A VC« MySpace is the AOL of Social Media/Web 2.0 | Main | M Ward Tonight at Webster Hall »
Could Post Roll Command Higher CPMs Than Pre Roll?
I posted my YouTube’s Potential Revenue piece to stimulate conversation about the best way to monetize the super popular web video phenomenon. I surmised that 100 million videos served per day could someday be worth north of $400 million in annual revenue and net YouTube something like $150 million in annual revenue.
It was a well-read post and I got a ton of great feedback on the post via comments and posts that linked back to mine. Most people got the fact that I was trying to stimulate a conversation more than project revenues for a specific company.
Here are my big takeaways from all that feedback:
1) Many people think that sticking a pre-roll ad in front of web video will severely reduce their popularity. I agree but it also depends on the content. It could work for SNL clips, it probably won’t work for kids falling off skateboards.
2) Many people think a $15cpm will never happen for web video on a service like YouTube. I think they are wrong and that it will happen.
3) Several people (including Umair, who I always pay attention to) think the whole purpose of a service like YouTube is to blur the lines between content and advertising. They should be one and the same. I totally agree with them.
I’ve been letting all of the feedback rattle around in my brain for the past week and I’ve been using YouTube a lot as well. And I’ve come to the conclusion that post-rolls ads may turn out to be more valuable than pre-rolls. Here’s why:
Watching videos on YouTube is like surfing the web. You watch one, find another, watch it, and do that until you get bored. The end of every YouTube video is a frame that features two suggested videos. I rarely click on those suggested links right now, but I think they are extremely valuable because over time they will become the natural place to find the next video to watch. I think the ending frame is the killer post-roll advertisement. I don’t think they should run a video post-roll, they should just monetize those links.
I think YouTube ought to take a page out of Google’s playbook (if they haven’t already) and make one of the links paid and one organic. I’d suggest the top link be organic and the bottom be paid.
The organic link should be served up by an algorithm that surfaces relevancy and quality, the way Google’s pagerank does for a search query. That algorithm should take into account the creator of the video you just watched, user generated tags on that video, popularity, number of comments, etc, I don’t have any brilliant insights into what that algorithm should be, but I know it should take cues from Google’s page rank, Delicious popular, Digg, Flickr’s Interestingness, Amazon’s front page recommendations, and any other web discovery algorithms that have become popular.
Over time this algorithm will get optimized by marketers like the search engine optmization business has optimized Google’s search algorithm. And what they will be optimizing is the organic link to another video. It could be an advertisement or it could be a videoblog, or some other form of video content. The lines between content and advertisement will blur. I agree with Umair completely on that.
The second link in the post-roll frame should be a bidded marketplace like the right column on the search engines, but it should be only for videos on YouTube.
Say you are a marketer, like Nike. You make a video, doesn’t matter how long or short it is, about your new Nike sneaker that works with the your toaster so breakfast is ready when your run is over. You post that video to YouTube in your brand channel (which already exists on YouTube) and you buy a bunch of tags (user generated keywords) that you want that ad to be linked to from the back frame. You pay on a CPC basis, only for the clicks that come from that link and lead to plays of your video.
If your video is great and the audience loves it, passes it around, etc, you can probably stop buying traffic to it and rely on the organic linking and viral nature of the service to keep the video playing.
Maybe there should be a third link in the post-roll frame that is controlled by the creator of the video. It can be given back to YouTube and monetized and the content creator gets the lion share of the revenue coming from that link. Or if the video is an advertisement, that third link is to purchase the product that was advertised in the video.
My gut tells me that when YouTube does this (and I suspect they are already putting the pieces in place to do it), the post roll frame is going to become very valuable real estate, potentially more valuable than the pre-roll.
But YouTube and the other web video services need to do one other thing. They need to embrace services like NBBC that offer a model for “licensed content” to run on their services. If NBC wants to run pre-roll ads in front of Saturday Night Live clips, that should be their perogative. Over time, content owners may find that other means of monetizing web video (like the post roll scheme I just suggested) work better. But until YouTube and others “clean up” their services by working with the licensed content creators, it’s going to be tough for them to monetize any of the video because of the risk of monetizing someone else’s content without their approval.
This post, plus the YouTube’s potential revenue post, and the NBBC post sort of summarize my current thinking on web video. It’s a huge phenomenon already. And it’s only going to get bigger in the coming years. It will be monetized. And I think we are beginning to see the path for how this all plays out.
September 14, 2006 Venture Capital and Technology | Comments (19)
Comments
So what you are saying is that's it's okay to steal someone elses content as long as you don't make money off of it.
I think that's still wrong, and that YouTube is den of immorality, nobody goes to YouTube to watch the kid fall of the skateboard, they go to watch SNL and music videos - stolen content.
Posted by: jackson | Sep 14, 2006 11:22:16 AM
Merging content and advertising - no thank you - if I'm viewing for entertainment reasons I want NO advertising anywhere and I suspect that applies to the huge majority of YouTube viewers.
Posted by: John Dodds | Sep 14, 2006 12:10:15 PM
Agreed on the post-roll links, Fred - not that I have the secret sauce to what # of links is best, etc., but I do agree that getting into the user's YouTube clickstream and ideally, their profile, and serving up relevant content (organic and otherwise) is a good idea. Maybe not links - maybe an Ajaxy recommended content list that shows to the right of the video. Perhaps a list suggested by the content being watched, and one suggested by the user's profile and clickstream. But I'm starting to wonder though whether YouTube has all the data to optimize that content feed - it has the video, but perhaps not the tools to target. Maybe it takes an Amazon or an Apple to do that? (More thoughts on this in my post on NBBC at my link above.)
Posted by: Rob Hyndman | Sep 14, 2006 12:30:44 PM
For John: So, I guess you never watched the Mac/Windows ads or the Coke/Mentors videos and you have no interest in watching the new Bond film because Aston Martin paid for the car placement... The truth is, advertising is unavoidable... so it might as well be entertaining.
And no, actually most of the YouTube demograpic actually defines themselves by the brands they consume... they don't want ads, but they want brand relationships. Ads suck. Entertainment (branded or not) rules.
Fred, totally agree with the postrolls...
Posted by: Charlie | Sep 14, 2006 12:37:46 PM
Like most of the others here, I totally agree with the post-roll. Pre-roll ads are not just annoying, they're ineffective. Does a "successful" pre-roll ad expect me to leave the video I came here to watch and go find out more about their product? By the time the video I watch is over I've forgotten about the pre-roll ad.
Post-roll is a great place to include these because I finished watching and I'm looking for something else to watch. If I just finished watching an apple parody ad, there's a very good chance I'd click through a post-roll to view Apple's new iPod nano ad.
If companies can produce ads or promos or anything else I'm interested in watching, and youtube can help those companies identify who would want to see those (through past viewing habits), youtube might be able to be profitable one day....
But I still don't see how they'll be able to move to that point from where they are now with all the copyrighted material on there. Major studios will be slow to release their content for syndication, and youtube can't really start monetizing until they get the rights to the video. Simply removing all the copyrighted material would make youtube much less popular (I don't go there to see people falling off of skateboards, I go to see the SNL skit about parents taking a class on myspace that made the front page of digg).
Posted by: Bill Erickson | Sep 14, 2006 3:27:21 PM
Post-roll inserts is currently being explored by Revver. The interesting aspect of this is that the ad is a still image. This has the nice benefits of low serving costs as well as low barriers for small-market companies to participate in the ad buying (e.g. mom-and-pop stores).
In any case, all of this discussion is moot until YouTube filters out content that is not "clean". That in itself is a huge operational burden at the scale the YouTube is operating at. They say they get 40k videos a day uploaded. How many people would have to look through that content? Nevermind the probably 1million pieces of content they have already.
Seems like Content Owners are starting to get testy (do they smell $$$?). In particular the RIAA and Universal Music Group are starting to play hardball with YouTube:
http://www.washingtonpost.com/wp-dyn/content/article/2006/09/13/AR2006091301877.html
Do I hear a plummeting valuation....?
Posted by: Chris D | Sep 14, 2006 3:47:51 PM
Why do they have to be ads? A good brand manager could come up with interesting clips that advertise via product placement. Post them on YouTube and you don't have to pay anything for the advertsing.
Posted by: Jon Smirl | Sep 14, 2006 4:22:03 PM
I think that any successful ad placements are going to have to take into account the market's desire for multitasking--for example a banner ad placed in the share it/email form, so that the user is exposed to the ad, while still able to do something other than just sit there and watch an ad.
Posted by: DWittSF | Sep 14, 2006 6:05:07 PM
Fred,
You misunderstand why post-roll is so effective for this medium -- it's unrelated to the content. The reason post-roll will work for online video is because there's no programming, no next show. You can't tune out a post-roll ad - you have to act - you have to either click to close or browse off or whatever - regardless, you're forced to recognize the ad - so. post-roll ads in non-programmed video demand more attention.
Posted by: David G | Sep 14, 2006 8:08:58 PM
the attractiveness of post-roll ads is exactly what makes them a non-starter as a business model -- post-roll ads are non-invasive, but effective advertising is, by definition, invasive.
google's ads would be clicked far far less, if at all, if they all were placed at the bottom of the screen, or below the fold. sponsors/advertisers would flee en masse if magazines stuck all the ads at the back of the book instead of the beautifully invasive avails in the middle of the texts and images. ditto TV advertisers...
i think its axiomatic: the more annoying (invasive) advertising is, the more effective it is, with the one huge caveat that invasiveness turns to affrontery when advertising compels (not requests, compels) you to action (which is why people who happily open literally piles of direct mail solicitations and catalogs get livid over telemarkerters calling during dinner.)
IMHO distribution may be the power behind the throne, but content is indeed king -- and advertising is the tax that keeps the kingdom solvent and voluntary or easily avoidable taxes don't get paid.
Posted by: steve | Sep 14, 2006 8:12:39 PM
I think your post roll suggestions are a great idea for YouTube to monetize what they've got. At the end of the day, people can always stop the video if they dont' want to see it, but by adding in a suggested video by the author people might be curious enough to stay until the end.
I also completely agree with Jon Smirl's comment that they need to appeal to an leverage the multi-tasking habits of viewers.
Heleen
PS As you're spending much time on YouTube you might want to have a look at a 5 minute stand up gig I did at the Edinburgh Festival the other weeek. It was my first time and purely on a whim so don't see this as trying to promote any future career! http://www.youtube.com/watch?v=BAvC0RQ5tI8
Posted by: Heleen | Sep 15, 2006 4:13:05 AM
For Charlie
Firstly I was talking about not wanting to consume advertising in the context of YouTube or similar and I don't think I'm alone (note that Virgin Mobile incentivises users with free minutes to watch its phone advertising - the reason being that without the incentive people choose to avoid the advertising).
Secondly, product placement is not interruptive advertising (though in some movies it does become so overt as to denude the experience) so the Bond movie example is invalid.
Thirdly advertising is not unavoidable because it is increasingly discredited and avoidable via various technical developments. Marketing on the other hand is, I agree, essential although to talk about people "wanting" brand relationships is post facto rationalisation.
Posted by: John Dodds | Sep 15, 2006 12:41:18 PM
this is a great post - one of the best i've read in terms of provoking thought about online video.
it's worth pointing out, though, that you're talking about apples and oranges.
the pre-roll unit is a video clip.
the post-roll unit is a link.
just to be technical about it... i think a view of a link to a video is unlikely to be worth more than a view of the video itself. in other words, the CPMs for links are NOT likely to be bigger than the CPMs for videos, on YouTube or anywhere else.
But that doesn't mean that the _revenues_ won't be bigger (at least for YouTube).
Again, great post!
Posted by: Nicholas | Sep 15, 2006 6:45:09 PM
Fred,
Great article on post-roll. I am in the process of creating a dynamic video based website with our own daily video content. I have been working hard to try to find a good revenue stream for the business (I know how old school:0)
You have a great blog and your insights really help us artists try to find a way to make it work for all.
T
Posted by: Taylor | Sep 16, 2006 12:46:11 PM
Taylor -
Make sure that you try Postroller (www.postroller.com). Postroller was built around many of the principles discussed in this thread and is the world's leading ad network for user generated video. My apologies for the shameless plug...
:-)
Posted by: Saar | Sep 18, 2006 2:33:32 AM
Post-roll advertising is valuable. Not only do you know that the user is interested in something enough to watch the whole video (factor in video length in computing user interest), but you know they're in the mood for that kind of content _now_ (or at least in the very recent past). There's value there, and it's a kind of value that sites like google lose - because they send the interested user away to another site. YouTube keeps the user around, and so gets a second bite at the ad placement cherry. I'd be charging significantly more for post-roll ad placement.
Posted by: Terry Jones | Sep 19, 2006 11:43:14 AM
Here's another move YouTube could make:
Ads that run during TV shows are obviously 1) effective, 2) not disturbing enough to lose too many viewers, and 3) familiar.
So YouTube could put ads in the middle of videos.
Sounds crazy, but..... it's easy. Give content providers the opportunity to indicate the places in the video where ads could be best inserted. There could be many, or none. The content provider may also provide keywords to describe the video overall and/or keywords for specific ad moments. YouTube just selects the ad to show. Ads are not mandatory. YouTube never inserts ads unilaterally. YouTube shares revenue with the content provider, who is incentivized to do a good (i.e., money-making) job. The intractable problems of knowing where, when and if to put ads in content are neatly solved by having the content providers do the work; plus, each is uniquely qualified for the task.
If this were simple (in terms of user interface), I think you'd see masses of people taking advantage of it, especially regular folk - not just corporations doing deals with YouTube.
Posted by: Terry Jones | Sep 20, 2006 11:57:37 PM
terry, you're absolutely right on this. the incentives on every part work together. users will develop add-friendly content to the extent it works /pays. some will fail, some will succeed. youtube has fewer opprotunities to make it work than the mass of its users. i hope they're paying attention.
Posted by: rustle | Sep 25, 2006 8:42:26 PM
Want to Save Google Videos to your hard drive to play later, Try this web tool?
Posted by: download google videos | Feb 13, 2007 2:55:46 AM
Main RSS Feed of All Posts
VC and Technology Posts Only
