Creating a Tableau Style Guide: How we achieved engaging data visualization at Exness

Introduction

Tableau is the primary source of data insights for both top and middle managers at Exness. In this article, I will share the story of how we developed the Exness Tableau Style Guide, the ultimate tool for standardizing dashboards that helps our Reporting team create high-quality data visualizations with ease.

At Exness, Tableau is the key reporting tool for both top and middle managers. It contains operational reports for top managers with the main company metrics, detailed information on metric behavior in each business domain of the company, and a strategy map with strategic objectives and KPIs. The sales and financial teams have specific sets of reporting tailored to their needs. Other types of operational reporting are usually handled using domain-specific tools or our separate self-service tool — Metabase.

We began using Tableau in 2018 with only four to five BI analysts who were tasked with creating dashboards while also juggling other responsibilities. Today, we have a dedicated Reporting team consisting of 18 visualization specialists and around 200 active reports. To top it off, the newest member of our Reporting team is a Data Designer. Their main responsibility is to make our reports clearer and more useful for our users by selecting compelling visuals, eliminating clutter, and highlighting important information. They also create Tableau assets such as color palettes, icons, pictures, buttons, and more.

Some members of our team had prior experience in creating a Tableau Style Guide. Therefore, at Exness, we built upon this foundation and further improved it by incorporating additional features and procedures. To access our Style Guide, you can follow the link provided at the end of this article, which will take you to our Tableau Public space where you can download it.

The Reasons Behind Our Style Guide

First and foremost, we need StyleGuide to assist our professional BI specialists in creating standardized and recognizable reports without increasing development time. As the company expands, the number of specialists responsible for report creation also increases. Even if they are all highly skilled professionals, they still have their own individual styles, which can lead to a lack of consistent and recognizable reporting standards. The beauty of data visualization lies in the consistency of style because it results in clarity.

“We have neat reports that are formatted similarly, and stakeholders are 100% certain that the reports are developed by the corporate reporting team, which works like an additional quality mark,” says Ksenia Ageeva, Lead BI Analyst at Exness.

Having unified standards can help focus the user’s attention on finding answers and insights instead of searching for necessary filters or understanding the meaning of colors, among other things. However, when developing the Style Guide, we aimed to identify aspects that could simplify report creation and reduce development time without limiting creativity. This is especially important in order to maintain and improve the quality of our reports. We also welcome suggestions from developers for improving the Guide and always strive to find the most interesting and useful visualizations while keeping the Guide up to date. Additionally, we have established a quality and design approval procedure to ensure adherence to the guidelines, which will be discussed in more detail later.

The second reason for having a Style Guide is to reduce time-to-market. With such a tool, your DataViz team can produce high-quality reports much more efficiently. Of course, some adaptation time is necessary, but after a brief period, everything will become automatic. Furthermore, the onboarding process for newcomers can be expedited.

Typically, the need for a Style Guide arises when we have a multitude of reports with varying layouts, color palettes, icons, and so on. Users begin to complain about the differences between reports and inquire about proper usage, descriptions, and the meaning behind icons.

Meanwhile, report developers also start asking similar questions regarding the appropriate layout, filter placement, icon usage, and more. In a worse scenario, they may not ask anything and instead copy-paste the style used in their previous job when creating dashboards.

At Exness, we have faced these issues as well. As the number of developers grew, some users outside the BI division with a Tableau license also began creating their own reports in an inconsistent style. In general, the quality of our reports was declining.

Prerequisites:

• Different developers with varying styles
• A large number of reports in pre-Prod or Pre-Release stages, resulting in technical debt
• Decreased customer trust
• High demand for quality from management.

What is Included in Our Style Guide?

Our Style Guide is a Tableau workbook that conveniently contains all necessary information in one place, making it easy to copy and paste. To determine the necessary topics, we considered the most common questions or issues encountered by Tableau developers.

We start our Style Guide with an Algorithm for Building a Dashboard, outlining the key steps to take when creating a visualization. Then, we provide some basic tips on Composition to improve dashboard clarity and user-friendliness.

The core elements of any guidelines are the implementation of the logo in different projects, typography, colors, and graphic elements like icons/images. If you have a design team, as we do, you can synchronize with them and obtain all necessary information from the company’s brand book, including typography recommendations, logo styles and colors, and more. If you do not have an in-house designer, you can still make some simple design choices and rely on Tableau Desktop’s default settings.

As a report developer, it’s important to know the different types of charts and when to use them. This knowledge helps you make decisions quickly and use common examples effectively. And for Tableau developers, understanding layouts is crucial. This includes how to use containers, sizing, spacing, and filters. These topics are common, and having some rules to follow can save time.

Let’s go through the main topics and why they matter

In the Key Steps section, we outline a series of steps to help you apply best practices. One important point to note is to create a simple sketch before starting work. This can help you understand the dashboard’s logic and identify opportunities and limitations. You can use a canvas to describe requirements or special tools like Figma to prepare the sketch. The main thing is to always prepare a sketch, even if it’s just on paper. This can save time in the long run.

Composition is a critical aspect of creating effective visualizations. While there are many rules to follow, we have selected the five most effective from our perspective. These rules can be applied to any type of visualization.

Layout is an essential component of creating a dashboard, especially for beginners. It’s important to establish a proper structure and size from the outset. We offer practical advice on how to build a dashboard effectively.

We recommend that every dashboard have specific sections, including a header area with a logo, a filter area, and a visualization area with various placements. This approach simplifies decision-making and improves the quality and speed of dashboard creation. Additionally, prepared examples can be downloaded and modified within the established structure, making it easier than starting from scratch. Proper sizing and container usage can ensure the dashboard looks great on different devices.

Spacing is another vital topic when it comes to design. Adequate “air” between elements can enhance the visualization’s transparency and improve the user’s perception. This approach helps to clearly separate semantic blocks and guide the user’s eye movement.

Chart Types is a section dedicated to selecting the most appropriate chart for specific cases. This is particularly useful for beginners.

Data can be represented in various ways, such as continuous numeric or discrete categorical. Each type has its own attributes that are best suited for understanding the data. In our Style Guide, we have examined all the key attributes to make it easier and faster to determine the most suitable chart type for data analysis. Moreover, there are many “chart chooser” tools available on the internet, including Tableau Public.

Typography covers font types, sizes, and colors, including headers, subheaders, tooltips, and actions. It’s crucial to have all these details in the workbook example, providing developers with a ready-made tool that can be easily copied. We use the standard Tableau Book font because it is convenient for dashboards, looks great, and is suitable for different purposes. There’s no need to change it in existing reports.

Colors is a section dedicated to providing tips and tricks on color usage, including various color palettes such as sequential and divergent sets, as well as RAGs.

Color palettes are a highly debated topic. Many users and developers have their own opinions on dashboard colors, while designers have strict rules to follow. It’s important to strike a balance between company branding guidelines and personal preferences. Creating a standard color palette with 6–8 colors and different light and dark versions can help achieve consistency. Additionally, a few RAG colors (typically red, orange, and green) should be included. To simplify installation, all these colors can be saved in Tableau preferences.tps file.

Filters can be a challenging aspect of dashboard design. The Style Guide provides best practices on how to use filters effectively and where to place them. It also includes technical advice on Tableau filters to improve user experience and dashboard performance. Knowing where to place filters can make a significant difference in how users interact with the dashboard.

Examples — Our experience has shown us that creating specific exemplary dashboards is not a sustainable practice. Instead, we recommend creating a “Collection” (in Tableau terms) of the best existing cases. Developers can choose the most suitable one for their use case and duplicate it. This approach is more efficient and saves time.

Pre-release Check List is a critical tool that enables a general check before publication. It contains main topics with links to the full list. By following this checklist, the approval process can be significantly streamlined.

What Processes Did We Implement?

After creating a Style Guide, the next question is how to ensure that all developers follow the rules. To address this, we implemented a process with the following aspects:

• Quality and Design Approval Process — We already had a code review meeting where we reviewed the SQL queries that we planned to put into our Git and use in our DWH as data sources. We decided to have a similar meeting for report design before publishing any report to the production environment. This process includes not only validating whether the Style Guide requirements were met but also a UI/UX test. This test helps identify incomprehensible data and highlight points that are unclear to someone who does not know the report’s details.
• Design Approval Group — We delegated the approval process to four specialists from our team. Once your report is ready, you must post a link in our internal Slack chat and tag @design-approval-group. You should receive at least two approvals before publishing to production. You have two options — you can either post a link in the chat or discuss it during the design review meeting if you have any doubts or questions. The Design Approval Group also serves as an additional step in the career ladder — membership signifies that you have top-notch design skills.

“A Design Approver acts as a stakeholder who is seeing the report for the first time and has little to no idea about the data’s intended meaning. This approach ensures that the real stakeholder finds the report easy to use, understands the figures’ meaning, and comprehends the intentions behind the charts. By doing so, we aim to enable a smooth decision-making process based on the report, which is our ultimate goal.” — Alex Izotov, Lead BI Analyst at Exness.

Visit our Tableau Public page and download Exness Tableau Style Guide

P.S. I would like to express my gratitude to Alex Barakov, my former colleague, who led the development of the Style Guide at Luxoft. Special thanks to Alex Izotov and Ksenia Ageeva for their contribution to driving the Style Guide at Exness. And lastly, I want to acknowledge our brilliant Designer, Kate Starova, who created this Guide and continues to develop it at Exness.