Let's start with some official YouTube terminology:
- Impressions = How many times your video thumbnails are shown on YouTube
- Clickthrough-rate (CTR) = Proportion of your impressions that turn into views
- Average View Duration (AVD) = Average amount of time viewers watch your video
TIP: If you increase your CTR, then you get more views on your videos. As simple as that!
Most creators will agree that first impression matters. For YouTube videos, your video’s first impression is its thumbnail (and title). It is generally accepted that optimising your thumbnail is a smart thing to do:
Higher click-through rate = More views
As we can all agree on that point, the question we will tackle today is:
Should thumbnails be optimised BEFORE publishing?
The ability to A/B test YouTube thumbnails after publishing has been around for some time, with tools such as TubeBuddy helping you do these tests in an automated manner. Even before the automated solutions were available, it has always been possible to test thumbnails on live videos for as long as YouTube has allowed thumbnails to be changed or edited.
However, this leaves a lot to be desired. The majority of videos experience an initial spike in views and receive a significant proportion of their views within the first few days. Existing thumbnail testing solutions often recommend that A/B testing should be avoided on new videos for this very reason. For example, TubeBuddy recommends waiting until a video is at least 5 days old before doing any A/B testing. This is a perfectly valid recommendation, with even YouTube stating that an initial spike often occurs for new videos:
“ CTR is often highest right after upload when some of your most passionate fans and subs might see the video at the top of their home page and subs feeds.  As total impressions grow beyond your core audience, CTRs are expected to decline until reaching a relatively stable percentage (this is totally normal).”
Source: YouTube Creator Academy
For those that have been reading closely, this means that existing solutions miss out on optimising and maximising the impact of the initial spike. This is the main reason why you would want to front load your optimisation process to before your video is published. By optimising the thumbnail before publishing, you give your video the best chance of maximising the initial spike and reaching a larger audience.
Other reasons why testing a thumbnail after a video is live is sub-optimal:
- Due to YouTube only providing data on a daily level, A/B tests on live videos need to be conducted over a minimum of 2 weeks to account for differences in viewership by the day of the week (e.g. viewership may be higher over the weekend when people have more free time). This means the earliest date your optimised thumbnail can be used is 19 days after publishing (5 days (can’t A/B test on new videos) + 14 days for A/B test).
- A/B tests require a minimum level of traffic or views. If your channel isn’t established, it may be difficult to collect enough data to make an informed decision as to which thumbnail performed the best.
- Lack of control — Live videos are influenced by many factors, many of which are outside your control. Any data collected on live videos will be noisy and won’t be purely driven by the thumbnail choice.
At this point, you may be wondering if it’s possible to test thumbnails before publishing, and if so, how? This is where Thumblytics comes in. Thumblytics takes different versions of your thumbnail (along with thumbnails from your competitors) and shows it to hundreds of people. The thumbnails you choose for your competitors act as the control in the experiment. By doing this, Thumblytics is able to calculate a click-through rate for each version of your thumbnail, giving you the data you need to make an informed decision as to which thumbnail will perform best.