Every website has points of failure. It is inevitable. The question is do you know what they are and are you doing something about them?
Nobody likes to think of themselves as a failure and no website owner likes to dwell too much on the shortcomings of his website. However, all websites have weaknesses and it is important we know exactly what they are in order to do something about them.
What is more, all websites develop new weaknesses over time. As content is added, structure is changed and designs are tweaked, we introduce problems into our sites that were not previously there.
We therefore need a system in place that continually monitor for failures so that they can be addressed quickly before causing too much damage.
Torian, Shutterstock
Implementing such a system does not need to be onerous. It simply consists of three simple reviews that should take place on a monthly basis. These are…
- Monthly user testing
- Identifying dropout points
- Analyse search queries
Let’s begin with user testing.
Monthly user testing
Steve Krug’s latest book Rocket Surgery Made Easy takes usability testing in a radical new direction that is both perfect for identifying the constantly evolving weaknesses in our sites and also inexpensive to implement.
He proposes an ongoing programme of quick and dirty user testing that takes place every month. The idea is that on a set date each month you schedule a morning of user testing with only 3 participants.
Your entire web team watches the user testing and analyses the results over lunch. By the end of lunch you will have identified the biggest problems that need to be addressed before the next month’s testing.
Jason Alley, Flickr
This is a great approach for detecting and eliminating problems on your website. It is…
- Lightweight – Only requiring one morning a month for testing and debrief.
- Regular – Ensuring that it picks up on problems that creep into the site over time.
- Action orientated – By the end of the session you have a plan of how to tackle the failings in your website.
- Fixed – Because the user testing is always on the same day every month it is less likely to be bumped for more ‘important‘ things.
- Prioritised – With only a short time to test and debrief the emphasis is placed on the most important failings rather than being caught up in nuances.
Although user testing is useful it should not be used in isolation. Testing only 3 participants enables lightweight and regular testing, but it does have its limitations. Ultimately 3 people are not going to be representative of the whole and even if they were, they are not interacting naturally with the website. For that you need to look at web statistics.
Identifying drop out points
Website statistics can be very enlightening when used correctly. Unfortunately few people know what they are looking for and so either give up try or never get past page views and unique visitors.
One aspect of web stats that is particularly interesting is exit points. Where do users leave your site? This may help identify potential points of weakness in the site and areas you wish to test in your next user test session.
Most statistic packages make it easy to view a list of top exit pages. However make sure you are viewing pages with the top percentage of exits, because popular pages will be exited more often. Look for pages that are popular but are also exit points. These are the biggest problem areas. Google Analytics actually allows you to view pages that meet this exact criteria.
Once you have this list ask yourself the following questions?
- Is this the final page in a call to action process? For example is this an order confirmation at the end of a purchase process. You would expect pages such as this to be a typical exit point.
- Has the user visited any other pages before exiting? If the user has viewed only this one page then the chances are they were simply at the wrong site. To make sure check the dwell time. If they were only on the page a few seconds before leaving then it is fair to conclude they had arrived by accident. Also check how they arrived on the site. If they came via a search engine, what search term did they use? If the term wasn’t relevant to your site then don’t worry.
- What type of content does the page contain? If the page is a blog post for example the chances are the user was more interested in the content than anything else you offer. They will read the one article and then leave. Obviously this is not what you want and should work hard to encourage them deeper into the site. However, the reality is that pages of this type will have a higher exit rate.
- Is the bounce rate significantly higher than elsewhere on the site? If so this could indicate a weakness in the page.
If users have viewed multiple pages and then given up on a particular page, it is a problem which needs addressing. Possible problems could include…
- No obvious next steps – Are you telling the user what to do next.
- Too many options – Too many choices can cause users to give up.
- The content fails to convince – Is the content of the page telling the user what they want to hear, or are you saying something that alienates them?
- Too much content – Is the page packed with large amounts of densely written copy. This can put users off.
- The content is not relevant – They have arrived from a search engine query and the page hasn’t provided what they want. Instead of looking elsewhere in the site they have returned to the search engine to view another result.
Unfortunately although web stats can be very good at identifying problem pages it is not so good at diagnosing the cause. That is why it is important to user test as well.
Of course another possibility is that a user has given up simply because the site doesn’t appear to have what they are looking for. That is where you need to analyse search queries.
Analyse search queries
What users searched for provides an excellent insight into potential failures of your website.
Take for example the top search queries that lead people to a page with an exceptionally high bounce rate. Which of those queries cause most of the bounces? What are the user expecting to find on the page and do not? The page must be relevant in someway otherwise the search engine wouldn’t refer them. However, perhaps the way you are presenting the information is wrong. Does the search term give any indication of how you could be presenting things better?
What about the terms that create the highest bounce across the site. Sometimes users find themselves on a site that a search engine believes is relevant but doesn’t directly address their issue. Could you retailor your content to more directly address these search queries?
However it is not just external search engines you need to be looking at. What about your internal search engine?
When a user arrives at your site after entering a search term into Google, you can be sure that at least some content on that subject exists. Otherwise Google would not have referred them. However, when a user types something into your site search, there is no guarantee it will be a topic you have addressed at all. This is a perfect way to identify content lacking from your website.
Repeat each of the top searches yourself and look at the results. Are you addressing the search terms directly? Are you addressing them at all! If not you have identified a weakness that needs dealing with.
Rinse and repeat
The secret to success when it comes to irradiating the failures on your site is to establish a monthly cycle of work. Each month you need to user test, delve into your web stats and analyse your search results. By combining these three techniques you should be able to establish a programme of work for the coming month. By repeating this process month on month you can slowly evolve your site so that its shortcomings have less and less of an impact on conversion.
Of course, this is not all that can be done to identify problems with your site. These are simply those that have worked best for me. What about you? How do you find the shortcomings on your site? Do you even do this kind of ongoing maintenance? Let me know in the comments below.
