Social media insights: Replicating behavioral data during crisis

As most countries have experienced or are currently in some form of a lockdown – to no-one’s surprise – an increase in social media usage has been reported. That being said, we set out to explore whether we would get the same results in an ad study we did 12 months earlier. If not, what are the main differences in the outcome when it comes to advertising?

EyeSee ran a validation test of a cross-industry advertising study by comparing the respondent’s answers from last year to their responses during the crisis. The cross-industry study specifications were as follows: we tested 18 ads in total – 3 ads per industry on 900 respondents. We used eye-tracking, facial coding, and survey to test the dynamic ads in a social media environment and take a more in-depth look into KPIs such as visibility, attention, emotional reach, ad recall, likeability, and others.

The goal of the replication study was to understand how stable behavioral indicators are, given the significant change of the circumstances and expected change of the respondents’ reactions to market research, and if the data provided nowadays is valid.

Winning ads still top their contenders

The best performing brands remain the same in each industry – but not only that, the order of the ranking of top 3 individual ads remains the same, with some winners even improving their previous results on some KPIs. Why do some of the results appear to improve? As we work on additional validation, we presume this is due to the fact that respondents are in a more alert state than usual or that they are spending more time on social media during social distancing. Either way, the stability of the results even during turbulent times, is what proves we can rely on the behavioral to judge whether our marketing materials are engaging enough and able to break through the social media clutter.

No significant differences for most other KPIs

Except for ad recall, when it comes to the different variables and their behavior in the repeated study, most of them (such as attention, dominant emotion, and emotional reach) stay comparable with the first study. Some KPIs show a tendency to increase (emotional reach, likeability, brand fit), but the effect is not strong enough to be statistically significant. This means we can use the same set of guidelines and best practices for dynamic social media ads:

• Showing people or showing your product (if you have one) in the thumbnail is likely to elevate the emotional reach
• Light colors in their thumbnail (first scene) will draw attention to your ad, and emotional reach was higher for ads that used more intense colors (light or dark) than those that used colors with medium intensity
• Lowest-performing ads are too long for social media – always adjust the length for the channel you are using.

You can check out the full blog and our podcast on the findings from the first study.

Ad recall is systematically higher across all ads

Ads are recalled with greater success in the second wave of the study, both individually and on the industry level. By and large, ads from all industries recorded a better recall, with an average recall uplift per ad at around 8.5 percentage points. For the two sectors that haven’t significantly improved in brand recall (yet also saw a slight increase), we saw a significant increase in visibility.

Overall, emotional reach and primary emotions remain stable on both the industry and brand levels.

To conclude…

Although the prospect of deciding whether to keep on advertising your products during the coming months may look daunting, our study showed that people are still open to communication from brands – this might be a good time to build brand equity and stay responsive to the needs of your customers.

• Whether you pre-test ads now or later – you will get the same winning ads. All study findings remained stable after 10 months- all trends and ranks stay the same so we can rely on the results as we continue to test ads during a time of changed consumer sentiment.
• Can we use norms or not? The first conclusion is that for most KPIs, the results are the same. For ads, we can except higher ad recall, so we need to modify this. For other types of testing, we expect the same conclusions.
• Facebook users scored higher on ad recall during the crisis – ads become equally (and highly) visible, regardless of the industry. Additionally, the retention rates of the advertised brands are significantly higher.
• Regardless of the situation and the slight differences in perception, social media ads are a powerful tool for brands only if executed right. Instead of focusing on the contextual differences in a state of crisis, companies should work on adjusting the content and format of their ads.

‍