Time on site and conversion rates

Have you heard the old advice about how you've only got a few seconds to capture someone's attention online?

Well you'd be a fool to ignore that after what I've got to show you.

Today I stumbled upon some interesting data while I was trying in vain to get some reliable intel on my video effectiveness. Unfortunately, while there are a few hacks using event tracking, Google Analytics falls short in that area.

Thankfully I've found Wistia which I'll share with you in the coming weeks. This will hopefully get me more of the data I'm after.

Anyway, while pushing some data around using the custom segments feature of Google Analytics, something struck me like a bolt of lighting. I was trying to determine how effective our video was at converting people, and if there were any particularly weak  parts of the video that people might get bored with and click away.

Take a look at the graph below.

What I've done is segment people based on how long they've spent on the site and looked at their corresponding conversion rates. I've broken it down into effectively 4 segments. Check out the results.

  1. All visits = 1.01%
  2. On page 1-5 minutes = 4.67%
  3. On page 5-10 minutes = 10.30%
  4. On page 10 minutes or more = 8.52%

This was a 1 page sales page with a 10+ minute video and had the first part of the sign-up form below the video.

What I wanted to see was whether or not the length of the sales video influenced the conversion rates.

There are a few interesting things about the data. For example, how it seems that people finishing the video didn't really impact the conversion rate. It actually had a higher conversion rate (although half the total conversions) for people who viewed only half the video.

Perhaps they were already convinced so we may as well have had a shorter video.

However, the most startling fact is how important it is that you capture people's attention IMMEDIATELY and engage them.

Let me break it down another way.

Here is our conversion rate segmented by people who stayed for more than 1 minute, and those who stayed for less than one minute.

Striking isn't it? Clearly we weren't doing a good enough job to engage them initially, however, once they were engaged, they converted VERY well.

Most of this traffic came in from large email promotions, so they should have been pre-sold to a certain extent.

We made sure there was strong continuity between the pre-sell and the landing page, so this should have been OK.

They bothered to click on the email link, so they must have been interested in the first place to a certain extent?

So what was the problem?

Here are a couple of guesses that could have been potential issues.

  1. People saw that it was a video and simply didn't want to watch it.
  2. The video was too long (and we had a video controller that showed the length) so they could see it was long.
  3. The first minute of the video was not compelling enough.
  4. The visitors connection was too slow to start the video within a reasonable threshold.

Whatever it was, it is clearly a problem. I wonder how it would have gone if we just had text and screenshots?

Do you have any other ideas?

As an aside, it has really driven home the point for me that when trying to improve conversion rates, we need to focus primarily on the things that matter in the first minute. They seem to have the biggest impact overall when you look at how much they have skewed the conversion rate from over 7% if they stay on the site for 1 minute, down to just over 1% for all visitors.

February 18, 2010

Comments (3)

Said this on February 18, 2010 At 10:26 am
Hi Allan,

This is all very clear...and I appreciate that you are testing and experimenting to help us, so we do not waste time.

Here is my problem. I generate a lot of clicks to my merchant partners. So people are obviously staying on my site long enough. For example, this month one particluar merchant (receives maximum clicks) has already received 158 clicks through my site...without a single conversion!!

I am at a loss to understand why people are not buying. And should I continue with this merchant or drop him?

Brendon Gearin
Said this on February 18, 2010 At 01:16 pm
Awesome use of segmenting!
that's a very functional/actionable report.

I've found depending on your audience.. when going into offices/b2b, the video playtime just tanks. (now the only way we could come to this conclusion.. was by segmenting there emails into personal/work & drilling down based on the time of day).

Having both the video and the text underneath has proved both really effective and completely useless at times for me.

Starting with a Image of the video and having it look like it needs to be clicked to play has worked well. When its clicked it writes in the video.

At least you know that they wanted it to play & and there not scrambling for the mute button - which has been a complaint before when asking users (asking... real high tech). It loads quickly to.

You really just can't assume anything can you... It was much nicer before i had data to back things up. ignorance was bliss ;) - that said its better to know.
Said this on February 28, 2010 At 02:35 pm

Surely video is the same as a sales copy?

Its' a matter of testing even if it is making a video shorter or a completely different video so you can split test. Find the best one and test again.

New comments are currently disabled.

Newsletter Sign-up

Sign-up Now!

How to make $1000 a month from scratch

Sign-up for our newsletter and receive our 77 page eBook "How to make $1000 a month from scratch".

Best of all it's FREE!

Your Host

Learn from an affiliate veteran.

Your host, , has been earning a good living from affiliate programs since 1998.

Want a BETTER article writer? We will write good articles just for you.

Content Kingdom