Creating a Design System for Wine-Searcher

Designing a system of components and styles that worked across Wine-Searcher's web and app ecosystem

Supporting images coming soon!

Background

Wine-Searcher has historically functioned with a code-first approach, the product originally built by a team with development backgrounds. While this did have its advantages (speedy reaction time, quickly-implemented projects), after over 2 decades without design oversight, there was no uniformity or cohesion throughout the web ecosystem. There was also almost no crossover between the web and app team, which resulted in 2 separate products with completely different component sets without any common ground.

My Role

I was hired to fix these inconsistencies and create a functional design system so Wine-Searcher could finally have a set of common components and design instruction to not only make the developers' lives easier, but also develop and sense of brand identity and improve user experience.

Challenges & Process

When I was first brought onboard, I learned the app designer had recently left the company due to health reasons, so I was tasked not only with creating this design system, but also filling her shoes. This was in the midst of the COVID-19 uncertainty, so the company wanted to see if we could make it work with just me doing both roles before searching for a permanent replacement (there was also a chance she would return). This turned out to be both beneficial and challenging – this system would have to be built in tandem with the app projects that had been stacking up in the pipeline, but due to this I got to see first-hand where all the inconsistencies lied and what we were missing.

This led to having to make very quick design system decisions that panned out like this:

"We're adding the ability to filter search results – a 'filter' button needs to be created. Do we have a standard component for this functionality? No? Let's create one – BUT – let's make this a component we can reuse. In fact, let's first establish what a button looks like (color, size, padding, corner radius, font). Before we can lock this in, we need to mock this up in other scenarios to make sure it's going to work. We need 3 versions of this same button: iOS, Android, and web. Let's find out what guidelines and restrictions we have on all 3 platforms and find how we can best make a universal component."

Pushback

With the company never having worked like this before, there was some initial pushback about why we were turning this small, seemingly simple task into something much bigger involving multiple teams that have never collaborated before. I had to advocate for why it was so important and necessary to establish these components and how it was going to be annoying at first, but it would actually make things much easier for everyone in the long run: No more guesswork and devs having to create things from scratch because the component we need would already exist.

I spent a lot of time in meetings with the C-suite explaining why certain decisions were being made and that there was a big-picture reason for why that [insert component/element] looked different and that there was a plan in place to roll it out site-wide alongside the other projects in the pipeline. This wouldn't typically be an ideal way of implementing a new design system, but given the resources we had, it was the only way it was going to happen while also addressing the app project backlog.

User Testing

Any time there was debate about a design aspect, I tested on UsabilityHub. Introducing UsabilityHub to the company was a game-changer – we could now quickly validate (or invalidate) design decisions and get user feedback on decisions we weren't sure about.

This involved a lot of A/B testing different ideas using mostly first click tests, where users are asked to locate something on a mockup of a page then comparing average time it took the user to find it. I'd follow with survey questions to get a better idea of user opinions.

The Switch to Figma

The designers were using Sketch when I first came onboard. While there's nothing inherently wrong with Sketch, it also requires the use of several other pieces of software for prototyping and dev handover. The company was paying for:

  • the Sketch license

  • Abstract for version control

  • InVision for prototyping

  • Zeplin for development handover

The number of platforms being used made collaborating with other teams confusing and difficult. It was hard to keep track of where the most up-to-date version of a project lived. Did someone forget to upload their latest iteration to Abstract? Has the design system been synced lately? What changes were made since this this prototype was uploaded to InVision?

I have been preaching the benefits of Figma since we made the switch at a previous company I worked for – it made life so much easier to have one tool for creation, collaboration, prototyping, and handoff. Since it's cloud-based, a design system can live in a master library that is updated in real time and syncs with components in every file the library is linked to.

It was a hard sell at first and I get it, nobody likes change and it is a pain to switch everything over from Sketch (as much as Figma tries to make it seamless). Since we were essentially creating a design system from scratch though, it seemed like the perfect opportunity to show how this could benefit everyone, so I started creating the new components in Figma and transitioning the Sketch files.

I tested a project designed in Figma with the app team and they were very receptive to the change. After determining that we could seamlessly complete a project using only Figma, everyone started getting on board. Not long after, New Zealand went into lockdown due to COVID-19 so we were all forced to work from home and since Figma is cloud-based, it proved to be instrumental in the collaboration process when everyone was working remotely.

Result

Wine-Searcher now has a functioning design system – with documentation – for the first time. The changes have been rolled out through most of the website and app, but development is still ongoing for some portions of the website.