June 3 - 6, 2024
opinion
Commentary
Personalized Video Ads Could Be Huge
- by Michael Ballard , Op-Ed Contributor, December 29, 2016
Two of the biggest current trends in advertising are 1) the growth of digital video, and 2) increased personalization. So it stands to reason that combining these two in the form of personalized video could be an obvious direction of potential growth for the industry.
So far, though, personalized video advertising has been more of a novelty than a legitimate category. In part, this is because of the current limits of technology and because the idea hasn’t caught on widely with marketers in general.
Yet, looking ahead a few years, it’s likely that personalized video will become a potent form of advertising that will combine the emotive power of video with the granular targeting that is increasingly popular with marketers.
Where the industry stands right now
While no one has yet launched a breakthrough campaign that will jump-start the industry, giants Coca-Cola and Amazon have already experimented with personalized video. In 2014, Coke launched a program on the UKs FOD platform, that which used each viewer’s sign-on information to create 4 million unique video ads that featured Coke bottles sporting each consumer’s first names. (A Cadbury campaign in Israel employed a similar premise, showing a handwritten note whose message changed based on who was viewing. it.)
advertisement
advertisement
Amazon, meanwhile, has been experimenting with personalized ads that are tailored on the fly to individual users. The company was short on details of how it would work, but it seems to be based on the same idea behind retargeted banner ads — if you searched for a miter saw on Amazon, for instance, then you might see a video about miter saws rather than a standard banner.
Nike, too, has worked with a company in France to make personalized videos for every runner in a recent Paris race. The videos featured the runners’ finish times and course, with the Nike swoosh.
Looking ahead
We are excited about these interesting early attempts at personalization and there’s no doubt that they have surprised and delighted viewers. It is still very much early days, however. Looking ahead a few years, it’s likely that personalization will get much more sophisticated and effective.
For example, right now a brand can use a consumer’s IP address to personalize the content she’ll see on a corporate website. In 2020, a consumer might analyze that same information and generate a personalized video instead. If the visitor has an address that indicates that she works at Cisco Systems, then the video might include information that other people at Cisco have viewed. A visitor to the website might also see a testimonial from a competitor. Imagine visiting a CRM vendor, for example, and seeing your rival talking about how great the system has been for his business.
Consumer-facing businesses will also be able to employ this tactic as well. As Amazon’s experiments show, one possibility is to tailor video messages to a consumer’s recent searches.
All of this requires that marketers start looking at video in a new way. Being able to personalize messages can mean shooting a lot of video to cover many scenarios and/or a lot of different scenarios and repurposing video that’s already on topic.
The real holdup right now is on the technical side. Real-time processing, the cloud-based infrastructure and streaming protocols will all need to evolve over the next few years to make this possible.
Given technology’s march, it’s a fair assumption that this will happen. In the meantime, as consumers get more used to watching video on their mobile devices and begin to expect that more of their marketing messages are to be tailored for them, marketers would be wise to starting looking ahead to how personalized video advertising could play a role in their overall media plans.
Michael: Just curious, when do you think the tech confluence will be totally in-place for
personalized video? 2018? 2020? Will brands first gain experience with audio personalization before moving to video or is it straight to video?
I thought I'd point out that the technology to personalize the video creative... even on a per-impression basis... is now well-proven. At Impossible Software, that's what we do. We've done campaigns for ebay with live auction data, Deichmann Shoes with remarketing, and so on. We've been at it for a few years now, with the fastest, most scalable delivery infrastructure available.
In answer to James Smith: the technology parts are available today. It's less a question of technology confluence than a question of workflow confluence. We expect you'll see what's possible in 2017!
I agree that the technology to make this happen already exists. Storybulbs is an easy to use platform that allows the creation and distribution of Personalized videos at scale. Much like a Mailchimp, we are entirely browser based and put full control in the hands of the customer. We even have the ability to pull data into the video through APIs, perform analysis and decisions as well as render HTML elements.
We agree that 2017 is going to be a defining year for dynamic video ads. Vehicle has developed some very interesting campaigns for our clients in the personalized video advertising space for quite some time now. I believe this is the year you will see rapid adoption by brands/agencies in the ability to deliver dynamic ads at scale with a large degree of variability. Brands are requesting to target a specific cohort based on location, behavior, product preference, etc. and all of this can be done today with Vehicle's personalized video platform. Here's to a very active 2017 with dynamic video advertising.
James, I think we are currently there with video. It's easy enough to change out text and images. The hard part is not making it corny. It's hard to say with audio (combined with video). This makes me think of Siri's use of customized words. I think due to language and pronunciation the risk and variables are way too high.
Nailed it Dan!
Exactly...but I say here's to dynamic advertising in general...not limited to video.
Michael - we've seen different approaches to audio. What does not work is trying to restrict the name to some text-to-speech engine and pre-recording everything else. It's incredibly obvious, and more likely to annoy than attract. What tends to work well is the (more expensive) approach of having a voice actor say the entire phrase that includes the name... for 100 or 500 common names.
Also - in terms of adding/altering content - there is a broad range of capabilities out there from simple, static text overlays to video-in-video, full-motion, in-scene, quasi-3D with masking. We do (up to) the latter, achieving photorealistic content insertion. (We've even dynamically put "tattoos" on people in the video.) Just as the audio can be annoyingly artificial if you go cheap, the customized video can look creepily bad with weak effects.
But - take it from a technologist - the technology is not the primary determinant of success. Campaign execution matters a whole lot more.
While making this technology less corny is possible, I can't see it ever being less creepy. And the more seemlessly "personal" it is, the creepier it becomes to anyone who's paying attention.
A lie told by a smooth con-man is still a lie, and a "personal" sales-pitch is just another con-job. I'm not saying it doesn't work, because it obviously does. I just wonder what percentage of your target demographic is driven away via their gag reflex?
That's a fair point. However, I think our creepiness radar (in general) is become less sensitive. This started the inception of retargeting (especially with Amazon) and that people are finally realizing that their entire life is public (thanks to Facebook). I think the more interesting use of personalization is through the use of product placement instead of your name.
Product placement has been talked about for years... the idea that a character in a video can pick up either a Coke or Pepsi can depending on who pays. However, there are numerous obstacles to dynamic product placement. Specifically, the producers own the creative rights - distributors may not change the content dynamically. Even if they got the right to do dynamic placement, distributors do not want to... because it changes their status under DMCA and therefore greatly increases their costs of compliance. And that's just the start of the legal, business, and practical challenges of dynamic product placement!
So, product placement occurs during production and is fixed thereafter.
Though product placement is supposed to be subtle (I think?), there was a not-so-subtle example a few years ago in a comedy film. I can dig up the name of the film if anyone cares.
Anyway, the two main characters were sitting at a patio table at a SoCal beach joint, with three beers in cans in front of them. The setup was a standard three-camera shoot; ... one wide-angle side shot, and one medium frontal shot on each character, showing the table and beers.
The goofy part was that from all three different angles, the beer cans had been turned to show their front logos. It was so obvious that some of those in the movie-hip L.A. audience began laughing at each angle change. Also so obvious that it had to be intentional, possibly a response to a placement contract dispute?
A little late to the party here, but better late than never. Our Cadbury video was mentioned in this article (by the way, that campaign actually took place in India, not Israel).
I did want to add a slightly different view: While I do agree that the creative and campaign execution are massively important drivers of success, I do believe that the technology is equally important. It seems to me that technology can only be ignored / assumed to already be there if you take a pretty broad definition of "video". Is a moving slideshow creatd using flash considered to be video? How about a GIF? How about a regular video with HTML5 overlays - is that considered to be a "personalized video"? And what about standard videos with placeholders for dyna,ic elements that are then added on using object tracking solutions?
All of these could be considered by some as being some sort of video solution. But in order to truly be adopted by the creative community, personalized video solutions must provide the same production value and creative flexibility to the agency as they are accustomed to with more traditional forms of video.
This is why we have created our Native Personalized Video solution, which handles dynamic elements within the post-production process in the exact same way that non-dynamic elements are handled. This provides the creatives the comfort of knowing that their personaliozed videos will not be tacky or of low production value, and it can only be achieved through the reimagining of the video rendering process (in our case, we have built a GPU-based video rendering engine that renders personalized videos at 10X real-time). Likewise, editing and distribution tools have to be recreated or tweaked for a world of dynamic video, since being able to deliver personalized video is not the same as being able to deliver it withn a very short timeframe and with minimal pain.
One last note about the creep factor: At the end of the day that is always a creative decision, and one which most marketers are acutely aware of. Furthermore, the creep factor applies equally to all forms of dynamic creative, such as dynamic banners or dynamic keyword insertion. Having said that, ee certainly do advise anybody using our technology to avoid being a creep!