Say it with a word cloud
Word clouds (also known as text clouds and tag clouds) are not just a cool way to show a bunch of tags they are also a great way to visually communicate all sorts of UX related information, from user stories to search keywords. In this article I outline what word clouds are, walk you through how to create them and discuss just a few of the many ways in which word clouds might be utilised.
What are word clouds?
Word clouds are simply a way of visually displaying a bunch of text as a shape. Usually that shape is a cloud, although plenty of word cloud generators such as Tagxedo now let you choose different shapes (like a square, love heart or even Abraham Lincon!). The frequency (or importance) of each word or phrase within the text is usually shown by its size. The more a word or phrase occurs in the text, the larger it will be within the word cloud. For instance, the word cloud at the beginning of this article was generated using Wordle by examining the text for all UX for the masses articles. You can see that words such as ‘usability’, ‘users’ and ‘website’ crop up quite a lot.
How do I create word clouds?
Creating word clouds is ridiculously easy as there are loads of good word cloud generators out there. Two that I can certainly recommend are Wordle and Tagxedo. Both allow you to copy and paste in text to create a word cloud, or to use an RSS feed. One thing to be mindful of is that in order to retain phrases, such as ‘UX design’ you generally need to replace spaces with the tilde symbol (‘~’). For example ‘UX for the masses’ would need to be input as ‘UX~for~the~masses’.
To make things a bit easier I’ve created a useful Excel spreadsheet that spits out the text for creating word clouds. All you have to do is enter (or paste in) a list of words / phrases with a corresponding weight (e.g. number of instances, % of instances, level of importance etc…). For example:
|Words / phrases||Weighting / Instances|
The spreadsheet will insert tilde symbols (‘~’) to preserve phrases and will output the right number of words / phrases to match the weightings. There is also a worksheet for utilising the advanced Wordle features that allow you to specify the colour for individual words and phrases (for the more hardcore word clouders out there).
If the word cloud generator doesn’t allow you to export the results as an image (shame on you Wordle) then you can simply screen grab the results using a tool such as ScreenHunter, or save it as a PDF using a tool such as CutePDF.
Suggestions for using word clouds
Here are some of the ways in which I’ve used word clouds in the past.
To show analytics information
Word clouds are great for showing analytics information and because you can usually export analytics reports to a CSV file it’s dead easy to plug the terms straight into the word cloud text generator.
The sort of analytics information that might be shown as a word cloud includes:
- Inbound searches (i.e. search engine searches that lead into a website).
- Internal searches (i.e. searches within a website).
- Most popular content.
- Most popular entry and exit pages.
- Visitor information (e.g. country of origin).
As an example of the sort of thing you can do the word cloud below shows the most popular Google searches that have resulted in visitors coming to this website.
To show user feedback
Word clouds are great for showing user feedback, such as survey and interview results (assuming a good sample size, such as more than 20 participants). The sort of user feedback that might be communicated using word clouds includes:
- Likes and dislikes for a product.
- Common issues with a product.
- Reasons for coming to a website.
- Chosen product descriptors (e.g. frustrating, professional, convenient) as captured using something like Microsoft’s product reaction cards (Word).
To show product usage information
Word clouds can also be used to show product usage information, such as how often features are used, or how popular content is on a website. For instance the word cloud below shows the various articles on this website by popularity (determined by number of views).
To show user information
Why not show information about users or their personas using word clouds? The sort of user information you could show includes:
- The different types of users (the larger the text, the bigger the overall proportion of this user group).
- Demographics information (e.g. country, language, occupational groups etc…).
- User needs and requirements (e.g. simplicity, reliability, speed, flexibility, easy to use etc…).
For instance, the word cloud below shows which countries visitors to this website have come from.
To show user stories
If you’re utilising user stories then you could even show them as a word cloud. The larger the user story, the more important it is to users. For example, the word cloud below shows some possible user stories for an ecommerce website.