Tuesday, October 28th, 2008
Announcing DoubleClick verified advertising in iTunes
This morning we announced at the Beet.TV roundtable event in New York City that we are now capable of using DoubleClick’s DART platform to dynamically serve and track advertisements in downloaded video within iTunes. The advertisements also travel seamlessly to iPods, iPhones and AppleTVs — although third-party tracking doesn’t work on these devices yet.
This is a first for video podcasts. We’ve been running advertisements in iTunes since last year but this is the first time we’re able to serve them dynamically and offer third-party tracking.
The importance of third-party tracking cannot be overstated. Until now there’s been no reliable way for advertisers to measure the success of their advertisements in podcasts. The best metric available has been downloads. The problem is that not everyone who downloads a video podcast watches it, and not everyone who watches it sticks around long enough to see the advertisements. This has meant that advertisers have been leery of spending money on podcasts. Advertisers need a way to measure the effectiveness and efficiency of their buys.
I should mention that Volomedia has a system that can do semi-dynamic insertion (insertion is done at download time) of advertisements in downloaded QuickTime. Volomedia’s system can also track impressions and clicks, but only when viewers have installed their Volocast plug-in. Kiptronic can also do semi-dynamic insertion but only offers tracking of downloads, not impressions. Our implementation does not require the viewer to install any software other than iTunes or QuickTime Player, and even works in non-iTunes podcatchers like the excellent Miro.
Our system is already in production running a Puma sponsorship on Golf Girl TV (link goes to iTunes so you can see the ad in action!) and a Skype campaign on Back on Topps. The system supports prerolls, postrolls, midrolls and overlays. All ads can be clicked — and clicks are tracked using DoubleClick — within iTunes or QuickTime Player. Clickability is not (yet) available on devices like the iPhone because of limitations of those platforms.
We’re not running any third-party ad networks in QuickTime because they don’t support the environment yet, so the ads we’re running in QuickTime are limited to those that we sell ourselves or that content creators sell. If you opt into run of network advertising on blip (click on “Advertising” from your Dashboard) we’ll serve ads into your QuickTime videos as they’re available. If you have your own sales force and would like to traffic your own campaigns in your QuickTime downloads you can e-mail support AT blip DOT teevee and let us know. We’d be more than happy to traffic your campaign for you (a self-service interface is coming soon!).
I can imagine that you may have some questions about this announcement. John Furrier (the former CEO of PodTech) did too. He wrote up a blog post on the subject. He seemed a little confused about what exactly we were announcing and why it was innovative. I figure that you may have many of the same questions as John, so here are some excerpts from my discussion with John in the comments on his post:
John Furrier:
thanks for commenting. are you turning on ads for all your videos or just select groups? what kind of metrics are you reporting? Views, clicks, and plays? do you guys do dynamic insertion?
Mike Hudack:
Absolutely. Any show on blip can opt into advertising from their Dashboard (http://blip.tv/prefs/). Shows that opt in receive a blend of ads from our direct sales force plus various ad networks (Google, ScanScout, YuMe, Adap.TV, VideoEgg, et cetera) for views in Flash. We have some daisy chaining and yield optimization technology that chooses the highest paying ad for any individual play.
In QuickTime we’re limited to ads that we sell and ads that our content creators sell. This is because none of the ad networks that we work with can serve into QT right now. So shows that opt in will receive some ads in their QuickTime views, but we’re not yet filling 100% of the inventory. Any of our 37,000 show creators can sell into their QuickTime inventory, though, and we’ll traffic their ads for them.
In Flash we’re reporting video views, advertisement impressions, advertisement clicks and engagement. The engagement is shown on a timeline — it shows how many people viewed each second of video. This is particularly useful for brand integration and product placement (we can see exactly how many people saw the brand integration or product placement and how many times).
In QuickTime we’re reporting video downloads, advertisement impressions and advertisement clicks.
In both Flash and QuickTime the metrics come from third parties (DoubleClick for ad impressions and clicks, Illumenix for engagement).
I think that the most important thing here is that, with both QuickTime and Flash, we’re measuring impressions according to the IAB standard — the client requests the ad, and the impression is recorded only at that time. We need no software on the client to do this. Just regular iTunes or regular QuickTime. There’s no need to download anything, and the viewer doesn’t have to be incented to allow measurement to take place — it just works.
John Furrier:
Mike thanks for replying this is great content and thanks for basically agreeing to do an asynchronous interview Q&A here on my blog.
A few questions:
1. An you sent me a pointer or particular publisher video playing in iTunes that you can measure
2. Can you measure while iTunes is in a disconnected state? If not, then is this just streaming iTunes, and who watches video this way? Perhaps you means QT player and not iTunes?
3. Can you deal with .m4v and .mp4 files?
4. you mentioned above “we’re measuring impressions according to the IAB standard — the client requests the ad, and the impression is recorded only at that time.” - are you saying that you record an ad impression even it they don’t watch it if it sits in the library of the users itunes. I’m asking to be specific between requested download, partial download, fully download, and actually watched
Mike Hudack:
You can find links to particular campaigns running in iTunes on the NewTeeVee story that you already linked to. Both are verified using DART.
We’ve found that between 50 and 75% (I know it’s a wide delta, it varies from show to show) of iTunes views happen in iTunes while connected. Apparently *lots* of people watch video this way. For what it’s worth, I do too. I subscribe to podcasts in iTunes and then watch them fullscreen on both my laptop at home and on my Mac desktop in the office. I find it to be a generally better experience for watching shows I like, rather than happening upon embeds on the Web.
Our implementation is also compatible with the standalone QuickTime player and with any software that uses the QuickTime player software (Democracy Player for example).
In terms of what file formats we work with, we deliver the videos and advertisements in a QuickTime container that’s fully compatible with the entire range of Apple portable devices and with the AppleTV. We have about 37,000 active shows using blip today (they release about three new episodes a month — each) and so as you can imagine we have to deal with a very wide variety of incoming video formats. Before we deliver videos we’re trafficking against to iTunes we transcode them to the universally compatible QuickTime format and then modify the container to insert the pointers to DART.
In terms of recording impressions, I’m actually saying exactly the opposite. Current iTunes advertising implementations (Kiptronic, Volomedia unless you download their iTunes plugins) record impressions as soon as the video is downloaded. This is a flawed practice because not everyone who downloads the video watches it (at least not while the campaign’s running and the ad is still relevant!), and not everyone who “views” a video actually sees all the ads. What we do is measure an impression *only when the ad is actually viewed*. This is what the IAB standards require. As far as we know no other implementation that doesn’t require a download by the viewer (and I’d be curious to know what the install base is for these measurement plug-ins is) does this — none of them comply with the IAB standards, and as a result they (unfortunately) overcount impressions.
It’s important to note, again, that for the purposes of advertisements we’re *not* counting downloads. And certainly not partial downloads. We do record those metrics, but for content creators, not to give to advertisers to measure the success of their campaigns. We are counting *impressions* - people actually seeing the advertisement. I can’t stress this enough.
on Thursday, October 30th, 2008 at 1:32 pm:
Sweet! Way To Go!!! :D