October 11, 2006
How Google Should Monetize YouTube With AdWords & AdSense
Unless you have been under a rock the past few days, chances are good you have heard about Google purchasing YouTube for $1.2 billion dollars. Yes, billion. And the natural assumption by many is that Google will somehow monetize this by bringing together their AdWords advertisers with the seemingly zillions of eyeballs watching videos on YouTube.
Unfortunately, it isn't quite as simple as slapping an AdSense ad unit on the page... after all, YouTube has sporadically tried that approach in the past. However, there are some different - and some unique - ways to integrate their advertisers with YouTube in ways that won't necessarily alienate all of those who hate watching commercials before watching what they came to see, an approach used commonly by television networks who show clips or even entire shows online, usually cleverly done so that viewers can't easily skip then without skipping the program they wanted to see in the first place.
Google began experimenting with video ads this year, and has done a great deal to promote them, both to advertisers to begin using them in their ad campaigns, as well as to publishers, encouraging them to use video-ad-friendly ad unit sizes and enabling image ads. And while Google hasn't said just how many videos are running on the AdSense network, there doesn't seem to be a short supply of them.
First, the obvious approach for video ads is what they have experimented with in their MTV video ad beta test - inserting paid commercials within the streaming video. These video ads were 30 seconds in length or less, done in interstitial style, meaning when you'd expect a commercial break if you were watching the programming on the television, you would see one of these video ads instead. Many of these ads were for branding purposes - just as the advertisers would have if you were watching television - rather than something that needed to be clicked on (which many wouldn't do since they would worry about leaving whatever they happened to be watching). This could easily be done on YouTube, particularly at the beginning of short pieces, or during "commercial breaks" of programming, such as in the recently announced deal with CBS that would see primetime shows being shown on YouTube.
But the problem lies that because YouTube's videos themselves are currently ad-free, you could alienate the viewership by suddenly tacking 30 second ad spots at the beginning of every YouTube video. However, there is definitely room for advertisers to create shorter video ad spots, say 10 seconds in length, that users would be much more tolerant of.
There could also be the advertising scenario where advertisers could pick and choose on-the-fly which videos they would like their ads appearing on, meaning the vast majority of videos would remain ad-free, while only those ones "chosen" by advertisers would have ads on them. This could be effective for when videos go viral and get a lot of views within a short period of time. Likewise, perhaps advertisers could select to only have their ads run on the front page "Featured Videos."
Google could also offer the ability for advertisers to pay for better placement of certain types of videos, such as infomercials or longer length commercials about their product or service. This could even be utilized by companies or individuals who want their videos to get better exposure. This could be useful for a network wanting to promote a new television series or a studio wanting to get better video exposure for a movie or video game preview. This could be done very similar in style to the sponsored listings versus organic listings in the regular Google search results, but done with videos instead.
On the non-video advertising front, Google could easily supply contextually targeted ads based on keywords added by members when they upload the video. However, as you have probably noticed, YouTube is also filled with keyword spam, where members stuff a huge amount of keywords into their video descriptions so they show up for terms the video has nothing about (most have something along the lines of "This movie has nothing to do about..." in the descriptions; example - click "more" on the description on the right to see it in action). This could definitely hinder how well contextual ads are targeted on the page, meaning more seemingly off-target ads in a network that is known for its great targeting... especially when everyone will be watching to see how Google approaches the issue of bringing advertising onboard YouTube.
But there would also be value if advertisers could site target their specific text-ads to appear on specific video pages, or for specific video search search results that would target YouTube only, instead of the entire Google or content network. So a snowboarding company could specifically target just YouTube visitors who search for various snowboarding videos by keyword, having them appear only on YouTube and not default onto Google search results or the whole content network
And speaking of the content network, let's look at how the content network could benefit from this YouTube deal. The most obvious would be to allow publishers to monetize from YouTube videos they post to their blog or website. So let's say Matt Cutts does one of his video posts and instead of using just Google video, he put it up on YouTube as well. And say he tied it into something with AdSense ("No, using AdSense won't get your site indexed in Google faster!") so I decide he's worthy enough to embed his AdSense video right into JenSense. As a publisher, it would be nice to have a nice spot to put my publisher ID into that embed code so that if there happened to be a video ad available for the video I put on JenSense (and I suspect there would be a lot of advertisers wanting their ads to appear before Matt Cutts talks shop!) I could make money every time a JenSense visitor viewed Matt's video about AdSense & indexing and saw a video ad first.
An alternative is to allow those who upload original videos to YouTube, to use their publisher ID to make money when someone views their video right on YouTube. Of course, you might have to threaten users with penalty of death (which I figure is definitely evil!) so they don't go on a video upload binge and upload every video they happen to have on their hard drive, regardless of whether they own the copyright or not. But if this had been available, the Numa Numa guy would be rich now.
There are a couple other issues to consider. First there is the whole copyright issue, with people uploading content that they don't have permission to upload, whether it be clips from a favorite show or an illegal recording from a concert or movie in the theatre.
What if someone uploads a naughty video and puts it on YouTube and Mr. Cranky Advertiser discovers his video ad playing at the beginning of the latest celebrity sex tape that has managed to find its way in pixelated goodness onto YouYube? That could definitely happen, even with the flagging system currently in place. That said, it sure isn't practical - or interesting - to have someone review every single video on YouTube, whether it be the latest Numa Numa or the video from someone's cousin's daughter's first birthday that captures all 30 minutes of a one-year-old not opening birthday presents.
USA Today also explores the issue of making money with YouTube in today's newspaper. I also wanted to clarify the quote I made in the newspaper that didn't quite come across as I intended. When referring to the number of websites running AdSense today, I was referring to consumer oriented websites, meaning sites offering things such as product reviews and other types of information-based sites such as content sites, community sites and blogs. I was not referring strictly commercial business sites (ie. sites selling their products or services directly to the consumer), although it seems as though more and more commercial sites are running AdSense these days too! But my thoughts on how AdSense can impact sites directly selling products and services can wait for another day :)
Overall, I think this YouTube deal could really push Google's AdWords video ads to the next level, and be done in a way to allow publishers to earn money (both for placement and for creating the video) as well as allow advertisers the ability to monetize all those YouTube eyeballs. And it gives AdWords advertisers a huge platform of diverse content where Google can control the implementation and overall ease of adding advertising into the YouTube mix to allow advertisers to track just how effective their campaigns are. There will be some interesting times ahead for Google advertisers and publishers.
Posted by Jenstar at October 11, 2006 12:36 AM
Btw,shouldn't that be "$1.65 billion"? Is it "$1.2 billion"?
Posted by: Ken Wong at October 11, 2006 02:46 AM
What you suggest sounds alot like Revver. They give share to the original uploader AND to the person who posts it on their blog. Of course, Revver allows download (with continued revenue earnings, even if reuploaded) and YouTube doesn't.
In February I had to set up a deal w/a social site and a video site and I had to put alot of time in dissecting both sites. I wonder if they patented this ad delivery mechanism? Could you even patent that?
Posted by: Robyn Tippins at October 11, 2006 05:58 AM
Wow, this is one of those posts that have to be read through a number of times to get out all the goodness.
I think you really hit on one great idea Google should implement. Content with ads attached that publishers could pick from ... providing a service to their readers and earning a commission for providing the exposure to Google advertisers. The "Matt Lessons" are a great example. Consider also something like a cooking blog where the chef could do video highlights of a recipe and then instead of just linking to his/her site, other "foodies" could link the vid, with ads, to their blog page. It almost sounds win-win to me ... I must be missing something ;-)
Posted by: Dave Starr at October 11, 2006 10:13 AM
Youtube will no longer for a long time give in its current form. Google becomes it also like so often kapitalistic. unfortunately :(
Posted by: Webdesign Germany at October 11, 2006 12:56 PM
Let's sum up Jen's entire post into one sentence:
"I am so excited, because now YouTube's videos are going to be filled with advertising!"
Um, yeah. Except there will be dozens of other sites that don't have advertising. And uh, well, people don't usually like advertising. So, um, well, Youtube's viewership is going to crater.
Why not do some ACTUAL reporting and show people that the only reason Google bought youtube was because they were back-scratching for their Sequoia VC masters, who just happened to make half a billion dollars with this buyout because they invested 11 million dollars into youtube 6 months ago to become the majority shareholder.
But I know reporting that kind of news isnt likely to help you when you want to speak at a Google ad conference again.
Posted by: Agreed at October 11, 2006 02:52 PM
Looking at Eric Schmidt's comment from the press release:
"Our companies share similar values; we both always put our users first and are committed to innovating to improve their experience. Together, we are natural partners to offer a compelling media entertainment service to users, content owners and advertisers."
Hmm, if users come first then that means YouTube stays as is with little or not advertising. Short term, Google can help YouTube save money by optimizing the back end systems to be more scalable. Long term, Google could work contextual advertising into the product but would need to do so in an inobtrusive manner if they really value the users first. I don't think they should insert ads into uploaded videos. I think they'd be better off matching ad videos to uploaded videos. They could do this by expanding the "related" section to include advertising videos. If these ads really did match contextually with the video the user's already watching this could work. The user would choose to watch the ad video. If the YouTube user could then rate the ad and post a text or video comment related to the ad, this would keep the ad creators on their toes and would lead to some actually compelling content in the ad videos. This would actually help the advertisers as their ads would then go viral. That sort of a solution might just fulfill the promise of this deal.
Posted by: Richard Ball at October 11, 2006 09:08 PM
Well "Agreed", since my blog is about contextual advertising, obviously if I bring up YouTube it going to be about how it relates to advertising. I have plenty of other non-ad thoughts about YouTube, but it doesn't fit the focus of the blog, so I stick to strictly the advertising aspect of it. But there are only about a zillion others commenting on all other aspects of YouTube, and I am sure at least a couple share your view, but I prefer to stick to advertising. And that is what people come here to read, not what I think the VCs might be thinking.
Posted by: Jenstar at October 11, 2006 10:19 PM
Google is getting set to unleash their flow-tracking capabilities. This could double adsense revenues.
YouTube may be the first place they test this. Bloggers will find their CTR going up dramatically once flow tracking is in place. It's a way for Google to serve ads based on the readers recent search activity rather than just on what keywords are on the page.
Posted by: Mitrax at October 12, 2006 06:33 AM
I like the idea of the users selecting what ad could go before their video. This would cause good ads to be shown.
Posted by: Bill at October 12, 2006 08:30 AM
Hand picking from a video inventory this big does not really scale. Neither for the advertisers nor for the publishers. Google is successful because it has found ways to accurately exploit a monstrous inventory of page-views.
Google will have to find ways to allow advertisers to explore and select the entire inventory of videos, not just the three most "popular". Ninety percent of the video viewings are down the long-tail, not on the top 10 list.
Posted by: Paul Salber at October 12, 2006 11:16 AM
I agree with your contention that "YouTube is also filled with keyword spam, where members stuff a huge amount of keywords into their video descriptions." Google could ignore the video uploader's information and instead really embrace this whole web 2.0 social media thingy. For instance, Google/YouTube could use del.icio.us as a collaborative filter. I suspect that the collection of tags created by users for a given YouTube video could be more relevant than the title/description/tags provided by the uploader of the video. Google could crawl del.icio.us, identify the universe of tags for a given video and develop algorithms to cull the most relevant tags. Much like Google uses "semantic technology to analyze and understand the meaning of the domain names" for their AdSense for Domains program, they could certainly identify the most relevant keyword tags.
This technique would be particularly effective for popular videos which would likely be bookmarked and tagged most. These are the videos that GooTube would certainly want relevant ads alongside. However, this could also work for the "long tail" of videos. How many users providing tags would Google need to find relevant keywords? Whether a video was bookmarked 1000 times or simply 10 times, I believe this technique might work.
Plus, somehow I think it would be funny for Google to monetize YouTube effectively by utilizing Yahoo! data. Even if you think this idea is crazy, why not play a game of del.icio.us TagMan? ;-)
Posted by: TagMan at October 12, 2006 11:53 AM