YuMe’s take2video Blog

Our friends from Google, a few miles south of us on the 101 hit the jackpot for several reasons, one of which was their allowing all advertisers, (but mainly small ones, and ones from SMBs) to sign up, login and start running text ads. Great idea. Let’s now all do it for video. Small issue though – It’s a few orders of magnitude harder to upload a video ad than to type in a 3 line text advertisement.

Also, since the availability of video content on ad networks is far, far less in terms of volume or impressions/streams, or whatever metric you want to use, it deosn’t work so well. The reason being even if we figured out how to get the video built, uploaded and targeted,  the next problem is that the target of the 94065 zip code or DMA would end up with a few hundred or a thousand people actually watching the video. In other words the ad network just isn’t big enough to get enough scale for this kind of SMB or small market video advertising.

Spotrunner tried their hand at it for the cable spot market by hiring dozens of people with green screens, G5 Macintoshes and video cameras. There are several companies, big and small, including Comcast (see article here) who are dipping their toe in this water and specifically pitching the fact that it will be easy to advertise online with video because there are great tools to make and upload ads and target the ads online.

Well, we support all these measures. And we think it’s a great opportunity to get the little guy the ability to advertise on video. We just think that doing it on a large network will afford more eyeballs than on a web-site or property-by-property basis. Maybe we should call Comcast.

- Jayant Kadambi

Targeting is a funny thing (funny strange, not funny ha ha). Every Advertiser wants targeting. They want to know if people are left handed so they can pitch them left handed golf club ads. And if they are interested in golf, they figure they can pitch them a really expensive, left-handed set of golf clubs.

The issue with targeting on video (or for that matter anything) is that as you get more granularity on the audience, the smaller the inventory pool.  Audiences (18-24 yr. old males) become niches (18-24 yr. old males who like golf) and niches become micro-niches (18-24 yr. old males who like golf and were visiting golf club sites).

So, the general problem is that once you find the micro-niche for the advertiser, they’re happy and in turn they’ll run the campaign. Then everything is great, until the campaign completes and it under-delivers. Micro-niches have this nasty habit of reducing the available pool of people.  

So what’s the solution? Reach and scale. In behavioural targeting terms, it’s called overlap. If you have a large potential audience (we do - 134MM), the odds of finding people who have been shopping around for a car (proxied by people who have visited car websites or been watching car videos, etc.) in your network for that niche grow higher. And you’ll get more of them as a result, which means intent-based or behavioural targeting opportunities will be more relevant.

We recently announced the capability to do just this across our  network. We waited until we had grown to a point where we could provide a meaningful audience for advertisers. Because few things are more frustrating than being sold the ability to find left-handed males who are golfers and then finding out that the network has 3 people.  It’s hard to go back to the client and tell them that their targeted spend under-delivered badly.

YuMe’s behavioural targeting solution is pretty simple… we have reach and scale on a national level, offering advanced targeting and optimization capabilities that make video ads perform… that’s just it.

- Jayant Kadambi

There was an article in the NY Times on the 21st of April “A Web Shift in the Way Advertisers Seek Clicks” on advertisers increasingly looking to display ad networks due not only to their reach and cost-effectiveness, but also due to their technology and targeting capabilities. Display ad networks seem to be growing at a faster clip than the general graphical ad market as a whole.

Obviously, this is good news for us at YuMe. However, in order for video ad networks to continue this trend, they need to provide a strong value proposition to video advertisers. This includes acknowledging that video is different, and that video advertisers need content-level transparency, strong targeting, cross-campaign reporting and flexibility in ad formats and creative design.   

In the coming weeks you’ll see news from us that will provide advertisers and ad agencies unprecedented visibility and metrics for their video campaigns, delivered run-time that will essentially provide a window into a campaign’s performance.   These metrics include not only the usual items such as impressions, clicks and views, but also offer items such as unduplicated reach, frequency, video completion rates and other measures of user behaviour that will help determine the effectiveness of a campaign.

Stay Tuned.

- Jayant Kadambi

When advertisers launch campaigns, they tend to target particular demographics. To identify that demographic, they usually refer to the reach numbers of a particular Web site or ad network but also turn to research firms that track and measure such data, such as Comscore. In other words, Web sites – or specific pages within a Web site – are targeted for the campaign. And this passes as an acceptable proxy method to reach a demographic audience.

Well, I think for video, TV’s got it right. In addition to site demographics, TV advertisers use the content or content category as a proxy for the demographic audience. Based on years and years of experience, we pretty much know who watches the Super Bowl (everyone), Dora the Explorer (kids), and American Idol (18-34). Advertisers should insist on being able to target channels, or specific categories of content, on the Web much as they do today on TV.

When an advertiser says he is showing up on Discovery Channel or the Oxygen network or HBO, we have an instinctive understanding of both the content and demographics. We need to add this way of thinking to the online video advertising landscape. In addition to the Web sites or domains on which the ad will run, advertisers also should be asking the ad networks or Web sites to name specific content or content channels on which the pre-roll is running and use this as a proxy for the demographics.

It’s a bit easier with TV and it needs to get that way online. It’s easy to show the client that their ad is running on the third slot at 4 p.m. on Sunday on the Lifetime network. In the online world, it’s a bit more difficult and it shouldn’t be. It just doesn’t seem right to tell an advertiser, “Please click on www.domain.com until you see the ad… it’ll show up, just keep clicking.”

Jayant Kadambi