Tag Archives: TubeMogul

What social media site refers the best video viewer engagement?

by

Social networking and bookmarking sites are a critical part of any online marketing effort utilizing video because you need to get your video seen where your key demographic is spending time online.

TubeMogul recently completed a research case study to find exactly what the title of this post asks: what social media site refers the least fickle viewers? They sampled 6,763,690 video streams over three months referred by links from Digg, Facebook and Twitter to come up with the findings. I’m going to highlight a few of the real key points to talk about but here’s the link to read the full results from their research report.

Results from TubeMogul

The results (below) are surprising: on average, viewers referred by Twitter tend to watch a video the longest (one minute, 58 seconds), compared to Facebook (one minute, 14 seconds) and Digg (58 seconds).

On average, audiences clicking on video links from Twitter watch a video 36.91% longer than viewers referred by Facebook and 49.98% longer than viewers referred by Digg.

My Analysis

This is an interesting study and the numbers are intriguing but there are a few things that the study doesn’t take into account.

Separation of social media sites & social bookmarking sites

I would have liked to have seen Twitter and Facebook (possibly even MySpace and LinkedIn too) go head to head and Digg go up against other bookmarking sites such as StumbleUpon, etc. My reason for this is that typically you are more connected with people on social media sites than on social bookmarking sites. Social bookmarking sites are cluttered with millions of links people are sharing with others they may not even know. So it’s a less direct form of sharing than say Twitter or Facebook where you (usually) have a more established relationship with the possible viewer clicking your link. Most bookmarking sites have a lot of users who are lightly “browsing” content and clicking on something that may sound interesting but then quickly clicking away if their interest isn’t peaked. On Facebook for instance if I share a video, only people who have some sort of relationship with me are going to see it and are therefore more likely to watch more of the video. So it would have been nice to see a comparison of apples to apples.

Yes its video…but what is the content?

This may seem like a stupid question but if 75% of the videos profiled were of a cat playing the piano…what does that actually tell you? It would have been great to cull out the user generated content and just focus on videos that have some sort of at least a vague marketing purpose, whether its a direct sell on down to the nebulous but humorous branding video. I realize this is nearly impossible to achieve, however including all that user generated content as part of the research definitely skews the numbers. Let’s face it…if you upload a video of your dog barking at the TV – you don’t really care how many people watch it to completion but if you put a branding video online with a call to action – that’s information you want to know.

Time of day comparisons

Just like email marketing where you have better days of the week or times of day to send your email to get ideal open rates or click through rates, social media works much the same way. It would have been interesting to see over a three month period what days of the week and hours of the day had higher engagement rates.

What the numbers tell me

Ultimately the numbers don’t matter. Well…they matter but its a giant brush stroke of the entire social media space, not necessarily YOUR demographic and how THEY are engaging in social media. So you have to keep this in mind when you delve into these numbers. If the key demographic you market to is predominantly on MySpace but you are just sharing your video link on Twitter because this research report told you to – you could be missing your mark.

Personally over the past 3 months, SmartMarket Media has had better engagement rates from LinkedIn (2 minutes 35 seconds) followed by Twitter (2 minutes 32 seconds),  Facebook (1 minute, 40 seconds), (StumbleUpon (0 minutes, 45 seconds) and Digg (0 minutes, 37 seconds). Obviously we have a much smaller sampling (hundreds of visitors rather than millions) but it just goes to show you need to know your customer base and engage where they are engaging.

What do you think? What do these numbers tell you?

How do people discover videos online?

by

Once again TubeMogul has released some pretty awesome statistical analysis regarding how people find videos online, from embeds on blogs to video search engines. For a two-month period, they recorded inbound URLs for a sample of over 35 million video streams from six top video sites. But which sources drive the most video views? For the full report from TubeMogul Industry Analysis, continue reading here. Here are some of the highlighted statistics that I found truly interesting:

45% of viewers find a video by direct navigation to a video site (i.e. going to YouTube and searching or clicking around the featured or related videos).

No surprise here given that over 10 hours of video footage are uploaded to YouTube every minute that going directly to the video sharing sites and searching would be the top method of finding videos.

In terms of individual web sites referring traffic, no single source dominated, here are the top 20 individual referrers:

Site Share of Video Referrals
google 7.19%
yahoo 2.12%
facebook 1.93%
myspace 1.55%
digg 1.49%
stumbleupon 1.13%
msn/live 0.92%
blogspot 0.78%
aol 0.43%
reddit 0.29%
truveo 0.22%
flurl 0.21%
blinkx 0.19%
ask 0.19%
comcast 0.16%
twitter 0.15%
wordpress 0.15%
cnn 0.12%
wikipedia 0.11%
ovguide 0.06%

However, since there are a limited number of players in certain areas online, TubeMogul was able to infer that:

  • 11.18% of all traffic comes from search engines
  • 3.66% comes from social networks
  • 3.19% comes from social bookmarking sites
  • 0.63% derives from video search engines
  • 0.05% is directed from Email/IM
  • 80.88% makes up the rest of the referred traffic…of this mix it is almost completely made up of blogs from the thousands of different blogs they scanned.

Here are the really interesting facts here:

Digg beats StumbleUpon by nearly 0.4% for video referrals

I wouldn’t have guessed that. When I share videos on both social bookmarking sites my traffic from StumbleUpon is nearly triple the traffic I receive from Digg. StumbleUpon is my #4 traffic source for the website (which of course does include my blog posts) bringing in 9.97% of my site traffic while Digg is my #10 source of traffic (also including my blog posts) accounting for about 3.85% of all my site traffic. About half of my bookmarks are for videos while the other half are for blog posts (possibly even this one will end up on both). Of course this is just me and I am not profiling over 35 million videos for my statistics.

0.05% is directed from Email/IM

This I find staggering to be so low. One of the easiest and most cost effective ways to get people to share your videos is through email marketing – particularly to an existing base of people who have opted in to receive your email newsletter. In a recent post about integrating video into your email marketing campaign I found that there was a significant 175% increase in click-throughs when video content was included in an email campaign. It sounds like a lot of people are missing the boat on this possible distribution channel.

Blogs sourcing most of the 80.88% of all referred traffic in this sample.

To those trying to make a video go viral, this should be telling you to reach out to relevant bloggers who could help you tremendously with the push for video views.

0.63% derives from video search engines

This is bad news to the ever increasing number of online video search sites that seem to keep popping up promising to help your video go viral or supposedly helping you search. With less than a 1% take, that doesn’t exactly fill me with confidence. I’ve long held that most of these sites have very little value to the online video producer – this study just proves my theory.

So the real take-a-way here…

…is engaging bloggers to work with you by sharing the video with them. If nearly 81% of video traffic is coming from blogs it only makes sense to try and engage relevant bloggers to share your video. The other real key that isn’t really discussed is to make sure you optimize a video’s meta-data to ensure it can easily be found by those who are searching.

 

 

How Much of a Typical Video Online Is Actually Watched?

by

One of the frustrating things about online video view counting is YouTube and most other video sites count a “view” regardless of how much of a video is actually watched. So that got the research staff at TubeMogul thinking…how much are people actually watching before they click away?

The results from their study are pretty amazing: most online video viewers watch mere seconds, rather than minutes, of a video. All going back to the point I try to stress with every one of my clients that brevity is key. Click the image to see an enlarged graph of TubeMogul’s study findings.

For the full report from TubeMogul Industry Analysis, continue reading here. Here are some of the highlighted statistics that I found truly interesting:

Most videos steadily lose viewers once “play” is clicked, with an average 10.39% of viewers clicking away after ten seconds and 53.56% leaving after one minute.

I found this one surprising but not a complete shock. Most of our online marketing videos fall under 1½ minutes. The fact that over half of all viewers they sampled drop off after the first minute is interesting. How many were because the videos were incorrectly described or tagged and how many were just “casual browsers” checking out videos randomly. Properly titling, describing and tagging your video is one of the most important steps to getting found by the right people who are more likely part of that 46.44%. If they were looking for your content, they are far more likely to watch it to completion…as long as you keep it short.

A three minute video that has a post-roll ad in the final seconds, for example, will only be viewed by 16.62% of the initial audience, on average.Another takeaway is that overlay ads should be displayed as early as possible in a video, preferably within the first few seconds. On YouTube, where most overlay ads appear at about 10 seconds in, 10.39% of a video’s initial viewers are not likely seeing the ad.

Alright I’m going to go off on a little tangent here – I am not a fan of pre-roll or post-roll ads on video content. Personally I just think it is too much of an interruption to the viewer who clicked on a particular video to watch…NOT the advertisement tacked onto the video. I wonder how many viewers are clicking away because they are annoyed by the interruption of overlay ads on the video they are trying to watch? Especially if it is interfering with the content. Social networking on video sharing sites is all about inbound marketing or letting the community find the resources they are looking for by properly tagging videos with the keywords they will be searching for. Overlay ads to me are more of the old school of outbound marketing like television commercials, print ads, etc. a shotgun approach to hit as many people as possible with their sales pitch without regard as to whether it is hitting their target market or not. Sorry…my diatribe is over now. Anyway if you are going to engage on overlay advertising, this is a staggering reason why you shouldn’t even consider post-roll overlay advertising.

TubeMogul has once again impressed me with their industry leading research and produced some impressive results from their study. What about your own viewing habits? How long do you typically watch a video? How do you feel about overlay ads?