The contextual video gold rush

We’ll treat ‘contextual video’ as video that matches its environment in both content and aesthetic. It runs on the principal that users are performing tasks online, providing them with content that helps them with this solution is more likely to gain trust and engagement. And advertising in this situation is more likely to hit KPIs.

Why now?

Barriers to watching video came crumbling down in 2017: network roaming charges were reduced, wi-fi on the Tube was extended and 4G became ubiquitous. Consumers were watching video from Marbella to Mornington Crescent. Meanwhile, social platforms tweaked their user experience and algorithms to give primacy to video.

The problem

Since out-stream video is considered interruptive, and at odds with user experience, the most desirable inventory remains in-stream. Whilst it has the widest reach of this unit type, YouTube has had some serious issues with brand safety, as a result, advertising dollars are being diverted towards new sources of video.

Brands quite rightly want their video advertising to be viewed by the huge, influential and high-spending online market. Television remains in limbo, and video campaigns remain the best way to spread brand messages.

Contextual video to the rescue

By placing video in an environment where it looks and feels natural to the viewer, it becomes much more appealing. There are video players on the market now which are totally customizable, slipping into web pages naturally.

As well as video looking like it belongs in a page, it must also contain relevant content. We should also consider the user behaviour at the time of viewing.

When matching videos to context we should consider:

• Type of page content (news, entertainment, reviews)
• Verticals referenced (beauty, sports, food, travel)
• Specific brand or interest data points (Lewis Hamilton, Gucci, Barcelona tourism)
• Consumer location (work, home, travel)
• Consumer mind frame (work, relaxation)
• Device (mobile, second screening, tablet)

This triangulation requires a trial-and-error approach, stitching together data and learning from many sources. That’s where ad tech comes in. Finding the right partner to deliver should be top of the agenda for media buyers as we move into 2018.

Machine learning

In 2018, access to machine learning technology will have an increasingly low barrier to entry. Discovering context through machines is now possible, and effective.

Sites can be crawled for words, images and sounds. These can be cross-referenced against libraries of content, enabling providers to place videos on pages algorithmically. This content can then be bought for advertising pre- and mid- roll.

This scalability will be everywhere in 2018. Advertisers will see the costs for in-stream fall, as availability rises. Video will no longer be expensive and difficult to execute.

The benefits of contextual video:

Contextual video is content which provides an in-stream opportunity. It matches both the information on the page, as well as the site aesthetic. To re-cap, here are the main benefits:

• Engagement: Increased clicks, views, and completed views
• Sentiment: Better user-experience for site visitors = happier customers
• Cost: A new found scale of this inventory type is creating value for advertisers

How do we know it works?

We can thank social media for providing the proof of concept. Facebook has done more than anyone to prove that creating an environment conducive to video consumption will increase views. Meanwhile Snapchat showed the appetite for advertising against video, and have innovated in deal types.

We’re in for a turbulent time in 2018 with regards to data, but contextual video based on page content offers stability.

Many were quick to criticise the video gold rush in 2017, and we’ll see a more thoughtful approach in 2018. We’ll all learn that video with context is a powerful place to advertise.

Steve Broadhead, SVP Global Sales, video intelligence.

Join Vi at The Drum's Predictions Breakfast on Thursday 25 January.