There is a weird kind of syntactic flow that pervades any industry in which the players all seem to converge on using the same words at the same time.  Right now, no one "does", they "execute", and its been years since anyone had any "problems"; they're generally too busy handling "challenges" and "issues".  But anyway, I digress.

Whenever I'm in a conversation that moves on to the subject of usability or conversion, I dread the moment  when someone who should know better trots out the phrase "engagement metrics" - largely because it is usually followed by a round of self satisfied nodding, and a chorus of bemused agreement before moving on to something else.  This is generally because once the phrase de nos jours has been delivered, no-one is entirely certain about exactly what it means or what to do about it.

It's all quite simple of course, and it comes down to measuring interaction, about analysing the way in which a website works to deliver a preferred outcome, and searching for ways to improve it. 

Where to start

Oddly enough, you need to start at the end. Define the goal that you want to achieve - whether this is a sale, a sign up, a click on an advert or a download.

Where to next?

Once you know where you want to end up, you need to think about where to begin.  In most cases, this is before the visitor arrives at your website on a search listing or another advert.  Once you know the two ends of the user journey, it is possible to define that journey and map out the various stages.  It's worth spending time on your own website and mapping out the various stages that the user will need to go through: what pages they need to visit, which links they will need to click, how many forms they need to fill in, and how they can find the information that they need to make a decision.

OK, so where's my engagement metrics?

Well, this is where they come in, and they're not as grand as you might expect.  Click through, bounce rate, click path, drop out rate, conversion rate. The same set of statistics that you always see on a website.  It's all about context.  Rather than measuring these different statistics in isolation, you map them against a specific goal.

Benchmarking each of these statistics is essential before you make any changes, as you need to understand the current situation with your website.  Each of the different statistics is impacted at a particular stage of the user journey through your website, and need to be mapped against them.  For example, CTR is a measure of how many people click on your advert, so it comes at the start of the journey; conversion rate relates to the other end of the journey.

Optimising the journey

To my mind, you should always start at the end of the process and work forwards.  My reason for this is that if you put a lot of work into getting more traffic to your website, but aren't able to convert it because you haven't got the back end working right, then you're wasting that effort.

At each stage of the process, you should test variations of the page in situ to examine what impact your changes have on the particular metric that you are looking at, but you should only ever make one change at a time, and leave it in place for long enough to get an indication of how the change has impacted on the statistic you are measuring.  

Be prepared for frustration...

In most cases, a thorough process tends to be a slow one - changing one element of a page at a time, testing different variants, and then testing the impact of further changes earlier on in the booking process before revisiting earlier changes again - and so on ad infinitum - will lead to plenty of head banging moments.

...and plenty of satisfaction

Sometimes, you will hit on something that will make a huge change to conversion rates on your website, and this is what its all about, I've worked on sites where the conversion rate has gone from 4% to 6% - which doesn't sound like much until you call it an increase of 50%.