Setting the standards
When leading an analytics department, it is important to find ways to promote good dashboard design that delivers a positive and insightful user experience to your customers. There are many ways you can approach this, a chosen methodology for how to approach development cycles, analyst training or quality assurance to name a few, however, in my opinion, one of the simplest ways is to support your analysts is to develop visualisation standards. A playbook in terms of how your team ‘does analytics’.
More than a....
No, don’t worry I am not about to burst into a high-pitched rendition of The Bee Gees ‘more than a woman’. Instead I will share my strong belief that visualisation standards should be more than just a template or set of statements relating to colour a font.
Instead, try centring your standards around your users; a wise company may even say visualisation standards should ‘help people to see and understand their data’. It would be so easy to create standards from a technical viewpoint, of course technical excellence is important element of achieving a consistent approach, however, in my opinion, there are far more important elements to consider. By seeking to simplify your users interactions with a Dashboard you will increase their engagement, their insights and forever become their friend.
Let us first consider how a user interacts with a Dashboard; through sight. Andy Cotgreave has shared with us how the eye is a muscle and just like any muscle it can be trained through repetition. By consistently approaching basic elements we can achieve a simplified, and positive, dashboard experience for users.
Positioning of filters and interactions
Almost every Dashboard you design will include some form of filtering, parameterisation or legend. If this is common across Dashboards then we should consider placing these elements together and in a consistent location on screen; consider this as a type of ‘helper bar’.
By placing all filters, parameters, highlighters and legends within a different coloured section you instantly allow your users to associate the background colour of this section with where to find interactive elements of a Dashboard. In addition, the placement of the ‘helper bar’ is also important. Eye tracking studies have demonstrated people naturally work from top left to bottom right when reading a Dashboard and for this reason you may want to consider placing your filters on the right of a landscape Dashboard, ensuring the first thing they interact with is data as opposed to options. For portrait Dashboards consider placing your filters beneath the header to ensure your users are aware of their existence, i.e. do not make them scroll to find the filters! That said, as demonstrated by a recent Twitter conversation, I do accept the placement can be a personal choice, so my one underlying message is wherever you place your 'helper bar' ensure you please it consistently across all of your Dashboards.
One of way to empower a user to explore a Dashboard is through the use of actions; whether that be filter, highlight or URL. Actions can help reduce the numbers of filters on a Dashboard and increase insight, however, they are only useful if a user knows they are there! How many times have you found an action by accident? Clicking on a chart only to find it updates another part of the Dashboard?
Consistently identifying the presence of actions through different coloured arrows, within a chart title, ensures a user instantly recognises this functionality. To an analyst a coloured arrow may seems simplistic, maybe even trivial, however to an end user it can be one of the most important aspects of their experience when using a dashboard.
Embed Dashboard information
How many times have you been asked ‘Where does the data come from?’ or ‘What is included in this Dashboard?’ People want to be confident in using Dashboard, they want to understand the basic construct and including an information sheet on every Dashboard view can help achieve this.
You can embed Dashboard information within a tooltip to avoid the need to take up valuable Dashboard real estate. Create an empty sheet and add a dummy dimension, for example ‘Information’. Hide the header and then change the marks card to a shape and use a custom shape such as a circle with an i within it. It is then possible to edit the tooltip of the sheet to share information with your users. Below is some suggested content:
- Dashboard title
- Version number
- Dashboard description
- Data sources
- Data last updated (if present in your data set)
- Business owner
- Analyst creator
There are many other aspects of Visualisation Standards that can help support a positive and insightful user experience. Here are some of the aspects we recently adopted as part of our JLL Business Intelligence Centre of Excellence Visualisation Standards:
- Header and Footer placement and inclusion of branding (i.e. logo)
- Mobile / device specific layout
- BANs (Big Ass Numbers)
- Use of colour, from an organisational style and accessibility perspective
- Number formatting
- Axis formatting
And now how visualisation standards can help support your analysts
If are you to adopt the Visualisation Standards as your way of working then it is fair to say the document will be referenced a lot by your analysts; you can use your standards to educate and support your analysts. Go beyond the what and where; help them understand the why.
When writing your standards try to include community content such as blog posts, white papers or webinars that can help explain a technique. For example, an important aspect of designing a successful Dashboard is understanding that less is more and white space is good space. Padding and containers can be used to help achieve this but, in all honesty, how many of us have used padding or containers on a regular basis? Do not just assume your analysts know every aspect of Tableau, support them by sharing knowledge and empowering them to learn through referencing community inspiration on every page of your standards.
A Dashboard can be ruined by the inclusion of just a single, poorly thought through chart. There are so many examples of chart libraries that can help educate and inspire your analysts to choose the right chart type for the right question. It is at this point I have to reference Andy Kriebel and say thank you. In 2018 Andy replicated the Financial Times Viz Vocabulary in Tableau and made this available for us all on Tableau Public. I can only imagine the amount of hours that Andy spent creating this and yet he openly shared it with the community, enabling it to become a learning resource as opposed to just a showcase of Tableau functionality.
As with any instance of using other’s work, if you do decide to reference Andy’s workbook within your visualisation standards, I implore you to ask Andy’s permission before doing so and to acknowledge his contribution within your Visualisation Standards. *
* Chart library small print : Other chart libraries do exist and are superb however the reference to Andy’s Viz Vocabulary is absolutely a personal endorsement, his work is absolutely amazing and should be cherished by everyone within the data visualisation community.
Promoting participation in community initiatives
If visualisations standards are genuinely to be viewed as more than just fonts and colours then why not use them as a way to emphasise and embed culture? If you are committed to supporting your analysts to continuously learn (if not, why not?!) then take the opportunity to include references in your visualisation standards to community initiatives that can help your analysts grow their visualisation skills and knowledge.
There are many such community initiatives; the three I would suggest to include in your visualisation standards, due to them all being directly attributable to storytelling, maximising insights and advanced use of Tableau are:
- Makeover Monday
- Workout Wednesday
- Storytelling With Data
Your analysts do not need to do all three, however by allowing your analysts to participate in just one of these initiatives on a regular basis you will quickly reap the benefits through better quality Dashboards.
Template Dashboards and custom Preference File
Let’s face it, analysts want to viz! They want to explore data and create great dashboards, they don’t want to be spending time applying headers and footers and formatting text. In this respect we should be striving to make it easy for analysts to adopt visualisation standards. There are a couple of easy ways to achieve this.
Firstly, you can create a template Dashboard, using Superstores data that already has the header, footer, corporate logo and consistent elements such as the ‘helper bar’ applied. Create example Superstore charts that are used as placeholders and then share this with your analysts as the starting point of any new Dashboard. By doing so you will save your analysts time and also create consistency in terms of the initial development steps.
In addition, you can also provide analysts with a customised Tableau Preference File that includes all of the hex values for your corporate colour palettes. Analysts can then load this into Tableau Desktop and instantly have access to your approved list of colours to be used when applying colour to measure values.
My last piece of advice is around how to embed visualisation standards and ensuring they are adhered to. By scheduling weekly or bi-weekly, team wide, viz review sessions you allow analysts to share their work, ahead of publication, and benefit from peer support and feedback. This approach also has the added benefit of ensuring the team are aware of each other’s work, helping to breakdown silos and enable Dashboards to be supported across all analysts post publication.
Ask analysts to submit their viz for review, in advance of the session, allowing others in the team to explore it and document feedback. During the viz review session the team can explore the feedback, discussing each element and agreeing potential improvements. Team collaboration is a great way to facilitate sustained excellence and to help create a sense of shared pride in everyone’s work.
If this post has resonated with you and you would like to be inspired by examples of visualisation standards then I have good news for you. The BBC published a fantastic example in 2018 and, my own organisation JLL, will soon be publishing our approach on Tableau Public. Watch this space!