top of page
  • Writer's pictureOmri Yaniv


By ICX Media - January 31, 2019

How marketers can anticipate audience reactions with AI

Gillette’s newest ad takes a strong stance against toxic masculinity and it’s been receiving tons of attention. The ad has elicited a deluge of responses from the public as well as from creatives in the advertising  industry.

As the dust settles on this campaign, marketers, and advertisers are taking a closer look at the ad and the responses to better understand the impact it had on Gillette and its parent company Procter & Gamble.

Gillette certainly isn’t the first company to use advertising to take a public stance on a cultural issue, and P&G CEO, David Taylor, has made it clear that this won’t be the last time you see their brands making a statement.

For marketers contemplating campaigns that could drive controversy, it is especially important to understand who is actually driving the massive numbers of engagements in order to better anticipate public reaction and ultimately campaign success.

Artificial intelligence, like ICX Media’s video platform, has the capacity to ingest and decipher complex engagement data -- allowing brands and marketers to understand who is engaging with their campaigns and how.

ICX Media compared audience engagement on the Gillette “Believe” ad with Gillette’s typical audience responses, and found striking results. Engaged viewers of the “Believe” ad were older and more white than their usual audience, with a 16% increase in engagement among viewers 34-55 years old and an 11% increase in engagement from Caucasian viewers.

Even more strikingly, ICX Media found that a projected 34% of the engagement with the Gillette “Believe” ad could be attributed to agent accounts. ICX Media's data science team used AI-driven language processing to identify such accounts, which include bots as well as human accounts that demonstrate either no engagement, perfunctory engagement, or disingenuous engagement. Agent accounts post manufactured comments designed to drive and inflame certain reactions and exhibit digital and linguistic signatures very different from those of everyday human commenters.

While many polarizing reactions to Gillette’s ad were certainly legitimate, ICX Media’s analysis suggests that agent accounts were effective in inflaming that polarization and over-inflating the perception of the public division.

ICX Media detected and identified similar agent activity in other controversial video campaigns such as Nike’s Colin Kaepernick ad last year; ICX Media’s data science team determined that as many as 30% of the comments on the Nike Colin Kaepernick ad likely came from agent accounts, a percentage of potential agent  accounts strikingly similar to the proportion seen in the Gillette “Believe” ad.

Gillette’s report that the “Believe” ad has had no negative impact on their sales serves to defend the theory that the swell of divisive reactions is largely inflated ‘noise.’

As marketers weigh the pros and cons of taking a public stance on controversial topics, they will need to anticipate and plan for the existence of these agent forces especially as they could impact measurement and analysis of public reaction to their marketing efforts. Because of the sheer volume of data generated in response to a campaign, today’s marketer needs to be leveraging AI and audience intelligence in order to effectively filter through reactions, decipher which are legitimate and which to ignore, and accurately measure the impact of their campaigns.

20 views0 comments

Recent Posts

See All
bottom of page