How AI tools help advertisers to come up with ads that work
AI tools can help marketers mitigate the risk of ad failure but these can't help in creating a groundbreaking ad that is certain to perform well, say experts
India Inc. has been spending close to Rs 75,000 crores on advertising every year producing hundreds of thousands of ads. Not all of them work though. Some fail to make a mark, land up in controversy or get into a legal tangle that damages the reputation of both agencies and brands or causes financial loss to them.
Not anymore. Softwares like Enteropik, Qualtrics and System1 are helping marketers to check how their future ad will be perceived by the people.
“Few Indian agencies have started using these AI tools. Their predictions are based on data analysis on various parameters based on human reactions. These tools can test both video and static ads,” says Rahul Vengalil, Managing partner of Isobar India.
Traditionally, research agencies have been taking qualitative feedback of ads from a set of volunteers but their sample size is often very small. AI-tools test the ads on a larger scale with real-time consumer feedback. They also offer various experimentation at low cost, says Saurabh Tyagi, Chief Client Officer, HiveMinds.
These tools help marketers to mitigate the risk of ad failure in expensive traditional media like TV and Print. Many programmatic buyers are also making key decisions through such algorithms while the campaign is going on.
How do these tools work?
Agencies have to upload the ad on the software to get the predictive scorecard based on its engagement levels and how much emotions, excitement or disappointment it creates among different cohorts.
“An advertiser can run a few creative options on the software for a week or two and then take the winner creative with the best engagement/positive consumer feedback. Another way can be to use all the information captured on digital to create the final version of the ad which runs on TV,” Tyagi explained.
Google’s Director Mix also offers mass customisation of creatives. Another service available by Google is to create a video experiment using YouTube data with Google Ads Data Hub. Based on consumer response to each of the unique ad combinations, marketers can decide which ads deliver better in which audience cohort. They can decide which ad to back with higher spends and multi-media activation.
“Multiple ads can be tested in a close environment and then the winner ad can be released for a larger audience based on insights gathered. This is almost like consumers are co-creating the winning advertising for the Marketer. With advancement in AI, tools like these will become more evolved & effective in future and marketers should start taking full advantage of them,” says Tyagi.
What are the drawbacks?
The tools are not perfect though. “The ad-testing process is a little long drawn,” quips Vengalil citing the week or fortnight-long testing process.
Ambika Sharma, Founder & MD, Pulp Strategy, opines, “India has diverse audiences and multiple languages. These tools can help but real-time feedback like an old school focus group cannot be discounted.”
"The best way to go about it is to use a combination of approaches from the past and the future," says Tyagi.
“Legendary ads like Apple’s 1984 Superbowl were created by people who understood what the consumer wanted. AI can help you optimize your tactical campaigns for effectiveness and ROI or choose between two copy alternatives but it can’t create groundbreaking advertising for you. At least not as of now,” he said.
