At ProOnGo (mobile + web QuickBooks-compatible expense reporting) we recently did the 100th upload to our YouTube channel, and it seemed like an opportune time to step back and look at the efficacy of marketing via YouTube -- what worked, what didn't, and what to strive for in the future. Maybe our reflections will save you from learning some lessons "the hard way", which is the reason for this post.
Our Channel So Far
By the numbers, our channel has been alive since March 9, 2010, and has generated 46,331 video views, largely this year (29,619 in 2012), with 78 public videos and 22 unlisted, and 13 videos with over 1,000 views each. That makes our channel but a small fraction of the popularity of some legendary B2B SaaS channels that we admire (DocuSign @ 566,990 views, Box @ 1,916,922 views, Xero @ 125,774 views), but we've got enough data at this point to make some conclusions about what's worked, and what hasn't.
Right now we're looking at three factors:
- Driving organic traffic to our channel
- Generating clickthroughs to our www.ProOnGo.com domain
- Measuring+raising the likelihood that those clickthroughs lead to subscriptions
That's far too much to cover in one post, so let me start with some of the lessons we've learned in terms of driving organic traffic (in some future post we'll get to the other two topics).
Driving Organic Traffic To a YouTube Channel
On the organic front, we're of course up against the black box of SERP ranking, which -- like Google searches -- is based on a variety of factors including popularity, relevance, and in the case of YouTube, viewer engagement. As ambiguous as that sounds, there are certainly a few techniques that have worked for us, to drive increased organic traffic.
Uploading a Transcription File
I'm always amazed when I encounter a YouTube video with no description, no transcript, and no annotations. Now, I know the search folks at Google are super smart, but they can't possibly show you extra love in the search results if your content isn't easily searchable. It's obvious that the description would matter, but do transcripts and annotations matter? Absolutely. To illustrate, in a video about our custom Excel expense reports, somewhere in the video I say the words "a way to get your expense information directly into quickbooks", and that verbatim phrase is in the transcript file that I uploaded in the advanced section of the video upload page. Try searching for that exact phrase in quotes in the YouTube search bar, and guess what the first search result is (excluding advertisements)? Yup, the video in which our transcript includes that exact text:
Of course, the particular phrase that I used in this example would be a pretty "long tail" search term, but the technique can be used on much more practical search terms as well.
Making Embedded Views Count
Of the 29,619 views on our YouTube channel this year, 16% of them (4,806 views) came from embedded players on our web site, like this one:
We used to use a more proprietary playback solution on our website, but the differences in feature set relative to embedding a YouTube video were not differences that mattered much to us, so we switched to embedding YouTube videos. In addition to getting rid of a small cost that we were incurring, it also leveraged the power of our most loyal viewers: people who were already knee-deep into using our solution, and just needed to watch a particular tutoral. Their participation in viewing YouTube videos both drove up our view count, and we suspect also contributed to raising our viewer engagement. Why? Because these viewers who were already knee-deep in our solution were surely more likely to watch a video to completion than a random visitor that happens upon our videos through some other means.
Which video would you click amongst the following search results? I suppose you could say "it depends what I'm searching for", but I'll say that I wouldn't be surprised if you clicked on either the one with the green background or the blue background with the cartoon-ish cutout. Why? Because in a sea of screencast-style videos (at least in our B2B SaaS space), a video with a bold background that looks different than the rest, seems to outperform the others:
I suppose if I were competing in a space where most videos were a live video recording of a 'talking head', I would consider making an animated video. And, I suppose if our competitive space were cluttered with animated videos, maybe I'd be looking at making a screencast (albeit colorful, if possible). Whatever stands out from the sea of bland search results will probably land a little bit of a clickthrough boost. Furthermore, when you upload your video, YouTube lets you choose the thumbnail from amongst three snapshots that it somehow automatically chooses from your video (anyone know how it picks them? I don't), and of course you'll want to choose whichever one you think will most stand out from the crowd:
For example, in the above picture, our three thumbnail options include one that shows a screenshot of the ProOnGo web app (we already have dozens of videos with a thumbnail indistinguishably similar to that one), a screenshot of QuickBooks Online Edition (again, we've got plenty of videos with that as a thumbnail). The alternative, perhaps only slightly more exciting, but at least different from the majority of our videos -- is a screenshot from a part of the video where we show off our Google Calendar => QuickBooks integration link. So, that "slightly different" screenshot wins (although I'd rather have something even more distinct).
Tuning Your Viewer Engagement
I've already mentioned that viewer engagement is now a major factor in YouTube SERP ranking -- so you ought to put some thought into how you can make sure that your viewers routinely watch videos through to the end, or at least close to it. Some of that is about creating content that leaves a reason for the user to keep watching, but part if it is just working on your feedback loop so that you know what's working and what isn't. Thankfully, in your YouTube Analytics page, under Audience Retention, you can see exactly what the falloff is on a time-series basis, for viewers of your videos (provided your video has some minimum number of views - I think 100). For example, here are the audience retention stats for this video:
And, below the graph, is a data table showing the average view duration by geography, like:
Based on the above specific data, if we had that particular video to do over again, we would make the fade-to-black much speedier, and would also cut down on the logo screen near the end. I'm pretty sure that if we did those two things, we could get our "average percentage viewed" up to nearly 90% - which would make it one of our strongest ever, in terms of viewer retention. You can also drive an improvement in viewer retention by adding Playlists, and adding them to your Featured Playlists in the right-hand-side of your channel:
Why bother with playlists? Because the folks that are interested enough to be viewing a playlist -- people that are literally watching multiple consecutive videos from your channel -- are people that you want to harness for their viewer engagement. You want them to influence your viewer engagement statistics as much as possible, and if you get them interested in a playlist containing several videos, you'll do exactly that.
There is much to know and learn about marketing on YouTube, and although I certainly don't claim to be an expert -- at this point I'm confident that just about any business could grow their organic traffic on YouTube by using the above techniques. I've left out the 1000 other techniques, theories, and ideas that we tried along the way that didn't work, and focused just on some of the ones that are really working for us at ProOnGo.
I hope to be back with future posts about driving traffic from YouTube to an external website, and posts about getting "the right viewers" to clickthrough -- folks who are reasonably likely to become paying customers.