First-party data is far more useful than just targeting advertising to the right audience. At the Times Internet in India, their first-party data is even more valuable considering the sheer number of active users they see on a daily and monthly basis across their many platforms.
“You will see across all use cases, overwhelmingly the first-party data is the most important piece of information,” Puneet Gupt, COO at Times Internet, told INMA members during a Webinar on Wednesday. “That’s like going to a cafe, meeting someone, saying hello, taking their name and their phone number. That’s first party.”
Gupt said the media company will see 140 million daily active users and 635 million monthly. They’ve had to make sure each unique user has a single identity across all their platforms to learn as much as they can about them.
“The interaction of identity and content creates data for us,” Gupt said. “Whenever people go to consume any piece of content — whether it’s music, video, games, content, or news — every click track is captured by Times Internet.”
By tracking every click, Times Internet can personalise news content, newsletters, push notifications, and advertising. The information is filtered to their ad server, which is also the home of their data management platform. Here, all the click track activity gets refined and even goes through an additional layer that helps drive higher engagement.
Times Internet uses first-party data in five major ways:
- Ad yield improvement.
- Content personalisation.
- Crosswalk management.
- Subscription intelligence.
- Lifecycle management.
Gupt encourages media companies to spend time focusing on finding out the identity of their users.
“Unique identity is the most difficult problem for news publishers because traditionally news publishers never had a logged in use case. People would come, they’d browse, and they’d go away.”
News organisations didn’t historically ask users for their e-mail or phone number and, if they did, they didn't account for the same user on desktop and mobile, Gupt said: “All these things can create noise in the identity system.”
So, he said, companies need to make sure their identities are clear and they can merge identity profiles.
“We have a big initiative on profile syncing because we collect about 1.2, 1.3 billion identities every month,” Gupt said.
Times Internet is using technology to link profiles and get that number down to 600 million active users. In fact, finding unique identities is so important to TI, they decided to incentivise users.
They made it easier for people who don’t want to create a login to still log in using Google.
“It’s a single click login across all of our sites,” Gupt said.
The same sign-in auto logs users into all the other Times Internet sites, and all the user’s preferences are saved. Another way they incentivise users is to make offers giving logged in users certain benefits. This has resulted in more than 40% growth of logged-in users in the last year and a half. They also run incentive campaigns to get people to share their demographic information, like offering free subscriptions if they share their age and gender.
Times Internet has also made strides in leveraging a user’s interactions to know them better. If the same user goes to a real estate site, books a restaurant, or listens to certain music, they get a better idea of the demographics of the user. The music interactions are key, Gupt said, and they’ve had great success classifying users in certain age brackets based on the decade of music they’re listening to.
“All these things get combined into a single identity that’s created by profile merging,” Gupt said.
Times Internet also has a consumer data platform or data management platform formed by all the data signals from a user like — whether they’re visiting the site from social media, what device they’re using, and their ad activity.
Gupt broke down the five ways Times Internet uses first-party data:
1. Ad yield improvement
TI has an ad tech engine they named Colombia that has a data management platform with a proprietary ad server that runs specific algorithms.
“This data is getting organised in 7,500 plus audience clusters,” Gupt said. “Each cluster lends itself to lookalike modelling.”
They then use technology to find more audiences like the ones getting clustered.
TI is seeing some major benefits and results with click through rates when it comes to advertising, Gupt said: “For every interaction that happens between an ad we have delivered using our own data management platform and the consumer, the CTR is about 30-150% higher compared to not using the data management platform.”
The lift varies by the category. Gupt said they can find a set of audiences across their sites whereas they used to have separate campaigns for different sites. Before using their data management system, an advertiser would ask to target a specific audience and they couldn't do it. Now with the system, they can find specific audiences.
Once they had the data collection down, TI built another product internally called ACE (automatic campaign engine).
“The ACE engine is able to pick the creatives and deploy the campaign directly on the ad manager without human interaction,” Gupt said. This saves them a lot of impressions, increases CTR, and helps find the right audience. They’re seeing a 40% higher CTR with all campaigns that are powered with ACE.
It’s not just advertising where they’re seeing big results.
2. Content personalisation
“We use data intelligence to create content widgets, and those widgets that are personalised drive 40-50% sometimes 100% higher CTR than non-personalised widgets,” Gupt said.
They also create personalised feeds in the apps, which give 30% to 40% higher session durations.
“A lot of our newsletters are now automated,” Gupt said. “We pick up the content, create the set of newsletters that we want, and find out which audience cohorts we should be sending this newsletter.”
Because of this strategy, their open rate doubled. They also saw three times the CTR with notifications after personalising them.
To get even more engagement, TI created a loyalty product called Times Points. For each action TI wants a user to take, they give them points. This could be reading an article or watching a video. They look to see what the user redeems the points for whether it’s food, movies, or groceries. They use all that data to help grow the quality of data for that user.
“We help users redeem those points across a bunch of strategic partner properties,” Gupt said. “The redemption programme of Times Points also helps us know more about our users.”
3. Crosswalk management
TI also uses first-party data for what they call crosswalk management. This is important for them since they have so many different sites. They try to move users from one site to another that has higher output.
“The more we know about the user and their preference, the easier it is for us to move a user from site to site,” Gupt said.
4. Subscription intelligence
As many news publishers choose to go from ad based revenue to reader revenue, Gupt wants to see first-party data help both revenue streams run parallel to one another.
“For me it is very important to say, ‘Can I use my first-party data and derive intelligence built on it to say hey, these are the users who have a propensity to buy a subscription product,’” Gupt said.
This is usually a challenge because it’s a very big change for newsrooms to go from fast content designed to perform well by search and social media standards, to longer form content that users will pay for. If you know which users you should not show a paywall to, Gupt said, you don’t show them a subscription service and continue advertising to them.
First-party data can also help you create a dynamic paywall that will lock certain articles to certain users when you know what they’re willing to pay to consume and you can decide if that user should pay per article or may want to buy a subscription.
5. Lifecycle management
Finally, when it comes to data and lifecycle management, TI created a product called “Growth RX.” Because of first-party data, they can create their own segments for example: people who have done (x), (x) amount of times over the course of (x) amount of days. They can narrow down those users and target them in many ways.
“I can create user segments of someone who installed the Times of India app but did not open the app for four days, so what is the communication I do with that user,” Gupta said.
Growth RX can send those users an e-mail, and if that doesn’t get them back, they can send a push notification or engage them in several other ways.