What do users think? 

YouTube ad length isn’t something I see written about much, yet it’s inevitably one of the first questions clients have when we talk about advertising on the platform.

A 2012 poll of US voters reported by Mashable suggested that the majority of users were willing to watch an ad up to 15 seconds long before they got to watch their chosen content on YouTube.

Which is interesting, but ultimately how long a user thinks they’re willing to watch an ad for, and how long is actually optimal for ad length are probably quite different things.

More recently David Waterhouse of Unruly Media found that the average length of the top 10 most shared ads of all time is 4 minutes and 11 seconds, while at the end of last year Greg Jarboe provided a roundup of the top trending YouTube ads of 2013, which come in at an average length of 1 minute 44 seconds.

Which goes to show, if you can entertain people and tell them an engaging story, then they are willing to watch an ad for much longer than they think.

In all the articles above, and in the majority of other studies I’ve seen, the success of a YouTube ad is judged on the number of Likes and Shares it received, and whether or not it went ‘viral’.

It’s great to be popular, but for a lot of my clients the bottom line of whether an ad is successful is whether or not it generated sales with a good ROI.

The test

With this in mind we set up a test to see if ad length has an impact on conversion rate and ROI, and the results were pretty striking.

We created two ads that were almost identical for the first 15 seconds, but the first ad ended at the 15 second mark, while the second ad ran for a total of 30 seconds.

We then set up an A/B test of these ads against each other in our YouTube remarketing campaign for a month, before analysing the results.

Firstly, in terms of viewer engagement, it seems users are slightly more likely to watch more of a video if the overall length of the video is shorter.

34% of viewers made it to the end of the 15 second video, while only 32% of viewers made it to the half-way point of the 30 second video. I know personally I’m more likely to let an ad run if I can see the progress bar filling up nice and fast, so I’m not surprised to see this is the case.

More impressively, despite a fairly even distribution of impressions between the two ads, we saw three times as many Earned Views (where a user goes on to watch another video on your YouTube channel after seeing your ad) from the 15 second edit than from the 30 second edit of the ad. 

It was when we looked at the conversion stats where the difference in performance became much more apparent. Overall the 15 second edit saw 51% of our impressions, but drove 83% of our total conversions.

Conversions by campaign

Even more impressively, the 15 second edit achieved this with a CPA almost four times lower than the 30 second edit, coming in at £45 compared to £169 for the 30 second edit:

CPA by ad length

It’s worth noting that YouTube conversions are reported against views, and not website clicks.

Arguably then, because the 15 second edit is more likely to be watched, simply because it ends sooner, you could argue that it’s not surprising that a higher number of people exposed to the video go on to convert.

However, the conversion rate was more than three times higher from the 15 second edit compared to the 30 second, meaning that not only did a higher number of people convert overall, but per view they were much more likely to convert if they watched the 15 second edit.

Finally one of the most striking stats for me was that the shorter ad actually had a higher clickthrough rate than the longer ad, with a 0.84% CTR for the 15 second edit, compared to 0.72% for the 30 second edit.

This surprised me mostly because I assumed that the longer ad might get more clicks simply due to having a longer duration onscreen for the viewer to interact with it, but ultimately it seems that users were more likely to want to interact with the shorter ad even though it was onscreen for half the time. 

Ultimately ad length is going to be heavily influenced by your video content and advertising goals, and I’m sure an account manager at a creative agency would have a different opinion to me (he’d probably be talking you into hiring Jean Claude Van Damme and two big rigs right now…), but if you’re not testing different ad lengths on YouTube I personally suggest you start, and start with 15 seconds.