Eric Guerin's Inane Ramblings

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?