"Most campaigns that ran on Facebook and Instagram had a positive outcome on brand saliency, associations, and motivations for consumers," said professor Andrew Stephen, associate dean of research and L’Oréal professor of marketing at the Saïd Business School at the University of Oxford.
Some even had a negative impact, but next to none had zero effect on consumer perceptions of the brand and their consideration of its products.
And there was a considerable difference of up to 35 percentage points between the most and least effective campaigns, indicating the content and style of the ad made a difference.
For the Facebook and Instagram arm of the report, Stephen, along with co-researchers, analyzed Kantar data on 235 campaigns for 110 brands that ran on Facebook and Instagram.
The campaigns were global, but mainly ran in the U.S., U.K., and Canada. Of the 235 campaigns, 49% were Facebook-only, 48% Facebook and Instagram, and 3% Instagram only.
In explaining the difference between the most and least effective campaigns on social, Stephen and his team eliminated the format type, number of ad units, number of creative types, industry, geography, and Facebook vs Facebook and Instagram.
"None of these variables significantly explained the differences we were seeing in campaign effectiveness, so we decided to examine a hypothesis that it was the brand's behavior that made a difference," Stephen said.
To test if it was indeed the brand's behaviour that made the difference, the team used machine learning to classify brands as "more or less human."
"We did natural language processing on texts from the brands' owned media on Facebook, and classified each brand as to whether they were more personable, emotional, or functional," Stephen explained.
The study found that brands that were more personable, emotional, and less functional – and so, more human – had a strong positive effect on brand awareness.
However, there was next to no correlation between brands that are more human and brand association or motivation, such as intent to buy, on the consumer.
"But I'd argue that these factors are indirectly impacted," Stephen said.
Digital advertising, in general, has an impact, too
Stephen and his team also analyzed the effectiveness of digital advertising on brand metrics across desktop and mobile.
To do so, they analysed Kantar's database of 8,811 campaigns that have been run around the world over the past seven years in terms of brand-lift effectiveness.
"We found that desktop has a constant 2% to 4% impact on effectiveness. Mobile, on the other hand really spiked in the earlier days, but with the maturity of the medium, it's now basically the same level as desktop," Stephen said (chart below).
In general, new formats outperform traditional formats initially, but then they drop off, said Jane Ostler, MD of media and digital at Kantar Millward Brown. "This decline is either organic, as people get more familiar with the platform and it loses novelty or commercial environment forces change."
The key takeaway from this part of the study, Stephen added, is that mobile and desktop perform equally well for digital advertising.
Is social and digital advertising worth it?
When asked if social and digital advertising was worth the time and money spent on it, Stephen said, "Yes, but a cautious yes."
"I think it’s good for advertisers to experiment with any different kind of media channel. Plus you need to be where your consumers are," he said.
It's up to advertisers and marketers to be more effective at using the medium, Ostler said.
"We've found that consumers do have expectations of campaigns based on the formats used and they are quite sensitive in their reactions to it," she added. "And of course, none of these campaigns run in isolation, they are contributing to a broader cross-media execution."
Key takeaways from the study
Digital is important for long-term brand building as well as sales;
Desktop isn't dead;
Things don't stay shiny and new forever;
Be more human;
Digital works best as part of the media mix.
This story first appeared on campaignlive.co.uk.