MACHINE LEARNING DRAMATICALLY INCREASES BRAND FAMILIARITY, BRAND INTEREST, AND PURCHASE CONSIDERATION ACCORDING TO NEW STUDY BY MAGNA, IPG MEDIA LAB, AND TRUE[X]
LOS ANGELES – September 19, 2018 – Machine learning technology makes advertisers dramatically more effective at identifying and targeting the most receptive consumers, significantly increasing purchase intent as well as other KPIs, according to a study published today. Entitled “The Humans vs. The Machines,” the study was conducted and released by MAGNA, the intelligence, investment and innovation unit within IPG Mediabrands; and IPG Media Lab, the media futures and advisory arm of IPG Mediabrands; and true[X], the leader in attention-based advertising products.
The study analyzed three live ad campaigns utilizing UP//LIFT measurement technology powered by true[X], to determine if more impact is delivered over the course of a month when the campaign is managed with or without the assistance of machine learning — in effect, humans versus machines. Machines working with humans bested humans alone in virtually every category measured – including raising brand familiarity (machines: +6.2%, humans: +3.3%), brand interest (machines: +8.6%, humans +1.9%) and, perhaps most important purchase consideration (machines: +5.7%, humans +0.8%) and purchase intent (machines: +1.1%, humans +0.5%). Machine learning was also more economical, with only 3.08 ad exposures per consumers vs. 4.13 generated by humans required for brand recall.
“The goal of right message, right person, right time is a lot more complex than marketers realize, with dimensions like platform, location, duration, and live vs. on demand viewing all factoring in,” said David Cohen, president, NA, MAGNA. “It is a market that necessitates real-time optimization and the use of machine learning to deliver the best results for marketers. We are excited to be able to quantify this impact with true[x] as a result of this study.”
A key learning from the study is that ad effectiveness surveys aimed at measuring brand lift should be shorter and to the point. Long surveys not only have high dropout rates but induce fatigue, generating poor data and creating missed opportunities with hard to reach audiences. Utilizing UP//LIFT’s “Brand Funnel Impact” (BFI) one-question survey overcomes these obstacles by asking consumers to self-report their position in a brand journey or “funnel” increasing survey completion rates from 67% to 93%.
“Streamlining traditional brand lift questionnaires into a one response brand funnel impact survey doesn’t just create a better user experience that can scale across devices, it generates more data, significantly increasing the representative sampling of a campaign’s reach,” said Jamie Auslander, SVP Research & Analytics, true[X]. “Higher volumes of observations then completely transform brand lift data from what have conventionally been performance notes post-campaign, to driving actionable data that machines can use to boost ROI in real time.”
“The Humans vs. The Machines” is the latest in MAGNA and IPG Media Labs’ media trial series. Recently, the companies explored the tangible impact of brand (un)safety in “The Brand Safety Effect”; harnessing the power of content creators with Twitter; the power of content targeting with Zefr; best practices for mobile ads in its “Battle of the Mobile Ad Formats” study; 360 video ads in its “The 360 Effect” report; the use of haptic technology in mobile video ads in its “Ads You Can Feel: The New Mobile Ad Experience” study and the impact of viewability on performance based campaigns in its “Pulling Back the Curtain: Viewability + Direct Response” report.