I was asked the following question recently and I thought it was worth sharing my answer in case anyone found it useful. The question is in italics and my answer follows.
Last year i saw your talk on designforconversion on that lovely boat in A’dam.
I’ve been thinking about the defintion of conversion rates and i think it’s flawed. I would like to hear your opinion.
The defintion i use now is a) number of users that completed a specified action / b) number of visits
a) makes sense
b) to use visits instead of visitors also makes sense because every visit is an opportunity to convert
However with visits that takes less than 10 seconds (bounce) there is not a chance for conversion. So it’s not fair to add these visits to total visits.
If you want to influence visits < 10 seconds your best chance would be to work on offsite promotion. F.e. your branding, seo/sea etc. The only onsite thing you could change to influence this metric is faster load times, design of the site (first impression) and clear headlines mayb. So that could be a reason to count these visits as well when calculating conversion rates.
What do you think?
Your definition is fine for specific purposes. I often use conversion rates segmented by engagement variables like time spent and single page views. However I stick to the standard WAA definition when someone asks for general reasons. Visits/converting visits or visitors/converting visitors.
To quote the document notes from conversion rate;
“Calculation of conversion rate requires consistency between numerator and denominator, both with units (visits or visitors) and with segments. For example, if a conversion is defined as “the number of visits where a purchase was completed,” then the appropriate conversion rate would divide those conversions by the total number of visits where a purchase could have been made (typically, total visits is used). If you had counted conversions as visitors who purchased, then the denominator would be total visitors. Further,if you are considering the number of visits from affiliate links that completed a purchase, then the denominator would be total number of visits from affiliate links.
More info can be found in the WAA standards document which can be found on the Web Analytics Association site.
Saying it’s not fair to add bounced visits is controversial. Bounced visits could for instance indicate a really strong disconnect with the search marketing advert and the landing page for instance or other poor relevance. In other words if your conversion rate is low because of high bounce it could be that you’re doing a bad job converting people.
This is where good design, but more importantly effective and clearly relevant messaging between advert and landing page is important. In my experience customers want to know the adverts that work and the ones that don’t and bounce is be a good KPI to pay attention to for that reason.
Hope this helps