Category Archives: Google analytics

Do you really want user feedback? Part 1

One of the current mantras for government online services is to say that they are customer focussed.

This is an excellent aspiration and it is becoming a standard protocol to build services around iterative user testing.

However once a service is up and running how do we ensure that we are still meeting our user needs?

This is not a definitive list but a description of some of the things that we have done with our website which might be worth considering.

Let’s break this down into indirect (this post) and direct user feedback (the next post).

You will probably collecting some of this information already. For example if you use an analytics package such as Google you should be monitoring the number of ‘page not found’ events.

Typically this is when some content has been moved or deleted and the person who bookmarked it uses the old link and arrives at a dead end. So every month I (and I do mean myself) go through these records and try and work out what the before and after for these links should be. I send these across to our hosting provider to be uploaded onto our server. If things go to plan, and they are, the number of these ‘page not found’ continually decreases.

Yes this is incredibly tedious but the benefits are that it does really focus on fixing the direct problems that users have. At least it does give sense of satisfaction of making something better. It also makes you more aware of your content what is still of interest – ‘why did we ever delete that page?’ Also any more fundamental problems become evident if the same kind of event keeps recurring.

As an extra touch we have added a small feedback box on the 404 page to ask if the user can say what they were trying to find. Hopefully this shows that we are user focussed and genuinely want to help them find our valuable content.

What else?

Do not forget that your other analytics data will tell you how long people stay on your pages and if they visit more of them or ‘bounce’ off to other sites. Did they come to your site by accident or deliberately. Was your content boring or fascinating?

There are also a number of remote monitoring tools that are useful unobtrusive and do not collect personal data.

Google has a handy ‘in page’ view that will show where users clicked on an individual page. We use software called Clicktale that does something similar though with a higher level of sophistication and range. It works very well for example to see how far users scroll down a page or what they click on. The latter can be very useful on forms to see how far people progress through them and when they give up. Clearly we should then act on these insights.

Of course the other key data is page views and downloads. This is your users voting with their mouse as to what content of yours they find interesting.

Another neat trick, which I have not tried yet, is attempting to measure if there are any pages that never get visited. To do this you need to be able to generate a list of all your pages and match against an export from your analytics. Take a look you might wonder why you ever set up that x page saying how great your organisation is about y?

You probably have other tricks so why not let me know your best ones.

Next week I will talk about the feedback that we have generated direct from users.

 

 

 

Why bother improving our web analytics?

I get a number of colleagues asking ‘have you got the web stats on our report?’

To which I answer ‘of course I have, if I give them to you what are you going to do with them?’

Naturally I am being slightly flippant.

Google Analytics is great for providing stats on individual outputs and most authors work on one output at a time.

In fact if I see a noticeable tread with a particular report such as Universal credit last week which was a bit of a record breaker – I try to pass the information on to the team and the press office for their information.

However how useful is this information on its own particular when another question is often ‘how does our report compare to others in the same sector or in a different sector?’

My next flippant answer pops out – ‘so if I tell you, will that influence which report you will write next?’ Well of course not.

So this is my digital dilemma – I have some of the data but either it is not in a great format, or it might not tell us very much. What to do?

Well we did spend some time a while ago logging some digital stats against each report. However it was rather time-consuming and I still came back to the same question – what is the point of these stats if we cannot change our behaviour?

However not being easily deterred we have tried a slightly different track after reading various articles about GA custom variables. So my excellent colleague Rob Skilling created a custom variable which in WordPress draws data from existing fields in the database in particular the audit sector of reports such as defence or health.

The resulting report looks like this for page visits for the top ten value for money audit sectors in the last two weeks:

CV sectors

So what does this tell us?

Well it shows the relative popularity of the different audit sectors which might have some limited value, but at least it is generated automatically, so no new work to update colleagues if they are really interested in the relative popularity of the sectors.

So what could we do next?

  • We could add in the report titles against each sector.
  • Following the inspiration of Peter Jordan at GDS we might then create a report for the other content that is more guidance based – then perhaps compare them against each other and see the relative popularity of audit work versus guidance.
  • Where the NAO reports against some bigger thematic topics which run across audit sectors we could label the content appropriately and see how popular this thematic content is.

Could be interesting?