I demonstrated how to produce vertical video ads that delivered campaign results 80% above industry benchmarks and designed one of the first web video editors specialised for that task.
When mobile devices became the predominant screen for watching content, everything changed for video.
- Ad budgets began to rapidly shift from display to video
- YouTube, Facebook, Instagram, Snapchat, traditional publishers, and ad networks responded by announcing new video ad products
- Brands began realising their TV ads are being skipped or scrolled out of screens way before the product even appeared
As a leading creative adtech vendor, Celtra needed to expand into exploding video market. Our assignment was to develop a mobile-first video ad product that would deliver superior campaign results.
I led the team from research through the development of a minimal viable product, roll-out, launch, and beyond, reporting directly to the CPO.
Data from thousands of campaigns running on Celtra’s platform and reports from other industry leaders pointed in the same direction: videos that perform well on mobile are short, vertical, and able to convey the message without sound. The complete opposite of classic video.
We set to investigate if 30 seconds long, widescreen, and sound-reliant TV ads could be adapted for the mobile context. The first step was to experiment with a wide range of existing video ads to understand the limitations and potential of such a video format.
Turning landscape video to portrait is not as straight forward as cropping a photo. Subject’s position within the frame often changes within the duration of the scene. To correctly position a moving subject within the shot would require each frame to be cropped differently.
This wasn’t a new problem. History of motion pictures saw plenty of changes in aspect-ratio standards. The challenge of adapting widescreen films for television was solved a long time ago with the pan-and-scan technique. Vertical is just the new kid on the block. Not easy, but doable.
Changing duration is an entirely different story. In most cases, image is inseparable from the sound. Re-editing the original would require professional video software and expertise that are out of our target user’s reach. There was no way we could make this bit simpler. The only way campaign managers could get a shorter video is by contacting the advertiser, to get the creative agency to commission the post-production studio for a new cut. We hit a dead end. Even if there would be a budget for it, there’s never enough time.
But what good is sound, if >90% of users have their devices in silent mode and publishers (auto)play videos with the sound off? If we eliminated the sound from the equation, editing would become orders of magnitude easier. Narration and dialogue would be difficult to substitute, but not impossible. (Hint: title cards, motion graphics and subtitles.) We could potentially build a simpler video editing tool — like a GIF maker, but for short-form vertical video ads.
These findings became the foundation of the product strategy:
- Brands are unlikely to pay for two separate shoots (landscape for TV and vertical for digital) — one size fits all approach is more likely to succeed
- Video needs to be shot in landscape but directed so that the subject can always be framed within a 9:16 vertical format
- Audio is luxury; the video should be able to fully communicate its content without sound (silent-first progressive enhancement)
Ideation: vertically integrated web video editor
Our creative research informed the technical one. Suppose we were to strip the video file of audio layer, cut the image to segments, and reconstruct motion graphics using web technologies. In that case, we could significantly improve loading time while retaining good-enough image quality.
Before developing ad-serving technology for standard video inventory (e.g. pre-rolls), we focused on the low hanging fruit: outstream video. A fancy word for video playing within articles and news feed content. We carefully chose appropriate formats and designed the video player UI to optimally serve specifics of outstream video user experience.
The final piece of the puzzle was the authoring tool — the video editor. It had to be simple enough to be mastered by amateurs, yet powerful enough in terms of functionality to cover as many adaptation use cases as possible.
During our research, we became very familiar with Adobe Premiere, and After Effects. As novice users, we felt the pain of the learning curve ourselves. A lot of product’s design was informed based on inadequacies of professional tools and simplicity found in emerging mobile apps for video creation. We even started the process by designing for tablets to constrain ourselves to the mobile-first approach, fully knowing it would most likely always be supported for desktop only.
Validation: pilot campaigns
Before committing to a long and expensive development of the authoring app, we decided to put the ad format to test first. If that fails, everything fails. If it works, there’s a chance the success would intrigue and persuade customers to give the unusual a try.
We developed the front-end and went into Alpha by running pilot campaigns with selected customers. The creative production of the ads was executed as a managed service using Adobe Creative Suite to simulate the end result of the product we were building.
“Audi Debuts New Vertical Video Ad Format by Celtra”
— PR Newswire
“Audi sees 80pc higher completion rate with vertical mobile video ads”
— Mobile Marketer
“Celtra’s new mobile technology is born from consumer insights and is perfectly merging media and creative. The campaign performance exceeded expectations by far – achieving results 80% above industry benchmarks”
— Giovanni Perosino, Head of Marketing Communication, Audi
After the initial validation, we went full steam ahead with the development of the minimum viable video editor product. We prioritised the functionality based on how essential they are to the creative production process and how expensive they were to implement.
In roughly six months, we had the first version ready for internal use. We pushed the product into a private beta stage by challenging ourselves to offer managed services to a handful of selected customers to stress-test the product. We systematically collected and processed feedback which informed our ongoing development efforts.
Another essential part of the product was video specific analytics that reported on standard video metric as well as proprietary ones designed to compare the performance of outstream with in-stream video formats that would follow later.
Roll-out and launch
Meanwhile, I also worked closely with product marketing, and customer success teams on the roll-out strategy by educating, evangelising, and producing content that became the cornerstone of pitch decks and webinars, and events we were preparing ahead of the product launch.
End of 2015 Celtra and Kargo announced a partnership:
“We are really only at the beginning of what can be done in mobile. Mobile is not another screen, it is an entirely different content experience. Reshaping the video experience from a mobile-first perspective will change advertising dramatically — from how agencies and brands produce concepts to how advertising is executed to what the consumer experiences —and that is exciting.”
— Harry Kargman, founder and CEO of Kargo
In May 2016 Celtra officially launched the product with “The Art of Short Form Video Creation” event which toured in San Francisco, Los Angeles, New York, and London. It featured a masterclass workshop by Paddy Bird — a world-renowned video editor and founder of Inside The Edit and attracted over 700 media professionals from Google, Vice, The New York Times, The Guardian, Hearst, MediaCom, NBC Universal, Time Inc, Twitter, Shazam, Foursquare, and others.
Product management: D. Rožac, M. Zegarelli, L. Dong
UI design: A.R. Urankar, D. Kužnik
Data analysis: L. Karelis
Product marketing: V. B. Brzin, Ž. Cotman, T. U. Mosquera, A. Benulič, Š. Buda, B. Jurjavčič, et al
Engineering: A. Kmetec, G. Dodig, R. Burgar, M. Uršič, B. Vizjak, S. Jazbar, A. Semiprimožnik, Ž. Tartara, D, Božidar, et al