Remember that ad you saw the last time you were watching your favorite TV show, sports match, or news report? Yeah, that one with the split screen, and lots of quick movement, and products flying in and out of the screen and people everywhere? Oh, you don’t? You’re not the only one.

Overly complex ads are nothing new, and somehow brands seem to think that the more features they put in their ads, the more the target audience will remember them. But in reality, it’s more like the opposite.

Studies from Pieters, Wedel & Betra (2010) as well as Pileliené and Grigaliunaite (2016) have both shown that high visual complexity in advertisements can be detrimental to brand awareness, furthermore, detrimental to the amount of attention paid to the brand and the overall attitude towards the brand.

So we know that ads that are overly complex aren’t good for your campaign's results. But still, it happens a lot. Weird, right? Well, as you might know, making effective ad creative is no walk in the park. When you’ve got executives, creative agencies, finance departments, and practically every other stakeholder dropping in their opinion about your ad, it’s no surprise that things end up being overly complicated.

Now, we don’t want to seem too forward here. But, if you’re looking for a simple, efficient and scalable way to measure the visual complexity of your ad so you can focus on results and not opinions – you’ve come to the right place.

What do you mean by visual complexity?

Visual complexity in advertising refers to the strength, number and movement of visual attention areas in the video. The best way to explain this is with a visual guide – let’s take a look at the below example of two still frames taken from separate ads in the beauty sector. The frame on the right has one clear focus area, one product and one visual element for the viewer to focus on. Whereas the visual on the right has many different focus areas, which are spread out across the image. In the case of the right image, it will be hard for the viewer to know what to focus on – they know there is obviously a lot of products available, but it will feel a lot like information overload rather than digesting a clear and simple message that will be more memorable.

In addition to the number and strength of visual attention areas, visual complexity also refers to the change of visual focus points over time. If you have too many quick scene changes, split-screen, and people moving all around the ad, it’s going to be difficult for the person to follow your visual narrative and therefore it is more likely you will lose their attention and very likely their preference for you brand.

How can visual complexity be measured?

So, knowing this, we built a score in our AI-powered YouTube ad testing platform (Junbi.ai) called Cognitive ease. This is essentially one score that will give you second-by-second insights into how visually complex your ad is at each moment, and benchmark your ad in terms of visual complexity against thousands of real YouTube ads.

But how does that work? How can you predict visual complexity with nothing but a neural net? Well, thanks to our clever data scientists from our parent company Alpha.One, we’ve been able to train an attention-prediction model that can predict visual attention on any image or video with 95% accuracy (and we called it expoze.io). The end output from this model is an attention heatmap just like you’ve seen in the example above. From this attention data, we are able to derive certain metrics that give us insights about effectiveness of advertising – such as the visual complexity (and other neat metrics like brand attention, too!).

How can you test the visual complexity of your own ad in Junbi.ai?

If you want to simplify your ad, the last thing you’d want is a complicated tool that is going to tell you how to make it simple. So, we’ve made the process of testing your ad in Junbi extremely straightforward – in just a few clicks, you’ll be getting those valuable insights you need to make your ad more memorable and send your ROAS soaring. Let’s take a look!

Step 1: Upload your ad

Drag and drop your YouTube ad into the ‘upload new ad’ field, give your prediction a name and then indicate which brand you are testing the ad for. This is an essential step for calculating the brand attention score, which we cover in extensive depth in this blog article.

Step 2: Select a viewer mode

The next step is to select the viewer mode in which you wish to test your YouTube ad. This is another crucial step, as it lets Junbi know in which context to place your ad when it creates the attention prediction heatmap. This will significantly affect the way attention is directed to your ad as it competes with the visual clutter in the surrounding YouTube environment. However, thanks to being tested in a realistic contextual setting, you will have a much better idea of you your YouTube ad will perform in the real world.

Step 3: Select an ad type

The final step before you can run your prediction – select the ad type of the YouTube ad you uploaded. Junbi will run a quick check on your video file to see how long it is, and will disable any ad types that might be unuitable for this ad. In this case, the uploaded ad is 30 seconds, which would only be appropriate for the ‘skippable’ ad type. Once your ad type is selected, this will let Junbi know which benchmark dataset to compare your ad with, to ensure your ad is only benchmarked against other ads of the same ad type.

Step 4: Create your prediction

Simply click ‘next’ to create your prediction, and Junbi will get to work processing your ad. This should be done in 5-15 minutes (depending on ad length). In the background, Junbi will create the attention prediction heatmap, test your ad in the selected YouTube context, and create your scores for our 3 effectiveness drivers: Brand attention, ad breakthrough, and most importantly – Cognitive ease!

Step 5: Get your cognitive ease prediction

When your prediction is complete, your results will be visualised on the output page. You will see results for 3 key metrics (or what we call ‘effectiveness drivers’). These are Ad Breakthrough, Brand Attention and Cognitive ease.

Firstly, you will see an ‘overall’ cognitive ease score. This is the score that you see within the circle that has a colour based on how well your ad performed. Let’s take a look at the example below from Hellman’s – this is a perfect example of an ad that finds a great balance between visual complexity and creativity with the concept and messaging, with an overall cognitive ease score of 92.

If you take a look at the graph below, you can see all 3 effectiveness drivers visualised over time. This is a frame-by-frame analysis of each effectiveness driver, and as you can see for cognitive ease, this ad does an excellent job of keeping the visual complexity quite low. You will see that the line ‘jumps’ quite regularly – this is normal, every time there is a scene change, there will be a slight jump in the chart. This is the temporal component of cognitive ease which measures the amount and frequency of scene changes. However, the amount and frequency of scene changes in this ad is quite good, so this has not very negatively affected the cognitive ease score.

The second element of cognitive ease is the size and number of visual attention areas. The great benefit of this ad is that it has quite a clear focal point for most of the ad, with only two characters set against a mostly white background within the fridge, it very clearly guides the viewer towards the main characters – and of course, the jar of mayonnaise.

Around second 21 in this example, you can see a lot of quick scene changes when the sandwich is being prepared. As visualised in the chart, this has negatively affected the cognitive ease score, as there is a lot of quick movement in these few seconds, making the ad harder to visually process.

However, with a cognitive ease benchmark score in the 92nd percentile, and an overall score in the 99th percentile – this ad from Hellman’s is an excellent example of how to do advertising right. Well done, Hellman’s!

Tips and best practices

There’s probably quite a bit to digest in this article, particularly if you’re new to the world of consumer neuroscience and measuring attention on your ad. But fear not – we always do our best to keep things simple here at Junbi.ai. So, if you were to take away only a few key learnings from this article, we would recommend you try to implement the following best practices in your next YouTube ad:

• Avoid putting too many visual focus points in each scene (people, products, scenery etc.) – less is more!
• Try to avoid split-screen and visually dividing the attention of the viewer
• Don’t include too many abrupt scene changes, make it easy for the viewer to adjust their eyes and know what they are looking at
• Don’t overdo the movement within the ad. Try not to have products, people or any other visual elements jumping around the screen and moving too quickly, this can make it harder to draw the viewer’s attention.

Sound interesting? We’d love to share more insights with you in a live demo with one of our specialists. Or perhaps you can’t wait to get started? Submit your YouTube ad for a free test and benchmark using our Free Youtube Ad Testing Tool.

Happy testing!