Experiments in video - Changing habits and behaviours

November 2015

There has been significant change in media consumption over the last few years, with audiences enjoying unprecedented control, watching content where and when they want across multiple screens. But while media consumption has evolved, the fundamentals of marketing have remained the same - marketers just have an incredible, new set of tools to plan campaigns, produce creative and measure effectiveness. At YouTube, we decided to make the most of these tools when we planned our latest ad campaign, while also embracing the agility and experimentation that lies at the heart of the platform.

The campaign set out to celebrate and promote the work of YouTube creators SORTEDfood and Copa90, whose innovative programming around food and football has already won them a close-knit community of millions of fans. We wanted to raise further awareness of these creators among 18-34 year olds, and in this article we'll share how we embraced the digital-first challenge, and share some learnings and best practices from our video experiments.

Reaching a Fragmented Audience

As media options have multiplied, audiences have fragmented. Viewers are more selective, making it harder to reach them through traditional channels. To counter this fragmentation, we approached our campaign a little differently, bringing all of our agency partners together into a single working group where ideas could flow, collide and evolve as collaboratively as possible. As the campaign was aimed at light TV viewers, a digital-only strategy delivered the reach and frequency we were looking for and also gave us the perfect platform for experimentation. We know that consumers embrace digital content for the choice it gives them, so we used affinity and contextual targeting to reach our target audience, and create personal, tailored connections.

Optimising for the Right Message at the Right Time

At the start of the process, we developed a series of hypotheses about how to best connect with our audience, make the most of YouTube's storytelling potential, and create a tailored set of TrueView ads that would allow us to test and optimise. Alongside standard :30 and :60 second spots, we produced creative tailored to the device on which they were being viewed, and that was contextually relevant to the type of content the ads were being shown alongside. To test the importance of 'second screening', we also also shot specific ads to accompany major media events like the Great British Bake-Off and Champions League football.

By establishing clear campaign goals prior to the start of production and giving extra time over to editing, we were able to produce a larger set of creatives without increasing the amount of time spent shooting video. Moving away from :30 and :60 second spots to more nuanced 'shorts' also allowed for reuse and repurposing of footage, expanding the campaign's engagement potential without significantly increasing its production budget. A great example of this - and a great argument for widening the traditional range of inputs on a shoot - came from a suggestion by our social agency. They proposed creating an outtake 'blooper reel' ad for SORTEDfood which, through retargeting, ended up being 23% more cost-effective than simply repeating the :30 second creative. Interconnecting the work of all our agency partners was one of the largest behavioural changes we adopted for this campaign, and in this instance it really paid dividends.

Meet Sorted Food: four friends who love to cook and (clearly) have a lot of laughs while doing it.

We were committed to measuring and optimising throughout the campaign, and not restrict ourselves to the traditional model of pre- and post-campaign measurement. To enable this, we implemented more than 30 BrandLift Surveys to test our hypotheses and monitor their impact on brand metrics in almost real-time.

Learnings and Results

Analysing the results of our experiments provided some clear lessons that we can apply to future awareness-raising campaigns, and that we hope will be useful for anyone looking to achieve similar goals.

1. Devices matter. Because more than 50% of YouTube content is now consumed on mobile, we created shorter, mobile-specific ads. Investing in device-tailored creative proved worthwhile, with awareness uplift running 18% higher for mobile-specific ads compared with desktop creative shown on mobile.

2. Context matters. We found that where possible, running genre-aligned creative really paid off. Knowing that there was a substantial cross-over between football fans and gamers, we created a gaming-inflected ad for Copa90 to run against content in that vertical. This performed 25% better in increasing awareness than showing a generic ad, while for SORTEDfood we achieved uplifts by showing time of day-related ads at breakfast and dinner. Similarly, we discovered that including prominent watermarked branding increased the effectiveness of ads by 20% without diminishing likeability, while including a Card on the TrueView ad increased click-through rate by 2x.

3. Storytelling works. Sequential storytelling and retargeting provided our digital-only approach with an enriched set of narrative possibilities. The experiments we ran serving various :10 and :30 creatives to retargeted users showed that in this scenario, following up with a short :10 ad was much more effective than repeating the initial :30 ad or serving a different :30 creative. Not only were the shorter follow-up ads better at raising awareness, their higher view-through rates also made them cheaper, leading to a cost per-uplift that was 37% lower than for their :30 counterparts.

4. Measurement rules. Running BrandLift surveys allowed us to be agile and responsive, making optimisation decisions based on real time data. Monitoring results during the campaign helped us make the most of our creative and media assets, and ultimately contributed to increasing the performance of the campaign.

Start Small, Start Now

By the end of their run, these digital-first campaigns yielded impressive results, achieving double-digit growth in awareness. But just as importantly, the experimentation and optimisation that went into their planning and execution revealed insights applicable to any advertiser using YouTube to raise awareness. While shooting 40 separate video creatives might sound daunting, it's important to remember that we set out to test a lot of hypotheses at once. Our advice is to start small but start now, by choosing one or two things to test and iterate from there. We hope these insights into our campaign will inspire you to join us in building hypotheses, running experiments and and using online video as a better way to build your brand.

The road to the Wembley Cup: EE's winning YouTube formula