To customize or not to customize  YouTube Videos.

Our team reviews content on a daily basis, and what we often see are video ad campaigns that are not designed appropriately for their target audience.

So what is the right level of customization for a video ad that can drive brand impact? It turns out that the team at Google’s Unskippable Labs have been trying to answer this very question through experiments with varying levels of customized creatives. The experiments revealed some interesting patterns on audience responses. Below is a summary of the experiments and their results as well as my thoughts on how to use such results in your quest for brand impact.  

Step 1 – Optimizing audience targeting

The core approach to target audience for video ads is changing. Video ads are no longer served based on simple “demographic signals” such as age range or sex. Today, the key approach to targeting audience is based on “audience intent” signals which are derived based on past searches by the specific audience in question. The audience signal approach enables a more refined video ad targeting. Results show an impressive 20% boost in ad recall and a 50% boost in brand awareness compared to targeting based on demographic signals. This is great news for brand impact. 

Step 2 – Running Experiments

The experiment setup by the team at Google nearly a year ago involves new ways of using audience intent signals on YouTube in order to better understand the implications of storytelling. The team partnered with 10 brands, across different categories and in various regions around the world, and designed a series of unique experiments to explore two questions. First, what is the right amount of customization for an audience-signal-driven story? And second, for which audience signals is it most important to customize creative?

The design of the experiment was as follows. First the team started with couple of hypotheses. These were: (i) more customization was better and (ii) the impact of customization would be greater for the audience who were going through life milestones such as starting a new job or buying a home. 

Each experiment was set up with three signal-based audience groups. For each group, three variants of creatives were shown for a given brand – a control video, a video with copy-only customization, and a video with customized copy and visuals. Below is an example set up of the experiment: 

The results of the experiments for the various brands were measured using Brand Lift with ad recall as the success metric. The impact of video’s performance for the signal based audience group was compared to a control audience group which was based on demographic signals alone

Step 3 – Interpreting results and formulating guidelines

The results of the experiments for the various brands were measured using Brand Lift with ad recall as the success metric. The impact of video’s performance for the signal based audience group was compared to a control audience group which was based on demographic signals alone

Pattern 1: For short-form video ads, the custom copy creative alone was sufficient to generate higher ad recall. 

This was exemplified with the Coty’s CoverGirl test where a 6 second ad was played in the United States. The ad contained both generic and custom creatives for three different signal-based audiences. In almost all instances, the copy-only customization was sufficient to drive ad recall. In fact, the copy-only customization outperformed the ad with copy and visual customization. The pattern held true for various countries and various brands. The takeaway was that for short-form ads, the required production for creatives need not be extensive. 

Pattern 2: For longer video ads, more customization tends to perform better.

A 15 second ad was run for a Czech auto manufacturer ?koda on signal-based audience groups. The results showed that the copy and visuals customization significantly drove higher ad recall compared to the control ad. This held true for other brands in different countries. The takeaway being that for longer “stories”, greater customization yield better ad recall. 

Pattern 3: Video ad customization has better effectiveness when targeting an audience for timely events.

Timely events such as becoming a parent or celebrating a graduation required highly customized creatives. In one test for Johnson’s Baby product, generic control ads were compared against ads catering to audience with intent-based signals. In all cases, the consistent pattern was that highly customized ads were particularly effective as they resonated with the audience’s timely event and decision. The takeaway for products catering to timely events is that a high level of customization is necessary. 

Below is a summary of the three patterns. 

Insight and takeaway

There is no definitive rules here but the studies prove the importance of experimenting and the level of creative customization will vary by brand and product. That said, I believe  the experiments reveal further insights in addition to the patter which include: 

  • Level of customization should be proportional to ad length
  • Running experiments on your target audience is always useful to help tailor video ads. 
  • Working smarter is always more cost efficient than working harder. So make sure you work with experienced and capable people.
  • Finally, length of ad should be longer for products and services that are more personal

I hope these patterns and insights will help you to maximize your return on investment for your creatives as you tackle the target objectives of building and improving brand awareness and impact. 

Write a response … Close responses

Leave a Reply

Your email address will not be published. Required fields are marked *

You might also like