Jivox, who is in the business of multi-screen interactive video advertising technology, released a recent report that, no surprise, says their particular brand of online video pre-roll ads are way more effective than those sort of standard video pre-roll ads you're used to. You know, the non-interactive ones. I'm not knocking Jivox, but when a company puts out research touting how great their own product is on their own platform with their own data, well, you know it needs some skepticism and that's my job. This is why you should always get an impartial third party or state that your results are not indicative of the entire industry.
Jivox added a new metric called “Brand Engagement Lift,” which is supposed to measure increase in engagement based on viewer interaction time with pre-roll ads.
Jivox analyzed for Brand Engagement Lift 259 pre-roll campaigns delivered on its platform between 2010 Q4 and 2012 Q2. The campaigns were randomly distributed across all vertical industry segments – auto, entertainment, consumer products, retail, technology, consumer electronics, financial, and telco vendors.
The Brand Engagement Lift was computed using data gathered about the “100 percent viewed event,” a timestamp that records when a viewer has played 100 percent of the ad video. To derive the campaign’s average elapsed time for the “100 percent viewed event,” the cumulative elapsed time to reach the 100 percent event was divided by the total number of 100 percent viewed events.
So basically it seems to be more or less than 100% time compared to the duration of the video ad. So if you've got a 30-second video ad and the viewer interacts with it for 45 seconds that would be 150% I guess. Oddly, in the chart for the results, retail and consumer electronics are not present from their vertical list in the methodology while mobile, which was not listed is. That, to me, means sloppy reporting. Just a minor discrepancy I suppose.
The Jivox Extended Brand Engagement
Across the 259 campaigns the average extended engagement time was 149% and the peak was 259% Weird, no? The largest peak of the data was at around 101-105% with another smaller peak at 142-146%
By Vertical financial saw the lowest gain around 130% while entertainment saw the highest at about 160%.
Still, I'm skeptical. I mean I understand how interactive video ads could grow time spent with the ads, especially if they're targeted to a highly relevant and interested audience. But when I watch video online, the thing I want to do the most, is get past the ads and watch the video content. Maybe I'm weird, I don't know. I did just play a trivia game twice on Hulu in order to watch the video content without ad breaks but I don't know that the 4-question trivia thing took more or less time than the ad would have as I didn't watch the ad or time the game. However, I did do it because it helped relieve the pain of 55 ads per viewer per month, Hulu's average according to comScore. I just find it strange that all verticals and apparently all the campaigns had higher engagement across the board. That's the part that seems odd to me. I'm also really curious who had the 259% engagement and what kind of ad campaign that was.
The Jivox Interaction Results
The whole point of the interactive video advertisement is to have the viewer interact with the brand and hopefully raise things like brand awareness, favorability, engagement, etc. Jivox's numbers say that the engagement with the brand sees massive uplift, but the report mentions nothing about awareness or favorability. So really, they're simply analyzing data in a vacuum without offering any insight into how this added time or interaction level impacted the perception of the brand or whether the viewers are more likely to purchase a product, etc.
The report also includes some information about interaction rate, which was 7% as opposed to 1.2% for non-interactive pre-roll ads (unverified number from the Jivox report).
By far, the most common interaction was a "custom" one totaling 71%. Of those, the "other" category took 53%and the largest of all interactions was to watch a trailer which was 15% of the custom interactions for 10.65% overall. The next closest interact was to like on Facebook at 9%. Next was becoming a follower on Facebook at 7.81% and then sharing on Facebook at 7%.
Clearly, Facebook is the interaction that should be included in all interactive advertising according to the Jivox data. Though I'm curious if there's overlap in liking something on Facebook and becoming a follower because when you now Like a page there you are technically following it, right?
Surprisingly, SMS info did well with 6% while, not surprisingly, Google Buzz, bookmarking and Myspace sharing all totaled zero percent. Embed the URL also got no love.
In terms of purchases, 6% of the custom interactions, or 4.26% of all interactions, was to purchase some merchandise. Again, I'd be interested to know what kind of products, obviously not the specific products. I imagine it was probably movie tickets as the "tickets and show times" custom interaction did almost as well at 2.84% of all interactions.
That's a Wrap
All in all the research results are extremely positive, yet have to be thought of as only pertaining to the Jivox platform, because that's where all of the data was collected and who it was collected by. Perhaps some vague comparisons might be made to the video industry at large, but it doesn't seem like there was any kind of control group for them to establish a baseline on non-interactive advertisements meaning that we can only guess at what the baseline would be. While the brand engagement uplift was stated as being 48% on average we can only assume that the baseline for that is 100% since we are talking about pre-roll ads. That then doesn't take into account any kind of abandonment of the ads, etc.
Also, the Jivox platform allows for changes of interactions on the ads mid-flight because their real-time analytics give you data that can help you optimize the campaign with a 2-hour lag time. So of the 259 campaigns they tracked for this research, how many were optimized mid-flight and what kind of uplift was seen between the original ads and the optimized interaction ads? It's all a mystery as none of that was addressed. One would presume that, since the campaigns could be optimized, they were optimized and so then all of the numbers in this report are optimal results.