joined HT in 2023 and quickly discovered the company had an anonymity problem: “A lot of the traffic on the Web site was unknown and we had no way to retarget the user. We had no way to create a profile of the user.”

First, they implemented user registrations and then moved on to identity stitching to track users across multiple platforms and Web sites. By connecting those activities to a user, HT could create a solid, targetable profile.

HT began by exposing login prompts across its properties, even before it had a fully developed value exchange. Garg suggested publishers often wait too long to introduce registration: “Just putting a registration wall or a login wall in front of the user… still helps you drive a lot of registrations because in general, people these days are used to signing up.”

HT implemented Google One Tap, built a native SDK, and integrated Truecaller to enable one‑click authentication.

The second part of that strategy was ensuring that users stayed logged in and didn’t have to sign in for different sites: “HT Group has many news Web sites,” he explained. “We ensured that we have a single sign-on, so if you log into one Web site, we will be able to track you across other Web sites as well without you having to log in and personalise your experience everywhere.”

These efforts paid off: Daily logged‑in users grew from less than 1% in 2022 to 10% within two years, a tenfold increase that Garg described as “much better than the industry benchmarks.”

HT also needed to unify identity across devices, platforms, and sessions. Even Google Analytics, Garg pointed out, “does not identify the user uniquely across AMP and non‑AMP.”

To solve this, HT built its own fingerprinting system, HTFPID, which allowed the company to track users across AMP and non‑AMP pages, across months, and across multiple HT properties. This also introduced client IDs tied to e-mail and mobile for logged‑in users, and captured device IDs across apps. This identity stitching created a persistent, cross‑platform view of each user. 

With identity in place, HT began enriching user profiles. During onboarding, users were asked about their content preferences. Engagement products such as quizzes, polls, Webinars, and surveys provided additional declared data.

HT also captured demographic attributes, professional and affluence indicators, and behavioural signals such as content consumption patterns, time of day usage, and frequency of visits. Vertical‑specific intent — such as auto, BFSI, real estate, education, and entertainment — was collected through actions like ad clicks, form fills, and product comparisons.

The type of data being collected was crucial, Garg said, put it, “whatever you track is what you can probably convert into a monetisation signal or a personalisation signal later.”

Turning signals into commerce

Once HT had a growing pool of first‑party data, the next challenge was making it usable.

The quality of the data layer determines the quality of the advertising engine built on top of it, Garg said, so “completeness of data is the first step.” With more than 1,000 pieces of content published daily, HT needed consistent metadata — content type, category, entities, and more — to ensure user behaviour could be interpreted accurately.

Standardisation was also critical. City names, vertical labels, and event structures varied across HT’s many properties. Without harmonisation, the customer data platform (CDP) could not reliably identify patterns or build cohorts.

Garg described the painstaking work of normalising values, unifying event schemas, and ensuring consistent data types across sources. Once this foundation was in place, HT began converting raw signals into commercial intent.

This conversion process was central to HT’s monetisation strategy. The team analysed behavioural patterns to infer intent:

• Heavy consumption of alternative investment content suggested investment interest.
• Repeated clicks on real estate ads in a specific city indicated a likely homebuyer.
• Comparing SUVs signalled auto purchase intent.
• Travel, education, BFSI, and entertainment intent were similarly derived from content consumption and engagement.

These signals were then grouped into high‑performing cohorts. HT’s auto cohort, Accelerate, became a strong performer for conversion‑driven campaigns. Its BFSI cohort, Nevaeh, and its real estate cohort, Home, delivered similar results. And Garg said HT has leveraged contextual cohorts as well — something he said every publisher can do:

“What we can do is contextually target users who land on [specific] content,” he explained, adding it could be someone who's reading about an AI code, or an online pharmacy, or a TV series. “All of those can be ways in which you can contextually target for related products.”

Activation across the full funnel

With identity, data quality, and intent signals in place, HT focused on activation and built a taxonomic cohorting tool.

This allowed sales and operations teams to filter users by attributes such as income, age, device, and interests, and instantly see cohort sizes, reach, and impression potential. The tool turbocharged sales conversations, allowing the sales team to “answer questions in real time and … close the client discussions and sales very fast.”

HT then activated these cohorts across multiple channels. Google Ad Manager remained the primary ad server, but HT also built its own in‑house server, Echo, to support non‑standard native formats.

Echo enabled personalised messaging, contextual targeting, fatigue management, and AI‑driven creative optimisation. It also powered interactive formats such as lead forms, vouchers, HTML widgets, and sponsored quizzes — formats designed specifically for conversion.

The company extended its activation capabilities beyond its own properties, using e-mail, WhatsApp, SMS, and RCS to reach users off-site. Thanks to its strong identity and consent, HT could personalise these messages with precision.

Measurement and attribution were the final pieces of the puzzle. HT integrated Appsflyer, Trackier, and in‑house pixels to track installs, transactions, and leads.

“All of that finally feeds back into the data layer for better optimisation,” Garg said. “This is how we are able to actually deliver all of the power of CDP at scale to our clients.”

The transformation was not quick, and Garg said building a CDP required significant investment in time, talent, and technology.

But the results have been impressive. Whilst programmatic revenue has stagnated or declined, HT’s performance and CDP‑based revenue have grown steadily. “In the past year, we have actually surpassed the programmatic revenue comprehensively with the performance revenue that we have set up,” Garg said, noting it prepares HT for a future where first‑party data and performance advertising drive growth.

But perhaps the biggest lesson is that a CDP is not merely a monetisation tool: “It is actually powering the whole ecosystem within your company,” Garg said. “It’s probably one of the most important driving forces for your success.”

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .