Tag Archives: web analysis

How to identify and analyse engagement in a system.

The principal challenge when designing a system is to create engagement. By engagement is meant,  the user not only spends a reasonable amount of time on your site, but also performs those actions that you want she to perform, for example, buying something, registering, or doing something off-line.

It’s no longer enough that the site attracts visitors, but it has to match a user’s needs, and this entails attracting the right kind of visitors and engaging them.

It could be said that the ideal site we want to create is one, which attracts a lot of users, which is efficient in engaging them and which encourages them to perform whatever we want. What follows is an approach, which can be used to analyse a system in order to assess what can be done to optimise engagement. It’s called the 2×2 engagement matrix.

The approach sets up a model with two variables: visitors and engagement, and the resulting basic  2 by 2 matrix contains four web-page types.

These four page types are characterised by how many visitors are attracted to them and the degree to which the visitors become engaged. They are labelled: bulls, sexy dolls, hooks and zombies.

Engagement matrix (by Rebeca Miranda)Bulls

These pages are completely optimised towards the user. They bring traffic to the site and they also foster engagement and lead to completion of interactive transactions.  They are characterised by the following:

  • High level of visits
  • High click through
  • Longer time spent on them
  • High bounce rate
  • High number o page views.
  • Low percentage of exits.

In general, the aim of an improvement strategy, will be focused on pages that are either sexy dolls or hooks.

Sexy dolls.

These are pages that receive lots of visits, but they are barely engaging.  They represent a good opportunity to improve the level of engagement because an audience is already in place. Basically the problem here lies in the content. Sexy dolls pages will likely show:

  • High levels of visits
  • Low CTR (Click through rate),
  • Lower number of pageviews
  • Higher percentage of exits


Writing relevant content. Perhaps the research and audience are not understood. Good research may be needed, as well as a proper segmentation of the audience.
Displaying other kinds of media such as videos, might also improve the level of engagement, but of course the media need to be designed to be ‘engaging’ for the specific audience.

Relating content with other parts of the website. This involves engaging the user more.  If the user is interested in a certain topic, then more content about it should be displayed.  Other example might be the use of contextual links

Making sure structural elements such as navigation menus are clear, intuitive, and simple.

Creating only a few elements to choose from. Navigation is more effective when there are not too many elements to choose from, and where the elements present are related to what the user is already familiar with.


These are pages for which there are few visitors but on which the user spends a great deal of time.  Users of these pages keep navigating indicating that their interests have been perfectly matched. However they represent a minority of the user base.

These pages are characterised by:

  • Low level of visits
  • High percentage of CTR
  • High number of pageviews
  • Longer times spent on these pages
  • High bounce rate
  • Low percentage of exits.

In addiction, the presence of such pages may indicate that:

A-    The audience has not been segmented correctly,

B-    The audience represents a different subset of the population than was assumed

In either case the starting point is a positive one: you already know that the content is relevant.

So what can one do to move these loyal geeks to became bulls?.


Improve the research and work on the SEO strategy – The content may not be sufficiently engrossing.

Place the calls to action in these pages

Try to understand why these pages are engaging in order to either:

–       Apply it to rest of the site, these pages may have elements that are absent in other pages, like contextual navigation, etc.

–       Promote this content across the site, it may be hidden, which may explain why it doesn’t have as many visitors. (e.g. users on these pages get specific information about certain services, and while trying to decide whether to buy those services, they may spend more time on the pages just to make sure, they’re making the right decision. Another reason may be they might be learning something through these pages and so spend more time on them.)


These pages are attracting neither visits nor loyalty. They are usually useful as a way of indicating errors: from 404 to just pointing out pages that need a review in order to establish:

1. If they are worthy to improve

2. What makes them so bad, in order to get clues to improve the rest of the site.

These can be recognized because they are characterised by:

  • Low level of visits
  • Low rate of click through
  • Short time spent on them
  • Low number of page views
  • High level of exits

Once one has identified which pages are problematic, it becomes easier to create specific remedies for improvement, as well as deriving valuable information about the overall performance of the site. Also by analysing the pages belonging to each of the defined categories across a system, we could obtain a ‘big picture’ of overall site performance, which could lead us to rethink how it should be developed.