• by Randy Kilgore , August 13, 2007

Is this what it feels like to be Paris Hilton? The non-stop buzz, the speculation about your every move, who you've been seen associating with, all eyes on you all the time? That may be a stretch, but there certainly is a lot of buzz about online video. We are without a doubt the "it girl" of online marketing.

With the flood of online video research being conducted at the moment, separating fact from fiction can be as challenging as keeping tabs on the Hollywood rumor mill. Do 30-second ads "really" outperform 15-second ads? Is creating original Web advertising "really" the only way to go? The arguments for and against nearly every data point published are healthy for a growing industry finding its way, but after attending a slew of events focused on our space, hearing opinions from agencies, marketers and the creative community, I am reminded of the phrase that Mark Twain made famous, "There are three kinds of numbers: lies, damn lies, and statistics."

Most of the opinions that I hear seem biased toward the business model of the company the person works for. That makes sense, of course, considering the limited amount of research available. And, as Mark Twain implied, there are plenty of ways to use accurate numbers to draw a less-than-conclusive conclusions. The broader question to consider, though, is what will we say in the face of data that contradicts our own opinions?

My high level view of the initial research is that things look pretty good for the online video advertising space. But it's important to understand the data and how it is being interpreted, while being careful not to get into a "paralysis by analysis" situation. Instead of talking about what is wrong, we should be digging into the numbers, considering their context, applying some common sense and experience, and driving our industry forward both for today and tomorrow. Despite the need for data and case studies to support our own initiatives, we can't afford not to heed Mr. Twain's warnings.

As many of you saw, I spoke on a panel at the OMMA Video conference last month. As you may recall, one of my co-panelists, a creative visionary for whom I have a tremendous amount of respect, followed me by saying, "I can't sit next to you any more. I'm in the complete opposite camp and think everything we're doing is absolutely horrible and we need to change everything." In case you need a reminder, or just a good chuckle, you can watch the video for yourself in MediaPost's video section.

As more data continues to pour into the market, we have to ask ourselves to do two things. First, take every study with a grain of salt. There are only so many factors any one particular study can focus on, and there's sure to be more questions than answers following each new report. Second, we need to accept the conclusions even when they don't necessarily fit with our own personal beliefs or corporate agendas.

While our creative cohorts may undervalue the power of pre-roll, or even loathe the format altogether, most of the recent data suggests that it's a powerful and effective ad format. The Online Publishers Association found that pre-roll video ads with static companion ads had the greatest impact on raising brand awareness. At a recent conference, Advertising.com's Aimee Irwin presented data that showed that pre-roll outperformed in-banner video across all metrics. In fact, the pre-roll ad generated six times as many clicks as the campaign's display ads.

Additionally, Irwin's data showed that higher frequency ( 3 or 4+ exposures) had a positive impact on message association and favorability. Counterintuitive? Maybe. But what will we say about frequency and the user experience the next time we get together?

Another common opinion that we hear is that creative produced for the Web will be better than repurposed content from other mediums. However, the OPA study, published in June, found that there was no major difference between the two. Yet at the same conference that Irwin presented her data, a co-panelist presented a slide that stated, "Content developed specifically for the application will always be more effective than repurposing broadcast assets." While I'd be very hesitant, especially in light of the short supply of research into this particular area, to draw any conclusions yet, the data says otherwise.

Lastly, the OPA study found that both :15 and :30 second spots were effective in lifting brand metrics. In fact, the :30 outpaced the :15 for ad likability. The IAB, as well as other individual companies, will be launching research into this area shortly, but a little common sense can go a long way in this particular area.

I'm sure we can all agree that a :30 second ad has no place in front of a short, user-generated clip on your favorite video sharing site. However, consider ABC's launch of its new high-definition video player. Is seeing two :30 second ads over the course of 50 minutes of programming, such as "Lost," an unfair value exchange? Absolutely not. It enables viewers to watch a program when they want, where they want, with fewer ads than watching on TV, while offering advertisers a very reasonable CPM.

I'm not suggesting that we don't explore new forms of video advertising. In fact, I believe some of the new formats available today are fantastic tools to add to our arsenal of ad opportunities. But while the data supports the use of pre-roll and leveraging our current video assets, advertisers should take advantage of the earlier adopter phase before the landscape is overly saturated.

With eMarketer estimating $4.3 billion being spent in online video advertising within the next four years, and more than half of online Americans watching at least some online video, there will be plenty of opportunity and funding for us to explore nearly every aspect of online video advertising. In the meantime, let's focus on what is good about online video advertising, and help our clients begin the transition so they'll be ready for that next big idea when it gets here -- rather than scaring them back into the stone age of static gifs.