Xaxis’ Bharat Khatri on the promise of programmatic & the ground realities

To many, programmatic advertising may seem a far-off thing in India’s digital future, but last year GroupM’s programmatic arm, Xaxis delivered 2,500+ programmatic campaigns with an average of 200 campaigns per month. Their clients include Manyavar, Valvoline and Frankfinn Institute of Air Hostess Training, for whom they delivered some of the most successful programmatic campaigns.

Bharat Khatri, National Director, Xaxis India, leads client development, overseas operations and publisher partnerships for brands in India. He has spent the last four and a half years growing Xaxis in India by leverage technology to drive efficiency and creativity in digital marketing at scale.

He loves working with brands to make marketing messages a welcome occurrence rather than an unwanted intrusion with data-driven approach.

According to Khatri, programmatic has undergone several stages of evolution in its journey in India. In conversation with Adgully, he narrates the journey of programmatic advertising in India and unique programmatic solutions that Xaxis has delivered for their clients.

Evolution of Programmatic

Programmatic came back a decade ago in India. When it came, it was about how people can automate their buying and selling of media. Rather than calling 600 million website owners to place their ads, they can directly use programmatic and buy inventory to reach the audience they want to target.

That was the flavour of the initial days of programmatic.

Then we moved to real time bidding, where you can optimise your bidding in real time. So, you don’t need to wait for weekly reports, analyse what happened and then do course correction. It is all in real time.

We started moving away from placement buying and began audience buying because most of the people on digital had the prior experience of buying on TV, Print and Out-Of-Home, so when they came on digital, they started buying placements.

Omnichannel Marketing

We don’t offer anything on traditional medium like TV, Print, Radio so for us an omnichannel strategy is where we are able to target the same audience across digital channels whether it be audio, video or any digital formats.

For any marketing campaign our first touchpoint is to identify the right target audience for the brand and what is their behaviour. What all do they access online and offline, we try to capture all the behavioural aspects and then we create this audience profile for our brand.

There is a lot of profiling that happens before we go live with the campaign. We have a proprietary data management platform called m/Insights. On this platform, we create all the audience segments.

Once we set up this audience segment on our data management platform (m/Insights), we then push these audiences on to a demand site platform (DSP) like Google Cloud Platform (GCP) or a random company. We use these DSP companies to target these audiences with inventory on OTT, mobile, desktop, audio, and late event display.

We tag these campaigns before going live. For example, in an offline store location, we need to tag that store location on our platform. We tag the latitude and the longitude of that store and we are able to map people who were in that vicinity in the last 3 days or 60 days, whatever window we want to input. Similarly, before going live on the digital effects, we tag the client website and landing pages so that we can attribute the performance of our campaign.

Then we initiate a deep level analytics to find out what RoI is brought for clients, how many engagements did we drive, how many people actually downloaded the app, how many people visited the store and what else do they do.

Executing a digital-only campaign is easy, says Khatri, adding there are three funnels that guide consumers in the campaign journey.

In case of the top funnel, what we do is engage the consumer with a lot of video ads and focus on whitelist OTT platforms such as Hotstar, Voot,  ZEE5 and YouTube. We run a top funnel campaign to make people aware of the product, communication, and the brand messaging all through video.

We re-engage with the same audience who liked our ads, who skipped it, who didn’t skip it, on different display and rich media formats by showing them product benefits, pricing and all that.

Then we do hyperlocal targeting when they are near the vicinity of a store or drive traffic to a client website or application to achieve the end output which is the store footfall or app download.

According to Khatri, there are all the more opportunities in DOOH because of the vast number of signals available to programmatic advertisers.

We are taking a lot of log level data from the platforms. Every impression that we serve, generates 2,000 signals. The moment we serve an impression, we get to know the impression was served by which publisher, how the audiences engaged with that impression (click, like, skip), at what time it was served, on what content it was served, time of day, day of week, etc.

A human mind cannot read 2,000 signals to build a variable model. We built proprietary tools using AI and ML to build models using this data and put it into context from a client perspective.

If we use technology stats like Data Management Platform, Demand Site Platform, Supply Site Platform in the OOH industry then a lot more signals will be available. For example, a weather trigger, a cricket live score trigger, and other signals that the stats offer.

In the DOOH industry, any screen which is connected to the Internet and is able to play digital ads is out of home. If you bring all these targeting elements on that screen and measuring the RoI then it will transform the way DOOH is looked at from an omnichannel point of view.

Role of Programmatic in Changing Media Landscape


Let’s say a brand wants to run a campaign on a particular channel. Earlier, they were getting 10 million reach but in the new TRAI regime they are getting 6 million reach. The consumer has power now. Let’s say the ICC World Cup is over and the next Cricket tournament is on Sony then viewers may choose to unsubscribe to Star Sports for at least another 2 months. With this change in regulation, it is becoming difficult for a TV planner to plan media and give a guaranteed reach and frequency to the brand.

Using programmatic you can deliver a TV+ campaign. In such campaigns on TV, we leverage OTT platforms from a content and targeting lens and reach out to maximum audiences in our target zone in terms of demographic targeting, geo targeting, hyperlocal targeting, interest based targeting, intent based targeting and then we reach maximum GRP and whatever is the balance we try to deploy it back on TV. We leverage the environments of TV and OTT because the audience is the same, thus the demographics are the same.

To balance or bridge the gap of 4 million, they can target the same audience on OTT platforms and achieve the same reach and frequency.

Audience Buying

Audience buying has increased efficiency and less spillover in ad monies. For example, MensXP is a website where the content appears to be catered towards the male audience. However, comScore clearly shows that around 38 per cent traffic on MensXP is female.

For example, if you run a campaign of Gillette Razor for Men on the website, there will be a spillover of 35 per cent because that campaign or creative will not engage a female audience. So, we have to move away from content and placement buying to audience buying.

Most brands have transitioned in this journey and are now engaging with the right audience. There are still a lot of brands who want to do a mass reach approach, for example, Pepsi. They will definitely invest in properties like the YouTube Masthead, which will give them massive reach in a short duration. It will work like placement buying. If you buy 100 per cent share of voice (SOV) on a YouTube Masthead and it delivers 100 million unique audience reach which is more than any TV channel in India and generate approx. 1.2 billion impressions in a day!

Digital Out-Of-Home

DOOH creates a larger brand impact compared to mobile platforms because of the larger screen that the audience sees on-the-go.

There are challenges, like the media partner is not keen on coming on to programmatic as they are used to traditional buying and selling. He refuses to let technology track the performance of campaigns. He is still happy with 5 per cent share in the overall AdEx for OOH.

There are challenges on the client side as well, as they want to understand how this medium will evolve, what is the efficiency and what is the pricing.

Companies like us work on the International Advertising Bureau (IAB) framework, which is a media institution which monitors all ad formats. IAB had recently defined guidelines that every DOOH company needs to follow. Companies like us need to be sensitive to IAB guidelines and ensure we are following them.

The challenge is to connect all these dots, brings all the partners together to drive innovation in this industry. The way digitalisation is happening in India we will soon see DOOH driving omnichannel strategy with other digital channels like audio, display, OTT will change the way we do media buying on digital.


Right now, in India programmatic has not touched the livestream scenario. There are a lot of challenges in the current ecosystem to make livestream programmatic a possibility. One of the challenges is the user experience. Since we deploy a lot of technology and parameters to determine the performance and RoI of a programmatic campaign, it will cause a delay in the stream. Because of these challenges we need to do manual campaigns on livestream where we need to manually place on the type of audience we want to target.

However, programmatic has covered some ground in the livestream space. Between 2015 and 2019, Hotstar has definitely seen an immense growth thanks to the Jio revolution and cheap mobile handsets that are easily available today. We have achieved a lot of digital penetration for livestream matches in Tier 2 and 3 cities. The concurrency numbers of Hotstar are 25 million+, which is a world record in terms of viewership on an OTT platform for a live event.

In the last 4 years, we have moved towards targeting. Earlier, livestream advertising used to work like TV where you placed an ad that was shown to the entire universe.

Ad Fraud

Ad fraud is there in the digital industry and there are ways to curb it. We use a third-party ad server. Thus, we can control our creatives and ensure we are serving ads only on OTT. First, we will not serve an ad on fraudulent inventory or non-human inventory. There are also tools to check whether the ad was served to a human or non-human. There is a company called Integral Ad Science with whom we work very closely. We deploy their technology in all our campaigns. With the help of this technology it helps us understand whether we are bidding for a human or a non-human impression. With the help of this technology we are able to detect whether our ads are watched by human or bots, time spent, and other factors.

Native Advertising & Branded Content

Definitely branded content is on the rise. Brands have taken a leap and have understood how the consumer is engaging with short form video, original content on OTT, there is definitely lot of brand integration happening on content.

From a programmatic lens, I don’t think this is something that will challenge growth of programmatic. e-Marketer recently released a report, saying that India is the fastest growing market worldwide with 64 per cent growth expected in 2019.

On the other side, branded content and native advertising is also on the rise. I believe both of these will work in synergy and drive amplification.

Branded content which a brand might have created in partnership with a content house can be used with programmatic to amplify that content to ensure that the content reaches the right audience. They can use a clipping of that content to promote on other digital landscapes.

Case Studies


If I talk about our latest campaign for Manyavar, we activated this campaign on digital OOH and the campaign was already running on a traditional channel which is a static billboard. We targeted these consumers on the mobile platform which is another channel. This was an omnichannel strategy where we were able to reengage with the audience who were already exposed to the first format. With the use of location data, we had already mapped the Manyavar stores. After running the campaigns for 30 days, we were able to analyse how many people were exposed to our digital OOH campaign plus the mobile campaign on one-to-one targeting who to find out how many actually visited the store.

The result: Manyavar tagged 400 stores PAN India for the campaign, however it tracked footfalls in Kolkata (21) Delhi (19) and Gurgaon (3) stores. The Unique User Reach of the campaign was2.13 million with an average delivered frequency of 2.5 Impressions per user. Average Click Through Rate (CTR) of the campaign was 1.18 per cent, which is above the benchmark of 0.60 per cent. There was a phenomenal jump in [store visit] uplift when audience exposed to OOH ad along with mobile medium of advertisement (1.98x).


We executed a TV+ campaign for Valvoline. Earlier, when they spent Rs 40-odd lakh on TV, they got 80 GRPs, but when they spent Rs 40 lakh in a TV+ campaign, they got 166 GRPs.

The result: The Valvoline campaign, undertaken in partnership with m/Six to drive brand recall among digital TV viewers, reached 180,000 unique users over a four-month campaign period. For this campaign, the client sought to branch out from traditional TV to reach its target audience of auto enthusiasts, adventure seekers, biking groups and social aspirers through digital TV channels. Without prior campaign learnings as reference for audience buckets or behaviours, Xaxis and m/Six leveraged data sets to identify selective audiences, targeting them programmatically with Xaxis Premium and Xaxis View. Using a robust digital media and creative strategy, the campaign successfully drove a 25 per cent lift in brand awareness, a 13 per cent increase in recorded brand recommendations, and an increase of 10 per cent in intent to purchase Valvoline products.

Franfinn Institute of Air Hostess Training

In the case of Frankfinn, we partnered with a telecom data provider zeotap, a company that monetises telecom data. Using that data, we were able to find out people in the proximity of the University area, they demographics, age group and the type of applications they are using on their phone (Swiggy, Uber, Ola). We targeted these audiences on Hotstar. When people run campaigns on Hotstar, generally it only provides targeting on the basis of demographic, geo, and content. Xaxis has an exclusive partnership where we overlay our own 3rd party data segments on Hotstar. In this use case, we were able to target relevant audiences on Hotstar with minimal spillover.

The result: The brand saw 50 per cent brand upliftment and was able to achieve 6 per cent upliftment in lead conversion through this campaign.


News in the domain of Advertising, Marketing, Media and Business of Entertainment

More in Marketing