Last week we had a very interesting roundtable talking about online conversion rate optimisation and the role that web analytics played in this. Anyone can download the briefing that came out of it .

One of the most interesting areas of discussion was around how to accurately benchmark conversion rates. The problem with overall site conversion rates, or 'average' conversion rates, is that they are skewed by so many variables that the term is almost meaningless.

For example, your conversion rates for any given product / service / outcome may vary considerably according to:
- Whether the user is a first time visitor or not
- Whether the user is an existing customer or not
- The time of day
- The source that referred the user to your site
- Seasonal factors (e.g. Christmas, Summer holidays)
- The weather
- World events (news, World Cup etc.)
- What your competitors are up to 
- Related marketing / PR activity you have going on
- Your pricing
- etc. etc. 

Given all these variables, how do you actually come up with conversion metrics which you can actually meaningfully benchmark or measure to assess whether you are improving things over time or not?

A few ideas put forward were:
- To measure micro conversions based either on goal achievements (e.g. did the customer get the right search results?) or on steps along the customer journey towards conversion that the user may take (e.g. browses product detail page and adds to basket)
- To use a baseline based on users referred to the site by "brand search" e.g. if Argos were to measure the conversion rates of customers who came to their site having searched on 'Argos' on Google. This baseline tries to group users together who might appear to have a similar level of brand knowledge, propensity to buy etc.
- To benchmark your metrics according to: your own site's performance over time; by reference to other sites you control; by comparing your performance with that of direct competitors. 

What ways do you use conversion metrics that you actually find useful and meaningful?

Ashley Friedlein
CEO, E-consultancy.com

The baselines you should use have to be internally driven. They are the only numbers which you can rely on if you are to base business decisions on the figures. 

The numbers we have published here are valid guidelines to form objectives perhaps, but they should be taken with the proverbial pinch of salt as there have not been any indications from any of the publishers as to what parameters were used to get those figures. Also the reasons Ashley suggests are all valid variables which differ widely across websites, even from competitor to competitor. So the only reasonable numbers you can use are the ones you begin with internally.

If you start with a 2% global conversion rate, your ideal situation is to improve on that figure, regardless of whether your competitor says they have a 5% conversion rate or not. 

You might even want to get more targeted and segment your audience. Many websites have different kinds of visitors. We defined people as suspects, prospects and customers for one client and measured conversion across those individual segments. The conversion rates were as you'd expect wildly different. However because we'd benchmarked in house what those conversion rates were we could begin to improve upon them.

This is where measuring things like micro conversions becomes important, as improvements in micro conversions (steps in a particular process) might improve the conversion rate of prospect to customer or suspect to prospect.