With the rise of digital analytics, KPIs have been a focus for marketers and unfortunately it is common for metrics to be reported up the chain to the board, without sufficient explanation or segmentation.
Specifically, aggregated site wide metrics such as bounce rate, conversion rate, and time on site can have excessive value placed on them, which in turn skew opinions and actions to the detriment of overall site goals and performance.
(Note: It is probably worth just explaining what ‘rate’ means, and in most analytics packages it is per session (or visit) rather than per user, this means that within a 30 minute timeout window, whether they completed an action and whether they did it over separate sessions.)
Bounce rate is bad if not segmented
My favourite example for a dysfunctional sitewide KPI is bounce rate – high bounce rate is bad, low bounce rate is good right? This engagement metric is typically calculated by someone only seeing one page, on most sites you want users to visit more than one page when they visit?
Critically if the information should be found in a single page visit then a high bounce rate is a almost a positive thing, but you still want to add in related links and information that helps them navigate further – is a bounce rate borderline useless in these cases?
As Avinash Kaushik, Digital Marketing Evangelist for says…
Personally I would say segment and test, never look at a site wide rates and don’t compare yourself to aggregated site wide averages without context – you can look for areas to improve from the 404 page right through to product landing pages to spot issues, but the moment you start looking at a site wide bounce rate is the moment it becomes a useless metric.
And don’t try to improve it, instead improve experiences and facilitate the user journey.
It is also worth addressing that long time myth that your bounce rate has an impact on your rankings, your Google Analytics bounce rate has no relation to your rankings. Whilst this is not a direct signal, however users who ‘pogo stick’ (a technical term for users being on your site for only a few seconds) might be a signal.
So, bounce rate – good for finding problem pages, improving PPC landing pages, optimising your 404 page; bad for reporting at a site level or making meaningful industry comparisons.
Time on Site
It is also worth noting that ‘time on site’ is also a terrible metric – users are no longer on dialup, they can leave tabs open in their browser and go make coffee – this, as a sitewide metric has so little value now. If you have a user replay tool on the site you will see just how often tabs are left open, users today who wander off or browse on another tab aren’t really spending time on your site or pages. Additionally the way it is calculated is limited and single page visits are excluded.
Conversion and Intent
The other popular sitewide metric is ecommerce conversion rate. Realistically, stock status and pricing has more impact than almost any other element on conversion rate (if you don’t have it, or it is overpriced they it will be challenging to encourage users to part with their money) — this is another metric which if considered sitewide loses all meaning.
On a typical ecommerce site activity should be directed at improving the number and quality of conversions which may be a different outcome to selling to every user on every visit — users may not be ready to purchase but are ready to signup for a newsletter or engage socially. Like bounce rate, there are times when as a metric it works, and this is where multivariate (or A/B) testing should come in, to truly answer if the actions you are implementing are improving performance.
Going back to intent, it may be that many users are using your site for something other than to purchase (e.g. the jobs section). It’s also worth addressing if people returning for product information (such as drivers or product instructions). There is an excellent article on Conversion Rate as a metric here.
Matthew Curry from LoveHoney explains it well, in a video and it is well worth watching (although if at work be aware this imagery comes from their sex toy product range). Matt argues that you should look at user behaviours over KPIs. He also explains how some users go to his site for the pictures rather than the products which can vastly corrupt the data if these users are factored into conversion rate calculations.
Stop reporting sitewide KPIs that don’t ultimately inform
Ultimately the question should be what changes can I put in place to improve visitor performance, which isn’t always about improving the number of conversions, bounce rate, or dwell time, but about improving profit (taking into account factors such as cost per sale and average order value).
So the key is to understand your users’ intent (i.e. what did they come to the site for?). We should be thinking about whether their experiences reduce the chance of them being able to purchase again, and would their journeys be satisfied with a single page view. From this point we can start to move the journey further.
If we aren’t using the data effectively it is ultimately useless, not every user comes to your site to purchase, not every user recorded in analytics is even human (there has been for a few years an increasing number of bots deliberately spamming analytics accounts), and it is challenging to segment out all users who are not there to purchase.
Metrics are critical, but targeting users by intent genuine means improvements can be made.
Segment users, analyse, and investigate your data
Without segmenting the data it is simply not actionable – reporting on this data sitewide is like reporting on the weather worldwide. So stop immediately, segment, analyse and investigate. Almost all digital analytics platforms now allow you to segment users (at least on one device) so you can start to see whether the activity you do in one channel is paying off.
KPIs for measuring user intent
Not everything can be measured with conventional analytics. User experiences often need to be addressed at by talking, listening, and conducting research with real product users (ideally both yours and your competitors). Hit level data combined with machine learning can also provide insights into behaviour and access to this level of anlaysis is becoming increasingly accessible.
Of course offline behaviour is harder to capture but should not be ignored. If you aren’t tracking phone conversions with call tracking, or able to correlate onsite behaviour with store activity you can only ever see a small part of the journey, however the scale of this issue may be known and quantified.
Increasingly we can start to spot hints towards the users true intent, right now we have access to their location, time of day, device and all the actions they complete within a session (and if we are lucky in other sessions), we maybe able to combine this with much more relevant customer information and we can start to build a much closer picture leaving to us being able to answer the question.
UX folks often talk about “delighting users”, maybe one day we can have a ‘Delighted Rate’?
Are we now moving towards user centric reporting rather than website-centric reporting?
What kind of metrics should marketers be reporting on instead?