Online advertising is thriving – in 2017, digital spending first took over TV ads, and since then the gap between them has only widened. Social media has a lot of potential for advertisers as their ads are shown in between emotionally engaging content and the targeting is much more precise. Nonetheless, a lot of companies do not reap all the benefits digital advertising has to offer because they fail to optimize their ads and creatives for social media.

The context in which ads are seen on TV vs on social media is completely different. TV ads can be “forced” on the watcher, while on social media the attention needs to be earned as consumers can scroll on or skip the ad… This fact has a lot of implications; unoptimized TV ads cannot be used online as the attention given to social media ads is on average 3 seconds. You need to communicate with and engage your audience in only 3-6 seconds. Additionally, the ads are typically much smaller in size, so your brand elements need to be much larger in the ad in order to be noticed. Social media requires not only a completely different creative, but a completely different way to test their effectiveness as well.

We believe that there are 3 fundamental social media ad tests rules that shouldn’t be violated:

• Context is king for testing – measure your ads in the correct context (not in full screen)
• Attention is scarce, so understand for how long your ad is seen – 1s or 6s is a huge difference
• It’s advertising, so emotions are key

Social media is different from TV in many ways. One of the key elements of social media advertising is that consumers are not “imposed” with the ad, but can choose themselves for how long they will look at the ad. The scrolling experience is extremely important as it will help to understand whether the consumer is hooked by your ad in a split second as it’s moving in front of their eyes. Using fixed exposure or flashing an ad (or showing a video of someone browsing through social media) will not capture this momentary reaction which is a key factor in the research. Testing an ad in a real or simulated timeline with free browsing and preferably, all the interactive features (zoom in, sound…) enabled shouldn’t be violated.

Online, the consumers get to choose whether they watch your ad or competing content – meaning, you need to EARN their attention. Given the content overload, they only spend between 1-3 seconds watching an ad – and if you are one of the lucky few, perhaps a few seconds more. But the difference between your ad being seen for 1 or 3 seconds is huge. It takes watchers less than 1 second to discard the ad as irrelevant content without storing any info about it. On the other hand, if they stay with you for 3 seconds, you can communicate 2-3 x more messages. 3 seconds are long enough to show a key visual + your brand, and perhaps squeeze in another message.

When speaking about such short timespans, understanding precisely where the attention of the viewer is focused on an ad is key. The cleanest and easiest method for doing this is using eye tracking, and we strongly recommend using it. On mobile (where eye tracking is really not possible) we might go for measuring the on-screen time of the ad as a proxy since there is limited other content on a small screen.

You might debate whether the functional or emotional messages are more important when presenting your product but, in the end, most ads go for emotionally engaging messages. Emotions are key because if people don’t feel good about a product, they will probably not buy it. Once you know that you’ve earned their attention, it is important to understand whether the ad resonated with them and evoked emotions in the consumer. When it comes to emotions, since we need to measure them during the scrolling experience, you have only a handful of methods available. Our preferred is facial coding, as it measures emotional reactions in real time and on a second-by-second basis. Most of the other research methods have difficulty measuring this granularity and you need to have people exposed to the ad for a longer stretch of time, which would, in turn, violate (1) the context again and yield unprecise results.

Additionally, there are two things we recommend having in mind for social media testing:

• Include Survey and RTM for more comprehensive results
• Choose wisely between the respondent’s timeline and a simulated timeline

Aside from eye tracking and facial coding which lead the way in this type of research, including a traditional survey combined with an RTM test make your study even more comprehensive and build the full picture. This way, you can understand the potential risks of the ad (e.g. does it communicate the desired message well). We would recommend using reaction time measurement as well, in order to check whether your messages/desired feelings are communicated within the first 1-3 seconds of the ad. According to Robert Health’s “Low Involvement Processing” theory, we do not draw explicit verbal conclusions, we process and store data through implicit associations – often without even realizing it. Awareness (and even more, top of mind awareness) is likely to be an offshoot of strong implicit associations. This would explain why some brilliant advertising campaigns scored low on tests that are using methodologies based on articulation and statements.

Depending on the particular situation and the task at hand, we can test ads both by injecting the stimuli in the respondents’ actual social media timeline or we test them in a simulated timeline. In some instances, the respondents might perceive logging into their own accounts as risky – which is not the case. Believing these statements will make the respondents feel uneasy about the study and will impact the collected data. This is why we strongly prefer testing in a simulated timeline with controlled conditions, which enables us to test even more variables and easily compare sales uplifts afterward.

Social media ads have already completely changed the advertising landscape and offer plentiful opportunities for previously unimaginable experimentation. You should leave no blind spots when it comes to the impact of your social media ads or other online content – in order to thoroughly understand it, we believe testing should be conducted in a real environment, with measuring their attention and emotional connection, including RTM tasks in your survey. Doing this will ensure the most precise results possible and the biggest impact of your digital advertising.