Machine learning to boost digital marketing in India


Recently, a multinational food and beverage company based in Thailand increased its consumer engagement through a digital advertising campaign strategy, promoting regional beverage flavours of Thailand using location and time of the day as personalisation triggers. All this was made possible with the help of artificial intelligence and machine learning as 125 unique variants of products were personalised for individual customers in different regions with different meal moments, image and copy variations among others. This resulted in higher customer conversion rate for the MNC.

Though AI and ML technologies have been around for some time, they have recently evolved into a very robust tool for marketers to uncover insights about consumer behaviour in real time by processing vast amount of data in milliseconds. Based on this data, brands can predict and recommend products that consumers are most likely to purchase by an understanding of their intent. This technology concept is becoming very popular in India, where marketers are taking cue from international user cases.

“Consumers usually turn to a device, typically a smartphone, with a strong need to research something, watch something or buy something. The decision to buy can be influenced by a lot of things, such as specific offers or what friends post on social media. Personalisation technology needs to very quickly interpret what the trigger was, in terms of consumer intent. Real time data feeds the machine learning algorithms. The analytics are critical to aggregating data that enable machine learning algorithms to assemble and deliver the best advertising creative and messaging,” explained Naren Nachiappan, managing director–India, of Jivox, a US headquartered firm which has a platform that connects brands with their audiences in a personalised way. This platform uses big data, machine learning and dynamic creative optimisation (DCO) technology to send relevant messages across different channels.

Experts such as Nachiappan said that with ML-based recommendations marketers are able to deliver precise, relevant and impactful digital marketing campaigns on a real time basis. “This technology can help from audience targeting to audience engagement and help accelerate conversions and increase the efficiency of campaigns. Using this many companies have been successful in targeting their consumers much better. Brands within the e-commerce, retail, travel and hospitality segments are already seeing the benefits that AI and ML can bring to digital marketing,” Nachiappan told THE WEEK.

Nachiappan said that content based recommendations focus on product attributes, such as categories, pricing and tags. “Machine learning technology can determine behavioural and contextual triggers and correlate patterns instantaneously, making it possible to engage the user quickly while they are in the market and ensure they get engaged or re-engaged with the brand. Automatic predictions about a consumer’s interest is based on similar products available from the brand. For instance, the Toyota Sequoia and Toyota Highlander are similar to the Toyota 4Runner, so users may also want to take a look at these. Collaborative filtering is used to create clusters of audiences based on their past browsing and buying behavior. The recommendation app views products from the standpoint of a consumer’s past interactions with them on websites—browsed, searched, clicked, rated, liked, saved in their virtual shopping cart or purchased,” added Nachiappan.

Jivox is already working with many companies in India and is trying to redesign their digital marketing campaigns. “In today’s fast paced retail environment driven by mobile users, it is critical to deliver the right messaging and creative about a relevant product or service in real time as it will determine the difference between a sale and a missed opportunity. This data driven technology is different from traditional recommendation engines, which operate in batch mode. In traditional methodologies, product recommendations might be processed when the user is no longer in the market for that product or service. ML technology can determine behavioural and contextual triggers and correlate patterns instantaneously, making it possible to engage the consumer quickly while they are in the market and ensure they get engaged or re-engaged with the brand,” said Nachiappan.