Reporting Suite Redesign
Creating connection and flow between KPIs and reporting at all levels of the organisation
Brief
Engage with stakeholders across the business to understand their key data and insight questions.
Create a reporting structure that supports a ‘choose your own adventure’ style experience for users where they can navigate through an interlinked set of reports to answer their own questions on business performance.
Design a set of reports with reusable modules that provide flexible insight
Findings
Users come to data intelligence with four different missions
Tell me the story of what is happening in the business/in my area, and what is driving it.
How is performance different between two [cohorts/time periods/attributes etc].
Help me find a specific fact.
Let me download the data tables to manipulate myself.
Distinguishing these missions clearly helped the Data Intelligence team design optimised reports and services, rather than the previous situation where reports ballooned with additional feature requests, and ad hoc investigations slid into regular production without consideration of the whole.
The Business Story - Reporting Structure
I developed a KPI tree that takes the overall measure of business success (total revenue) and breaks it down by the component measures, e.g. subscription revenue + transaction revenue + hardware revenue, and so on down the chain of measures.
Organising metrics in this way allowed the business to really understand its drivers. It also allowed me to structure a suite of reports that group metrics together in meaningful ways.
The Business Story - Reporting Designs
The Data Intelligence home page gives an overview of the most important KPI in each area and acts as a navigation hub for users to explore stories further.
Each story & dashboard contains a core set of common dimension filters which allow the user to drill down to their area of interest.
Each area contains its own story, but they all have similar structures.
What is happening?
Is that good or bad?
Why is it happening?
What contextual information do we need?
Mobile friendly customer segmentation dashboard
Brief
The client wanted a mobile friendly Tableau dashboard that provided a standard set of metrics for any customer segment, placing the segment in the context of the overall business performance.
Tools
Pen, paper, and Tableau.
I follow a UX development approach for visualisations of all kinds, which means sketching and testing multiple ideas before building anything in a tool. The first image shows the final set of sketches in a series of iterations and improvements made during one afternoon. Using this approach gave the client a chance to feed back on multiple versions, clarifying the message and thinking each time.
Analysis
The client wanted the template to work equally well whether they were focusing on a marketing segment (e.g. affluent older people), or a data defined segment (e.g. those who research online and then shop in store).
I combined indexed fields, percentage changes, and total sales values to answer the two key questions: how is this segment performing vs. other customers, and how important to the overall business is that change?
Interactive D3 Organisation Structure
Brief
The client has a large and complex data organisation, with teams in multiple markets contributing different roles and products to the overall Big Data programme. They saw a static visualisation based on some IBM research and asked me to create an interactive equivalent to describe their own Big Data set up across multiple products, roles, and markets.
Tools
D3.js, and some help from David Johnstone's wonderful gradient picker for generating a smooth colour palette.
Data
The client wanted non-technical team members to be able to update the data easily, as different markets came on board the programme, so we created a simple CSV data set.
The data used in this public version replicates the original IBM image.
Comic Industry Survey Infographic
Comics laureate Hannah Berry commissioned me to create an infographic to share the results of a survey that looked at the state of the comics industry in the UK.
I also make comics and illustrations. You can see them here.
Interactive Discworld Explorer
Brief
This was a self initiated project that came from my long time love of Pratchett's books.
Tools
Tableau for the data visualisation, and brush and ink, Adobe Illustrator, and Adobe Photoshop for the illustrations.
Data
Gathered from the books themselves and the Discworld Wiki.
Data Visualisation Style Guidelines
Brief
I have created for multiple organisations, a set of data visualisation style guidelines and templates across spreadsheets, Tableau, Power BI, and presentation software that extends corporate branding guidelines in a way that is suitable for presenting data clearly and effectively.
Solutions
The spreadsheet / Power BI / Tableau guidelines generally cover how to best present
Performance headlines
Various kinds of bar charts and line charts
Specialist charts such as waterfalls and Sankeys
Fonts, colours and sizing
Effective labelling
Google slides or Powerpoint guidelines generally cover
How to structure a compelling data presentation
Some options for clear layouts
How and where to place commentary
Colours, fonts, spacing
Author Gender Infographic
Brief
The client wanted to update their visuals created for the annual survey of diversity in literary magazines. Previously, they had been using PowerPoint, they wanted something that was more impactful, but still felt familiar to the audience.
Analysis
The analysis is in three parts:
The articles in each magazine are analysed for the year by gender and type of article.
These values are compared to previous years performance.
For the magazines in the 'Main Count' section of the survey, there is an additional section providing demographic details as recorded through a survey.
Survey data can be complex to analyse, so a significant chunk of the time was spent putting it into the right structure in Excel.
Tools
Excel & Tableau.
Using Tableau parameter functionality was critical in this work, as we needed to create the same analysis for around 50 magazines. Parameters allow users to select the magazine of interest from a drop down list.
