By Łukasz Abgarowicz, VP of Agencies at RTB House
In 2022, consumers’ appetite for video content is stronger than ever and demand is still growing. According to 2021 WyzOwl figures, people now watch an average of 19 hours of online video per week, an increase of 8.5 hours per week across the past 4 years. When asked how they would prefer to learn about a product or service, 73% said they would choose to watch a short video.
As a result, 86% of businesses now use video as a marketing tool, a 41% increase since 2016. Video content has become an indispensable, multi-purpose ad format that can boost brand visibility, spark customer interest and increase conversions for any product or service. However, with the upcoming deprecation of third-party cookies on Google’s Chrome browser, marketers are continuing to face challenges in terms of segmenting, targeting, remarketing and accurately measuring the performance of their video advertising campaigns.
For future video campaigns to be successful, advertisers will need to develop first-party data strategies and/or depend on next-generation contextual marketing. This does not mean a return to the contextual marketing of old, an imprecise technique based on targeting keywords and domain names. Modern context marketing relies on advanced AI, known as Deep Learning, which allows advertisers to pinpoint messaging to match the content of both ad and website. The result is more effective and profitable ads, which are more relevant to the viewer’s interests.
Deep Learning explained
In basic terms, AI refers to any technology (usually computer software) which attempts to simulate some aspect of human intelligence. Machine learning is a subset and application of AI that uses data to inform decisions within predefined parameters. Lastly, and most importantly, Deep Learning is a subfield of machine learning that structures algorithms in layers to create an ‘artificial neural network’ that can learn and make intelligent decisions autonomously.
Classic machine learning is not able to make decisions without human input to direct its actions and struggles to handle unstructured, granular or dynamically changing data sets – it just can’t ‘improvise’. By contrast, Deep Learning technology is capable of conducting its own decisions, needing only the desired objective as input.
So how can Deep Learning be used to improve the effectiveness of video campaigns? Here are just three ways in which Deep Learning can help:
- Better contextual targeting
Deep Learning improves the efficiency of contextual targeting strategies by making the most out of every single ad impression. The technology can analyse masses of user information in fractions of a second and decide not only which ad creative to present, but also whether its display is likely to trigger the desired action. Deep Learning tech can make literally millions of these decisions in a short space of time, something that is far beyond the capabilities of even the world’s best media planners. Our own research found that campaigns utilising Deep Learning in this way are up to 50% more efficient compared to those using standard machine learning approaches.
- Better timing
With Deep Learning, multiple targeting strategies can also be supported, including contextual, behavioural, and demographics, that adjust content to the preferences of the target group. For example, different creatives can be used based on target group interest; sporty and exciting SUVs for active persons and safe and reliable family cars for parents.
By identifying a customer’s browsing habits, Deep Learning helps advertisers become better customer journey orchestrators when it comes to sequential targeting. Using the data from Deep Learning, advertisers can identify specific touchpoints in the user’s interactions with the brand, to create an advertising strategy that, when executed, will make video ads more effective.
- Better performance
By applying the power of predictions in real-time, Deep Learning helps advertisers to pick the right time and the right context to deliver each ad to the right user, increasing brand visibility and optimising the cost per completed view (CPCV).
CPCV is a metric for an advertising campaign’s efficiency assessment, but it is also an advertising pricing model – in that sense, CPCV advertising means that companies pay each time a video has been viewed through to completion. In many cases, it may be a better way to invest the marketing budget.
This metric allows you to calculate how much you have to pay for presenting the entire video to a user. Information of this type can be used later to optimise your video campaigns towards the completed views rather than clicks and impressions (as those two don’t really tell you if you have reached a potential customer with your message).
In the end, Deep Learning delivers more efficient media buying and real-time, continuous optimization of your buying strategy, ensuring both value for money and maximised brand impact.