Kickstarter Stats 101: Does the Kickstarter Video Matter?

It’s common knowledge in the Kickstarter community that having a project video is important to the success of a Kickstarter project. But very little data exists to substantiate this claim in more detail.

My intention here is to speak directly about the data that does exist, and try and illuminate any key insights.

In this blog post, I want to discuss the statistics behind the infamous project video, and identify whether or not a significant difference exists between the success rates of projects with videos and those without.

I also want to jump into whether or not the length of a video has any bearing the success rates of Kickstarter projects, and end with a few of my own thoughts.

This data greatly influenced my thinking when I created the video for my current Kickstarter project, so hopefully it can add some value to yours as well.

Okay, enough intro. Let’s get into the nitty-gritty details.

Kickstarter Success Rates Based Upon the Kickstarter Video

From the bar graph below, it’s pretty obvious to see that there exists a clear difference between success rates of projects that have a video, and those that don’t. But this is no surprise to many of you.

So the next things I would ask (in conjunction with the thought process used by great philosophers is this: “Are all videos created equal?” Obviously, a video that doesn’t exist is the same as a video that doesn’t exist, but for projects with videos, does the data say anything about them?

Kickstarter Success Rates Based Video Length

The graph below shows the success rates of projects correlated to the length of their video. You can see from this bar graph that Kickstarter videos with a duration between 1 minute and 2 minutes actually have the highest success rates while projects with video a duration greater than 6 minutes see a significant decline in their overall rates of success.

* Something to note about the graph above is that it does not show a 0%. This is a common method in statistics but an obvious consequence is that it makes the differences between the values “appear” to have greater variations visually, that are not actually accurate to the their numerical values.

Determining Statistical Significance

It’s easy to see that most of the success rates differ according the video duration, but what we really want to know, is if these success rates differ “significantly”. In order to answer this question I ran a simple ANOVA (Analysis of Variation) Test.

The three different colored bars in the graph above show the corresponding levels of significance according to the ANOVA. The green bars show span of video lengths that vary significantly higher than the others while the one red bar shows the duration that varies significantly lower than the group as a whole.

The level of confidence used in this significance test was 5%, which in a very rough and dirty way simply means – these results would only come out by chance less than 5% of the time. (The chance you would roll a 20 on a D20 with one roll)

With this in mind, the blue bars (though some variation is apparent) represent video durations that are not significant with respect to the group as a whole. The significance test shows that these variations could exist by chance alone with a frequency greater than 5%.

I should add, this 5% value is a pretty extreme guideline, which is commonly used in the medical and public health industries, and it may be overkill for our purposes here. But, just one more reason to trust the analysis.

Total Number Projects Verse Successful Projects with-respect-to Video Length

Just in case you’re curious, I thought it would be good to also add a graph showing the total number tabletop game projects verse the total number of successful projects for comparison.

My Personal Interpretation

Nothing I say here should be construed as truth or fact, rather it should give you a good reason to leave a comment explaining why think I am a big dummy and your own conclusions from the data.

The main point I would like to make is this: None of this data really matters if you have a great video, and for that matter, a great project.

After measuring traffic to my Kickstarter pages, I noticed that only small a proportion of the people who visited the pages actually watch the video anyway.

For Linkage, the video length was 2:15 and the completion rate was right at 33%. We had roughly 2,200 video views when the project closed (I didn’t check it the day it ended so the number is rounded down). With 616 backers total that’s a conversion rate of 31%.

The Linkage video wasn’t very exciting either. It was mainly just me talking about the game and explaining why I designed it.

With Peptide the video was 1:30 in length and the competition rate is currently right at 59%. We currently have 147 backers and 454 video videos which is a conversion rate of 32%.

The Peptide video was really funny (according to feedback from my backers – I am obviously biased here) and I had a lot of people say they watched it multiple times so it did seem to be a big hit.

It would be great to have a big set of data from other projects showing the number of video views and then correlate this to the number of backers, and see if the video duration has any bearing on the actual conversion rate. Maybe I’ll try that at a later time.

One thing that’s apparent is that there does seem to be a sweet spot right between 1 minute and 2 minutes. The current advice I hear most often going around the blogosphere is that a video between 1:30 and 3:00 is best. The data does give support to this, but a small drop is seen in the 2 -3 minute range.

Personally, I think 3 minute is just pushing it. After asking a question in the Kickstarter Best Practices Facebook forum, I had many people say that they don’t watch Kickstarter videos anyway because it’s usually a waste of their time.

I should also point you to Jamey Stegmaier’s blog about the Kickstarter Project video. He does a great job of laying out what should be put in a Kickstarter video and how it should be laid out.