Ooyala put out theirLongtail Video dropped some stats about video consumption through their products. I've been digging through the data trying to glean some interesting insights and I've got a handful for you today, video viewing by country and region so you can optimize your pre/mid/post-roll campaigns.for Q2 2012 recently and
For Ooyala's part, they pulled the numbers from their network of publishers. They see over 200 million unique viewers a month in more than 130 countries so their analytics should prove useful. Longtail compiled theirs from JWPlayer installs and Bits on the Run usage. That means that none of these numbers are industry-wide, but when we start compiling them I think we get a clearer picture of where the industry is going and how video is being consumed digitally.
Optimize Online Video Ad Campaign Tip #1: Mobile and Tablet
I've already mentioned on several occasions that you should be optimizing your online video publishing schedule to match the target device because there are certain times of day that each (PC, mobile, tablet, TV) are used more than others. Here's a quick look at some Ooyala info in regards to mobile and tablet which further reinforces the fact that you need to optimize your content for these platforms. This is pretty much just what percentage of video viewing is done in each country on them. It's amazing to see how much Japan watches on their mobiles and how much the UK watches on tablets.
As you can see, even in the US and Canada, we are using our tablets for video more than we have in the past. Longtail shows that tablet and mobile usage for video is reaching up to 10% on their platform. You can also see that there has been a rapid rise in video viewing in the last year on tablets.
Tablets are clearly growing in video viewing, but if you're going to optimize for them, you might as well do yourself a favor and optimize for mobiles as well.
Optimize Online Video Ad Campaign Tip #2: Time of Day
Both sets of research also showed that video consumption varies not only by platform but also by time of day. We already had an idea about that as I stated, but these findings reinforce the need to daypart your campaigns or run multiple campaigns at different times of day and on different days of the week.
Ooyala found that, by region, video viewing follows a very similar trend. EMEA and LATAM both watch a bit more later at night while APAC has an early morning spike (most likely commuters).
Longtail went a step further and broke down the usage numbers by device finding that a larger portion of video viewing on tablets was done at night versus other platforms and that desktops rule the daytime hours.
Device usage by time of day. Light blue indicates late night / early morning hours, pink indicates morning hours, yellow is afternoon / early evening, and green is late evening / night.
Ooyala broke theirs down by counties in each region as well, which I think is extremely informational, especially if you've got campaigns you're running outside of the U.S. and need better results.
US, CAN, UK
Optimize Online Video Ad Campaign Tip #3: Placement
By placement I am talking about in-stream: pre-, mid- and post-roll. There has been a rise in mid-roll completion rates recently, most likely because more long-form content is being watched. But that isn't true across all devices as there is a major difference in how much long-form is consumed by device. So as part of your online video advertising campaign optimization you should look at where to do a pre-roll and where you can probably do more mid-roll, or a mixture of both.
According to Ooyala, tablets rank second only to Connected TVs and game consoles though desktop looks to be almost even (66% to 62%) in regards to long-form (10+ minutes). In their research, tablets gained a pretty good chunk in this past quarter. Long-form gained almost 25% on tablets alone and showed more modest gains on other platforms.
The continued trend of more long-form consumption means more in-stream video ad inventory that will need to be filled. It's probably safe to assume that you can do mid-rolls or a mix of pre- and mid-rolls on desktops, tablets and connected TVs. From the looks of Q2 2012, you might even see some good success with mid-roll ads on mobile, though it's only 50% of all video consumed on the platform now.
Average video play lengths on tablets are in the 200 second range, 3-4 minutes. Surprisingly, for desktops, it's closer to 150 and mobiles are around 100. Connected TVs average about 500 seconds per play, about 8 minutes. Clearly, mid-roll is going to be a major portion of inventory on that platform for in-stream ads. But don't limit yourself to just that, see myfor more.
Optimize Online Video Ad Campaign Tip #4: Day of the Week
These types of research, by themselves, are interesting. But when you start compiling them together, you can start to get a better view of the online video landscape. Turning that into optimization tips seems the next logical step for me. We know that viewers watch different content for differing lengths of time based on device and we know that they do so at different times of the day. Longtail even did it by day of the week (below) for some added insight.
This gives you a bit more information on what day you might want to push more ads. You can see that weekends generally account for just 25% of video viewing on most devices and the rest of the time is pretty much evenly spread across the weekdays. It does seem like viewing slows later in the week, perhaps that's because people need to get more work done by the weekend or they're spending more time in the bars.
That's a Wrap
While the numbers in the research might not match up entirely, it's to be expected as both Ooyala and Longtail did their research based on their own products. That being said, it's still valuable information, especially when we look at it in a more combined way like I did here.
For the full stats post from Longtail head to their blog. You can download the full Ooyala Q2 2012 Global Video Index. Neither of them was specifically talking about online video advertising, but I thought their insights could be applied and should be.
Interested in more articles of this type, just let me know and I'll dig some things up!