# Simple & Pretty – full content > Simple & Pretty is an independent product design studio led by Matt Homer. Product design for startups, established businesses and design teams: research, prototypes and production-ready products, from idea to launch and beyond. ## About Matt Homer Founder of Simple & Pretty, product designer and design leader. I work across AI products, design systems and mobile apps, and have led design in-house and as an independent consultant, taking products from conception to launch and beyond. I'm an interdisciplinary product designer, used to working with many stakeholders: reporting to the board, working closely with engineers and leading projects across departments. What I'm known for: - Creating user-friendly experiences - Leading designers and product teams - Working directly with company leaders Before starting Simple & Pretty in 2013, I designed apps at Tigerspike, where I met [Judit](/judit). --- ## COVID-19 Testing at scale URL: https://simpleandpretty.co/case-studies/nanoscout Summary: AI software platform to accurately and rapidly detect viruses Outcome: A 5-level proof of concept and a two-phase product vision: cloud software for testing labs, then a brandable self-service testing kiosk. Client: Deep Innovations · Role: Design lead · Dates: Oct 2020 – Mar 2023 · Team: Matt Homer & Judit Layana Ilundain What we did: Research, Service design, Product spec & strategy, Proof of concept, UX & UI concepts, Hardware concepts Tools: Sketch, Google Docs, Google Slides, Flowmapp [Image: ViroScout analysis app on a tablet showing virus particles boxed in red in a microscope image, with a SARS-CoV-2 Detected positive result panel] #### Background Shortly after the COVID-19 outbreak, **[Deep Innovations](https://deepinnovations.co.uk/)** started to build and train a vision-based, deep neural network algorithm that could identify viruses by using cloud-connected SEM microscopes in a spit sample. #### Objective Envision a scalable, flexible, TAAS (Testing As A Service), which would facilitate rapid, low-cost virus testing. For use at health centres, transport stations, public venues, anywhere where a negative result is required for access for public safety. #### Solution In order to deliver a solution to market faster, we divided it into 2 phases: **Phase 1:** Cloud-based, software solution, for use medical centers, hospitals, testing labs. Leveraging existing processes and hardware **Phase 2:** A brandable, self-service kiosk for automated, on-location virus testing #### Deliverables POC, Customer journey mapping, Software & Hardware concepts #### Project overview We began by researching common medical testing procedures, technologies and hardware. This included speaking to medical professionals and lab assistants. This allowed us to map out the end-to-end service lifecycle, including both the physical and digital interactions amongst the different actors (testee, test assistant, lab assistant, SEM operator, health consultants, organisation administrators etc) [Image: ViroScout lifecycle map: a service flow diagram connecting test scenarios, status updates and service steps] [Image: Simplified system model showing how the system, client and testee components connect] [Image: ViroScout simplified testing flow in five photographed steps for Phase 1, from spitting into a tube to getting the result, and for Phase 2 at a self-service kiosk] Working closely with our team, we wrote the MVP Product spec and Strategy documentation to clearly define the product solution, before conceptualising the UI. [Image: The ViroScout MVP Strategy and MVP Product Spec documents open in Google Docs] We worked with the developers to define a proof of concept (POC) to rapidly build and explore the core functionality. We defined 5 levels to the POC, starting with recreating our internal tech demo in app, then each level introduces another piece of functionality e.g. capture image on SEM microscope, multi-SEM processing, multi-user, text scanning and tracking. #### Concepts The system was composed of 4 main areas, which area a user can access depends on their user type and access level: [Image: Phase 1 Tracking app: three phone screens for pairing a sample tube to a testee by QR code and tracking it] [Image: Phase 1 Management app: tablet screens with a test data table, trend charts and a map of infection rates across England] [Image: Phase 1 Analysis app on a phone and tablet, detecting virus particles in microscope images in real time] [Image: Phase 1 Diagnosis: three phone screens of the testee's results page, showing test in progress, positive and negative states] [Image: Phase 2 Self-service kiosk concept shown from the front, side and back, with a list of its features] [Image: Phase 2 Brandable kiosk concept in three wraps: British Airways, Manchester United and Sónar] > p+p]:mt-4"> > > "Having worked with Matt and Judit for 3+ years, they are a great team to work with, very productive, and contribute a lot to the quality of the product, delivering best results from proposal to production. Lots of ideas!" *[Faris Dabour](https://www.linkedin.com/in/faris-dabour-898981176) Co-founder & CTO, Deep Innovations* --- ## Designing an award–winning, pool safety system URL: https://simpleandpretty.co/case-studies/poolscout Summary: From POC, through MVP, Beta, to market launch Outcome: Won three industry awards in 2022 (SIA, ESX and CEDIA Expo), after a 160-person beta and market launch. Client: Deep Innovations · Role: Design lead · Dates: May 2019 – Oct 2023 · Team: Matt Homer & Judit Layana Ilundain What we did: Product strategy, UX & UI design, Design system, Beta program, Customer journey, Website Tools: Sketch, Zeplin, Abstract, Jira, Confluence, Google Docs, Google Sheets, Squarespace, Zapier Integration, Flowmapps [Image: PoolScout app on a phone showing the live pool camera view with a person and a toddler detected, in front of a mother and toddler in a pool] *PoolScout app uses real-time AI to keep pools safer* **[Deep Innovations](http://www.deepinnovations.co.uk/)** specialise in AI and real-time video analytics to identify and alert of unsafe scenarios before accidents occur. Its leading product, [**PoolScout**](http://onelink.to/vttcfz), is the only pool safety system that distinguishes between toddlers, other people and pets in real-time, raising timely alerts before accidents happen. The system includes an outdoor IP camera, an alarm unit and a cross-platform mobile app. When we (Matt & Judit) joined the team in 2019, PoolScout was a basic, but functional, proof of concept app. It connected to a camera livestream and had AI detections. Over the next three years we defined, designed and launched the app, from a 160-person beta to market launch, along with the design system, website and launch materials behind it. The result won multiple awards in 2022, including the Security Industry Association’s New Products and Solutions Award for Smart Home Solutions, an ESX Innovation Award, and the grand prize in the Launchpad TechStarter program at CEDIA Expo. [Image: Before and after: the 2019 proof-of-concept app with three call buttons, next to the current app in alert state for an unattended toddler] *PoolScout prototype proof of concept (2019) vs current app* We were brought onboard as the first designers in the company, reporting directly to the CEO. Together with the product team we defined, designed, built and launched PoolScout, and helped the CEO build and lead the internal product team that took it forward. **Our responsibilities:** - Gather UX requirements for PoolScout Beta and Live v1 (hardware & software) - Write Product Spec doc - Define and design the app UX - Create production-ready UI designs and assets (led by Matt) - Create and maintain the design system for PoolScout multi-platform, multi-product offering (led by Matt) - Work in production with developers to implement and test designs - Gather beta user feedback to improve detections and alerts, app features and general user experience (led by Judit) - Map and design PoolScout's end-to-end customer journey, from product discoverability, to purchase, onboarding, support and post-purchase experience - Provide product and content strategy across the customer journey - Design PoolScout marketing assets: Branding, Website, Social media, video production, flyers, banners, business cards and product packaging (led by Matt) [Image: Product spec document open at the User requirements section, listing primary goals, secondary goals and user needs] *The product spec, focusing on User Requirements section* [Image: Five hand-drawn sheets titled PoolScout Ideas, sketching monitoring screens and a high-severity alert scenario] *Early PoolScout concept sketches* #### The PoolScout app **The pool view** provides users with constant reassurance that the system is active and working. The Live Stream displays AI detections’ boundary boxes and visualisations overlaid on the stream displaying the view of the pool area. When it's a safe detection, a green floating bar displays a live counter. [Image: Three app screens: the live stream in safe state, the alert state for an unattended toddler with emergency actions and contacts, and the Timeline view] *PoolScout app, showing livestream view in safe and alert state, and Timeline view* When an unsafe scenario is detected, users are alerted and the app changes to alert state, playing an in-app alarm sound and displaying a red icon in the header and the alert type in the red banner. If the camera alarm is enabled, it also triggers a siren that goes on until a member of the PoolScout account marks the situation safe in the app. The contextual actions change to display the most relevant shortcuts in the case of an emergency, as well as the user’s predetermined contacts, so they can reach those people when it matters most. Below is an example of how the system responds to a high-severity scenario (Unattended Toddler). [Image: Flow diagram of how the system and users respond to an unattended child alert, from notifications and SMS to marking the event safe] *Mapping the experience of a high-severity event* The **Timeline view,** displays a chronological list of events & updates, grouped by day. This lets users browse back and replay past incidents and events in a period of time, which also helps pool owners to learn from previous incidents. [Image: Close-up of the Timeline view on a phone, listing the day's video events and updates such as Marked as False Alert by Sam and Police contacted] *The Timeline contains the last 7 days of video events and updates (user, settings, system)* #### The design system Living in Zeplin, it contains all design tokens, components, views & flows for PoolScout and all other Deep Innovation's software products. It creates a unified language and visual consistency across products and channels and provides the ability to replicate designs quickly by utilizing pre-made UI components and elements. [Image: Zeplin design system page for the Bottom Sheet component, showing the Pool Alert ambulance variant and its properties] *The design system in Zeplin, focusing on Bottom Sheet Component* #### Production To quickly build the initial version of the app, we used an external development agency. We led the production, working directly with the front & back-end developers, providing designs (view and components) in Zeplin with accompanying UX documentation and guidance. To allow us to efficiently, scale production and keep more knowledge in house, we worked with the CEO to build and lead the internal Product team. To work efficiently as a design team of two, we created internal processes, a D.A.M, automations, document templates and a design system. As this was a very complex experience to build, we created many breakdowns docs to improve design-dev communication and ease implementation effort. [Image: Spreadsheet breaking down each alert type by trigger, conditions, notification content and user actions, colour-coded by severity] *An experience breakdown doc, created to collaborate with dev* #### The Beta program During the Beta program (August 2021 to June 2022), the latest PoolScout hardware and software was available for free to 160 people for testing. The goal was to gather early and often feedback to unveil any critical issues (with the hardware, the technology or the app UX) so we could improve PoolScout as much as possible before market release. We created an automated signup process on the website, that allowed users to apply for beta, and have the hardware shipped to them and an installer to install it for them. [Image: The Apply for Beta web page headed Help Test the Future of Pool Safety!, next to the in-app PoolScout Beta settings page] *The Apply for Beta webpage and the in-app beta page* Feedback surveys were sent to participants after each new release and 1-2-1 follow-ups with testers took place to dive deeper into certain aspects of their feedback. [Image: Beta feedback dashboard with pie charts of feedback by component and by sentiment, above a list of logged feedback items] *Dashboard for tracking the beta testing and actioned insights* As a result of the feedback provided by the beta participants, the product team improved significantly the livestream performance and the Timeline experience. Detections and safety alerts were also improved with the introduction of the 'activity zone' feature, which was requested by several participants during the beta program. #### The customer journey We mapped the end-to-end customer journey to identify the goals and needs of the customer in the different stages of the journey and its touch points – from the moment they discover PoolScout to buying the camera and becoming loyal subscribers – It also represents the relationships between different user types and actors involved in the decision of purchasing PoolScout and becoming a customer (pool owners, PoolScout partners, pool professionals and other influencers). [Image: Two customer journey diagrams with connected stages, from discovering PoolScout through purchase, installation and onboarding] *Mapping the end-to-end customer journey* #### The website We designed and built **[PoolScout website](https://www.poolscout.co/)** in SquareSpace, that is core to PoolScout's customer journey and its touch points. The website is the place where people can: - Find out about PoolScout, its unique value proposition and how it works - Purchase the product and get it delivered home - Learn how to setup the system - Get assistance with the hardware installation - Download the app - Get assistance during onboarding and beyond - Manage PoolScout accounts and subscriptions (2023 release) #### Support We also worked with the Support team to provide a better in-app experience. This is including integrating Zendesk for Support forms and for specific events that may require their intervention to resolve, hopefully before the user knows they have an issue. We also defined and prototyped a secure, multi-tiered Support web app with tools, real-time system alerts and database access. [Image: PSR Admin support web app on a tablet, listing cameras with their status alongside details of a selected camera] *PSR Admin - designed for the PoolScout Support team* #### Launching to market During the run up to launching, we designed assets for the rest of the business. We made decks, animations, videos, app store assets, product packaging, flyers, backdrop banners, business cards, caps and polo shirts [Image: Three PoolScout team members smiling at the PoolScout trade show stand in Las Vegas] *Daniel (VP Sales & Marketing), Gabriela (Partner Success Manager) & Sam (CEO) in Las Vegas, after winning their 2nd Award 🏆* #### Testimonials > p+p]:mt-4"> > > "Matt is a hugely experienced and passionate Product Designer that had a massive impact on bringing to life the product. He's a great asset to the team and continues to play a crucial role in the ongoing product development" *[Sam Weitzman](https://www.linkedin.com/in/sam-w-8923396) CEO & Founder, Deep Innovations* > p+p]:mt-4"> > > "Judit took a new concept and expertly made a user journey that was clear and easy to follow for both users and engineers. She also ran our beta testing, which helped position us for the successful launch we're having. Judit will be a strong addition to any UX team" *[Uriel Weisz](https://www.linkedin.com/in/uriel-weisz-bb83606) *Former *Head of Support, Deep Innovations* > p+p]:mt-4"> > > "As a developer who works a lot with Judit. One is able to see her talent both in the ability to understand the technological flow and interpret said flow an amazing UI/UX > > Additionally, the ability to do QA and see that things were indeed done according to the plan: the keen eye and understanding to think about edge cases and to understand the faults that were found during searching and completing QA testing. This alone bring enormous value to me as a developer and to the company as a whole." *[Israel Mastboim](https://www.linkedin.com/in/israel-mastboim) Full Stack Developer, Deep Innovations* --- ## Modernising financial regtech software URL: https://simpleandpretty.co/case-studies/imtf Summary: Designing a flexible, multi-product, multi-brand design system Outcome: A multi-product, multi-brand design system with child themes, delivered and documented in three months. Client: IMTF · Role: Product designer · Dates: Aug 2019 – Oct 2019 · Team: Matt Homer & Judit Layana Ilundain What we did: Design system, UI audit, Theming & multi-brand, App concepts, Documentation Tools: Sketch, Abstract, Zeroheight, Google Docs, Google Drive [Image: IMTF file manager interface showing recent files such as passports and bank statements above a list of all files] #### Background **[IMTF](https://imtf.com/)** is a global software leader in the regtech space. They develop a large suite of different solutions that range from anti-money laundering (transaction monitoring) to KYC, customer due diligence, transaction screening and onboarding solutions and are with over 1700 clients one of the key players globally. #### Objective To take advantage of synergies and reduce inefficiencies, they were in the process of modernising and standardising their codebases. To increase the speed of rebuilding their products and to increase design consistency, IMFT wanted to build a single, multi-product, multi-brand design system. It would provide a consistent, familiar style, but allowing for stylistic differentiation between the products, to better match their individual, specific brand and experience requirements. #### Deliverables - Design system - Design system documentation - ICOS3 app concepts #### Solution We started by meeting the design and development teams to align on the priorities of their redesign and how the design system would be utilised. Google Material was chosen as a starting point as it was customisable, scalable and the team had previous experience using it, which would give both design and development a head start on the UI. We then presented the approach to leadership for sign-off. Inside Sketch we stylistically modified all of Google’s Material library to create IMTF base theme. To allow the internal designers to be able to design faster and make the system more useable, we added smart layout, symbol overrides and nesting to all the components. With the help of IMTF internal design team, we conducted a UI audit of all their products. The goal was to identify all product-specific styles and common components and patterns. Design components such as lists, cards, side sheets, tables and navigation bars were then added to IMTF base design library. Also common objects such as 'Documents', 'Cases', 'Tasks', 'People' and 'Companies' were identified and differentiated visually. [Image: The IMTF base library in Sketch, open on the Tables page with dozens of table cell and row components] With the base library completed, the internal team was ready to create child libraries with their own themes for each of the products in the family, importing the base components with their custom tokens (colours, typography, icons and styles) applied to them. We also added product specific components and variations, which had been identified in the UI audit. [Image: Card types from the design system: meta lists, comment lists, check lists, timeline lists, forms and image top cards] Using the child library for ICOS3, IMTF flagship product, we were able to quickly mock-up the app, so we could visualise the design system in action. Seeing how the components worked together in different contexts allowed us to make some important changes at the component level. Additionally we made the library customisable to allow it to better match client brands. [Image: ICOS3 Cases screen listing 100 cases with status, type and owner, and an onboarding case open in a side sheet] [Image: ICOS3 Case Manager re-themed in red, listing KYC review and onboarding cases] [Image: ICOS3 Case Manager in a light blue client theme, with a case's comments open in a side panel] We used Zeroheight to document the design system’s libraries, components and style tokens, before handing over to the internal design team to continue. [Image: IMTF ICOS3 style guide in Zeroheight, open on the Cards component page] > p+p]:mt-4"> > > "Great experience to work with Matt & Judit, great design input and ideas, on-time and professional collaboration and pro-actively driving things forward!" *[Gion-Andri Büsser](https://www.linkedin.com/in/gion-andri-buesser) CEO, IMTF* > p+p]:mt-4"> > > "We hired Simple & Pretty (Matt & Judit) to help us to build our design system at IMTF. With their help, we managed to launch our first version of the design system in six months (while starting from scratch). It has been a great experience working with Matt & Judit, very responsive and accommodating with an infectious positive energy and can-do attitude." *[Niko Antonogiannis](https://www.linkedin.com/in/itsniko/) *Former *Head of Product Design, IMTF* --- ## Growing Instaread beyond book summaries URL: https://simpleandpretty.co/case-studies/instaread Summary: From app conversion to an audio player on publishers' websites Outcome: The embeddable audio player went live on sites including Britannica, Dr Axe and Entrepreneur's Handbook. Client: Instaread · Role: Product designer · Dates: Feb 2019 – Nov 2019 · Team: Matt Homer & Judit Layana Ilundain What we did: Product strategy, UX & UI design, Conversion, Prototyping, Audio player, Website Tools: Sketch, Zeplin, Google Docs, InVision, Userzoom [Image: Three Instaread iPhone screens on a blue background: the Big Debt Crises book preview with an Unlock this book button, the Try Instaread totally free for 7 days pop-up and the Time's up membership pop-up with monthly and yearly prices] #### Background Instaread is a learning app that gives people the key insights from 800+ best-selling nonfiction books, in text and audio. In 2019 it was growing beyond book summaries, adding Instaread Originals and Articles from trusted sources like The New York Times Magazine. #### Objective Instaread's first problem was conversion. For every 100 people who installed the iOS app, 25 opened a book preview and only 2.5 started a free trial. They wanted to double conversion from Book Preview to trial, from 10% to 20%, and find new ways to reach people outside the app. [Image: Line chart of the Instaread user journey from download to remaining subscribers, comparing the current user drop-off with the estimated drop-off of the proposed solution, against the current 2.5% and challenge 5% conversion rates] *User drop-off in the existing experience, and our proposed solution* #### Deliverables - Design challenge: Book Preview and trial redesign, with an iPhone prototype - iOS Book Preview, CTA testing options and onboarding - Embeddable audio player for publishers (MVP and iterations) - Articles landing page concept - Website homepage and All Categories updates - Article narration landing page, widget and profiles - Brandable playlist widget - Ad-supported in-article widget and global audio player #### It started with a design challenge Instaread invited us (Matt & Judit) to take on a design challenge: redesign the Book Preview and trial pop-up so that 20% of people previewing a book became trial subscribers. We audited the existing flow, mapped where users dropped off, and looked at why people read summaries and what makes them pick a book. Two assumptions came out of it. The Book Preview was hiding the content that helps people decide, and the opt-out free trial (card details first, cancel before day 7) was putting people off. The details backed this up. A carousel hid the summary and contents behind the cover. CTAs were small and easy to miss. Users didn't know a book was locked until they tapped it, and the Apple Pay sheet made a free trial look like a payment. **Our solution:** - A single scrolling Book Preview: cover, useful metadata, summary and contents, with a large CTA that makes it clear the book is locked - Reviews, other books by the same author and similar Instareads, to keep people exploring - An opt-in trial, separate from the subscription, so people experience the product before being asked to pay - A post-trial pop-up that repeats the value proposition, with clearly labelled pricing We delivered it as an InVision prototype for iPhone, along with how we would validate it with user research and A/B testing in a real project. A month later we were designing it for production. [Image: Two iPhone screens: the Try Instaread totally free for 7 days pop-up listing Get Smarter, Find Inspiration and On the Go, and the Time's up membership pop-up with monthly and yearly subscription prices] *The opt-in trial pop-up, and the post-trial pop-up shown when the trial ends* #### Part of the product team From February to November 2019 we were Instaread's design team, working remotely with the product lead and developers on a run of focused projects. Each one started with a short project doc covering background, business and user goals, research, what the design needs (and doesn't), and the design intent. Short video walkthroughs kept everyone up to date between calls. **Book Preview, for real.** We took the challenge into production, working within constraints we learnt from interviewing the iOS developer: no ratings or rankings yet, and one visual treatment for books, Originals and Articles. We researched what makes CTAs effective and gave the team button labels and microcopy to test for each scenario, such as "Start my free trial. 7 days. Cancel anytime." We then reviewed the build against the designs in Zeplin. [Image: Three iPhone screens: the Big Debt Crises book preview with likes, reviews and listening and reading times, the free trial pop-up and the post-trial membership pop-up] *The Book Preview as a single scrolling view: cover and CTA, summary and contents, then reviews and similar Instareads* **App onboarding, Website and Discover.** We redesigned iOS onboarding with new illustrations, updated the homepage to reflect the wider offer, and designed an All Categories page, with an illustration per category, to shorten the long list of carousels on Discover. #### Taking Instaread beyond the app The biggest new piece of work was an audio player that publishers could embed in their own articles. Instaread converted their written articles to audio, and the player put Instaread in front of new audiences on other people's websites. We benchmarked 13 embeddable players, including Spotify, SoundCloud, Simplecast and Apple's, then stripped it back to what short spoken-word audio needs: play and pause, progress and scrubbing, speed and share. No waveform, no chapters, no repeated article title, because the player always lives inside the article it reads. It kept a fixed height and scaled down to 300px wide without changing layout. When the team asked to add the article title and source name to drive more clicks, we showed what that would do on mobile and proposed alternatives, including a paused state that promotes Instaread without hiding the controls. We also tested the build across Android, iOS and desktop and fed back every issue to the developers. The player went live on sites including Britannica, Dr Axe, Entrepreneur's Handbook, The Emotion Machine and The Quintessential Mind. Outside of the app, we designed: - An Articles landing page for listeners clicking through, planned in stages from a static beta page to one personalised by the referring source, author and topic ([video of the Articles landing page concept on Behance](https://www.behance.net/gallery/258252265/Growing-Instaread-beyond-book-summaries)) - An embeddable audio widget for narrated 3rd-party content [Image: Hand-drawn sketches dated 20 Feb 2019, titled Instaread concepts: embeddable audio widget (3rd party content), marking which parts of each player layout are fixed and which are variable] *Early sketches: what stays fixed and what flexes as the player scales* [Image: Widget concept sheet titled Narrator citation, showing the player in its initial, loading, playing, paused, share, speed and played states] *Widget states with narrator citation* [Image: The audio widget inside a Sam Harris article on desktop, tablet and phone] *The widget in an article on desktop, tablet and phone* **Narrator profiles.** To spotlight narrators and encourage them to narrate more articles. [Image: Narrator profile page for Jaysone Nawalla on desktop and mobile, with articles narrated, listens and followers, and articles narrated for Encyclopaedia Britannica and Dr. Axe] *Narrator profiles, linked from the player. Optimised for desktop and mobile.* **The Playlist widget.** For continuous, hands-free playback and quickly accessing additional content from the same source. [Image: The playlist widget in five publisher colour themes, shown in its initial state and while playing an article from the list] *The widget is brandable to better match the 3rd-party website it's embedded into* #### 2023: an ad-supported audio player In 2023 Instaread brought S&P back. With AI voices now able to narrate articles, they wanted to relaunch the player with an ad-based model: banner ads first, then audio ads once listening numbers were high enough. Matt designed two connected components: - Updated in-article widget: listen now or add to playlist, always in sync with the global player - Global player: minimises to a floating button with an animated waveform while playing, and can be pinned left or right so it doesn't clash with other floating elements. It expands into a bottom sheet with Now Playing, Playlist, Recommended and Popular tabs Both are responsive, support three banner ad sizes, and can be themed with the publisher's own colours, falling back to an Instaread ad when no advert is available. The player never shows an empty state: pressing Play loads the next unplayed article in the playlist, or the most popular one. The playlist works without logging in and is only remembered between visits if the visitor accepts cookies. Everything was delivered dev-ready in Zeplin, with a spec covering every component, state and flow. [Image: The global audio player expanded with its playlist on the Dr. Axe website on desktop, minimised on an Earth911 article on a phone, and expanded on the Dr. Axe mobile site] *The global player expanded on Dr Axe, minimised on Earth911, and expanded on mobile* --- ## Redesigning Mullvad VPN URL: https://simpleandpretty.co/case-studies/mullvad Summary: The safest, easiest, most reliable way to use the internet Outcome: PC World: “Mullvad remains our top recommendation for a VPN service.” Client: Mullvad · Role: Product designer · Dates: Nov 2016 – Feb 2018 · Team: Matt Homer & Judit Layana Ilundain What we did: User research, Product strategy, UX & UI design, Design system, Usability testing, Beta releases Tools: Sketch, Zeplin, Google Sheets [Image: The redesigned Mullvad VPN app open from the macOS menu bar over a dark map, showing a secure connection to New York, USA] #### Project background [Mullvad’s](http://www.mullvad.net/) mission is to make all internet censorship and surveillance ineffective, by helping all individuals to be secure online using their VPN service. Their VPN was ranked among the top 5 most private and secure VPNs in the market worldwide, but was technically complex to use and mainly targeted to high-proficient tech users. When it came to user experience and offering, Mullvad had a competitive disadvantage against sleek VPN clients with less security standards, but with easy-to-use interfaces across platforms, such as [VPN Express](http://www.expressvpn.com/) and [Encrypt](https://encrypt.me/). We came into the project to help them delivering a user-friendly experience within the required security standards. #### Mullvad's existing VPN clients Window left & macOS right [Image: The old Mullvad Windows client: a grey settings window with account, ports and country fields and checkboxes] [Image: The old Mullvad macOS client: a grey status window listing account time left, connection status, server and version] #### Project objectives Designing a cross-platform (Mac, Windows, iOS, Android and Linux) app, which combined Mullvad’s high security standards with a more intuitive and engaging experience. #### Business goals With the new VPN client, the project team was aiming to: - Increase conversion after free trial - Increased active paying users - Decreased customer support requests vs active users #### Challenges - Most people don’t use or know about VPNs - App would not be native and needed to share a common interface across platform - No existing user data or analytics - No UX team, culture or process in existing team - Preconceived solution requirements In order to help Mullvad, we had to develop a better understanding of the company, its team, technology, future, competition, existing and target users… And much more about internet surveillance, censorship, security and VPN usage. [Image: Judit placing sticky notes on a wall-sized customer journey map] *Judit mapping out Mullvad ’s customer journey.* #### How we learnt about the problem We started by conducting **stakeholders’ interviews**, followed by some project team workshops, while doing market and customer desktop research. We interviewed Mullvad’s **customer support team** to identify the needs and problems existing customers had with the service. We **looked into their competitors** to identify Mullvad's strengths and competitive advantages, as well as weaknesses and room for improvement. As Mullvad’s has a strict ‘No user tracking or logging’ policy, to get a better understanding of their **target audience**, we interviewed many internet users: some existing Mullvad customers; other VPN users and non-VPN users. To cast the net wider, we spent a lot of time in VPN and security online forums on platforms such as **[Reddit](http://reddit.com/)** to gather users impressions and opinions. [Image: Wall of sticky notes grouping customer reviews into a pattern analysis] [Image: Printed benchmark analysis grid comparing VPN services on privacy, technical and business criteria] #### Research insights Internet security is complex and deep. While there was an increasing public awareness, most understanding was vague and superficial. In fact over **70% of internet users either didn’t know what a VPN is or had never tried using one**. Whereas, many of Mullvad’s existing users seemed very tech proficient, with specific security requirements fulfilled by Mullvad’s existing app offering. By **making being safe online simple enough**, Mullvad had a bigger **market opportunity** to reach, not only VPN existing users, but also break into the 70% of non users… At the same time allowing most experienced users with specific security requirements have access to the functionality they needed. [Image: Three handwritten persona sheets: The content streamer, The privacy advocate and The censored] *Mullvad ’s preliminary personas, based on user types and motivations to use a VPN.* Reaching the 70% would not be easy though. The apps would have to do a better job **bridging the gap** most security products have with normal internet users. The experience should take inspiration from non-security apps and appear where they do — Not be trapped on the Mullvad website and security forums. The experience would also have to **feel fast and easy**, so the user doesn’t feel confused or intimidated, and doesn’t get in the way of what they were doing. #### How we helped defining the solution. We presented our insights and recommendations from the discovery research followed by a Q&A to discuss impressions and clarify what we’ve learned. We do this, so we can **align with the team**, and everyone understands why design decisions are being made, which reduces subjective or biased discussions later on. [Image: Sticky notes on a wall grouped under solution requirements, needs and use cases] *No-logs privacy policy, speed, ease of use, control, automation and scalability were key solution requirements for Mullvad .* #### Solution requirements - Simple, fast, intuitive interface, that can work on all platforms, with minimal platform specific variances. - Focus only on primary info and functionality in UI, hide complexity. - Encourages ‘Set and forget’ usage. - Show don’t tell — use colours, animations, interactions and transitions to clearly communicate to the user whether they are safe or not, and whether they need to do something. - Inclusive — should be usable by users with common access needs. - Available in OS’s official app stores [Image: The New Mullvad VPN Client product requirements document open in Google Docs] *Mullvad VPN app Product specifications document. In Google Docs for team collaboration and feedback.* #### A collaborative approach As the Mullvad team contains many technology and security experts, we felt it was important to define the product UX requirements very collaboratively. We did this by running a workshop with the team to prioritise app functionality and creating the product specification document in **[Google Docs](https://www.google.com/docs/about/)**, so we could then iterate collaboratively as a team. To prevent chaos ensuing, we used commenting and the product lead moderated changes to the doc. #### An iterative strategy As often the case, the product spec document grew long. To prevent feature creep and development taking forever, we encouraged a single-platform MVP first. We could beta test and iterate on it before launch, release on the other platforms and afterwards add more functionality. Due to team capabilities and technical constraints, it was decided to go MacOS first. We helped the team to break up the product spec into milestones and then populated a Trello board with the prioritised tasks, into sprints. [Image: Design plan table for the Mullvad VPN client MVP, with research, design and production phases across learn, build and test loops] *Design plan for Mullvad ’s MVP cross-platform app.* #### How we created the user experience. We first started mapping and breaking down the experience based on functionality. We identified key user flows and optimised them to **the least steps and user interaction** possible. [Image: Flow diagram for first time connect: login, not connected, server selection, connecting and connected] [Image: Flow diagram for switching server: connected, server selection, connecting, connected or not connected] Through many sketches, mockups and prototypes, we repeated ‘design > test > learn’ iterative loops to refine the experience, so it was **simple and intuitive** enough for most to use, but at the same time containing functionality that more advanced users required. We created the interface and its interactions with the goal of providing users with **reassurance**, **ease and speed of use**, **convenience**, **clarity** and **control**. ##### Reassurance Using Mullvad’s strong visual brand, we made the UI inside and out very clear to whether the user was connected or not. With **just glancing** the app or menubar icon, users could see if they were safe (green, locked padlock) or not (red, unlocked padlock). With a big button to disconnect (red) and green (connect). [Image: Hand-drawn state diagram of the connection animation between not connected, connecting, switching and connected frames] [Image: Paper sketches of the not safe and safe app states, with red and green headers and connect buttons] [Image: The not connected, connecting and connected app states, each with its red or green padlock icon] ##### Ease and speed Beyond connecting and disconnecting, the next most likely thing a user would want to do is **switch location**. The standard VPN experience is to disconnect, select server, reconnect, which is bad for both security and the user experience. So, we designed it so users could select and connect to a different location **in two taps**, keeping the user secure and connected throughout. ##### Clarity To show and teach the user what was happening and what state their connection was in, we used **large typography** for key information like location. We also used a **simplified map** in the background to show the user where the internet thought they were. ##### Control We wanted for users, regardless their technical proficiency, to understand what is happening when using their VPN so they can be in control of their actions. Labels, instructional text and contextual help on functionality were written in **plain language** across the interface, with a particular focus on Settings, so features could be better understood by the user. ##### Convenience We wanted the app to run on ‘**auto-pilot**’, so the standard user could just install it and forget about. So we prioritised launch application on-startup, auto-login, auto-connect and auto-reconnect, so the user was **safe by default**, without doing anything, with option to change if they preferred. #### How we executed the user interface. To manage the migration from the existing VPN client to the new, the back-end development team needed to update its tech-stack, while maintaining the one that was powering the existing apps. This meant front-end development could have a head start, working along side us, which allowed us to collaboratively build the front-end, test it with users and iterate before plugging in back-end functionality. [Image: The Mullvad VPN prototype in InVision, showing the first time login screens in sequence] *Tappable Mullvad VPN app prototype in InVision.* #### Dev collaboration We created the design system and screens in **[Sketch](http://sketchapp.com/)**, design specs in **[Zeplin](https://zeplin.io/)**, prototypes in **[Invision](https://www.invisionapp.com/)** and UI demos in **[Codepen](https://codepen.io/)**, so we could communicate efficiently with the front-end developers and execute the interface the way it was meant to look and work. We coordinated our workflows, so we could build and test different prototypes quickly, with daily builds and catch-ups, and ongoing communication on **[Slack](http://slack.com/)**. [Image: The Mullvad VPN style guide in Zeplin, with the colour palette and text styles] *Mullvad VPN app Design system shared in Zeplin.* #### User research & testing All user research was **conducted remotely**, running one-to-one interviews with testing via screen sharing and weekly diary studies with follow-up interviews. We recruited participants to gather qualitative user feedback through Mullvad’s website and their networks. To gather **fast quantitive user insights**, we used **[Usabilityhub](https://usabilityhub.com/)** to conduct click tests, nav tests, design surveys and preference tests to make sure all design decisions were informed with evidence. **Research findings** were shared and discuss with the team via **[Slack](http://slack.com/)** and **design actions** were agreed during stand-up meetings or specific team calls. These then were translated into tasks on the project **[Trello](http://trello.com/)** board to keep track of ongoing work and production sprints. [Image: First click test heatmap showing where participants clicked to switch server in the app] [Image: Preference test result comparing two settings account designs, with the first preferred by 80% of participants] [Image: Diary study checklist where testers ticked any issues they ran into while using the VPN each day] #### Release We conducted several internal, private and public **beta releases** of the MacOS app before official release to make sure the app was stable and secure. During the public beta, we continued to work closely with the development team, designing and implementing new functionality and extending the experience in multiple, iterative releases. We also worked with their **marketing** team to **produce assets** for their website and social channels. #### What the press said... [A fantastic VPN has a great new look](https://www.pcworld.com/article/3313320/software/mullvad-2018-vpn-review.html) > p+p]:mt-4"> > > “Mullvad has changed where it needed to and everything else has remained the same or become stronger. With its commitment to privacy, anonymity (as close as you can realistically get online), and performance Mullvad remains our top recommendation for a VPN service.” *[Ian Paul](https://www.pcworld.com/author/Ian-Paul) from PC World*