The year is 2017. A new technology startup in the Denver area is born, and two UX Designers are hired on the same day in December. They are employees number 4 and 5. There are no users, no market-ready product, and no design process.
Where do they even start?
This is the story of how my co-designer Brenna Zumbro and myself, Klew Still, created a design process from scratch at the Golden-based startup, Active Oversight. It is part case study, part to-do list, part post-mortem and luckily only a very small part cautionary tale.
The first thing the newly-minted Active Oversight Design Team ever did was level-set on all our core design decisions. I mean the nitty-gritty: our fonts, our colors, our pixel grid and standard elements, even our file sharing software. We had some fairly complete branding packages, material from the alpha, and most importantly the support of a very design-first organization. These things were indispensable that first week. Working physically side-by-side with our Product Manager - also indispensable - we nailed down all of the base design decisions that would let us flow into a beta product with a consistent look and feel. This was a continuous coworking process that spanned about two weeks. For remote teams, I would strongly recommend coming together in a physical space for this kind of “design kickoff” event.
Those early days where Product and Design sat around one big table and occasionally wore costumes
Right away, we started implementing an abbreviated goal-based process. We knew we had discrete deliverables and we worked towards those first, working things out on the way. The only process chain we had defined at this point was basic but functional: we had a roadmap item to ideate, mock up, test, refine and pass to development.
These points in the process did not yet involve the wider team, and our test activities exclusively starred our friends and spouses as users. We knew that we had to nail down a more fully functional design process before we onboarded our first user. So we defined what I affectionately dubbed “Baby’s First Asset Pipeline.”
Halfway through forming our asset pipeline: using Sketch for mockups and good ol’ fashioned paper and pencil for prototyping
The tools we defined at this point have, in large part, persisted in our organization since that time. So be careful with you what choose - it just may stick! I am fairly tool agnostic and won’t recommend one asset pipeline over another, but in the interest of transparency I will reveal that our asset pipeline consisted of Sketch for canvas creation, InVision for prototype creation, and Adobe Creative Cloud for image and icon creation - with support from Product Team tools like Trello and Notion.
With that, the Design Team had a working but unrefined process to generate consistent and validated work.
Refining Our Process
That abbreviated and somewhat shallow process would not serve us for long, and we knew it. It was time to get robust with ideation and testing.
Under the leadership of our Product Manager, Kimberly, we started to expand our ideation process into more formal exercises, and we also began to involve team members outside of Design and Product. This was a huge leap forward for us.
A Mind Map, one of many ideation methods we began to use as our process matured
We experimented with a ton of ideation formats, such as:
- Brain Dump
- Crazy 8’s
- Design Studio
- Mind Mapping
- Journey Mapping
- … and about a million variations of Sketch Dump
These exercises opened up our creative and diverse thinking, and gave us a more structured and consistent approach to what was before simply parallel designing. But by far the most useful aspect of this new, richer ideation process was the practice of including team members from diverse areas of discipline such as Development and Operations. It cannot be skipped!
Including non-designers in timed Ideation exercises was a huge improvement at Active Oversight
We also matured our testing process somewhat, investing time into writing standardized test scripts. This was extremely helpful in forcing us to phrase more informative, more neutral, and less leading interview questions. We had a lot of early failures with poor test question phrasing, so I’d recommend taking a short UX Research course if you are at this point in your own journey. We also worked hard to access better testers. What is “better”? More representative of our target users and more likely to give unfiltered, honest feedback. Friends and family had helped out, but the time had come to get more real with our fake user.
Where did we find these new test users before we had built client relationships? Brenna was successful in finding volunteers who were more appropriate candidates by looking to users of similar software products in adjacent industries. We were also able to start sourcing engaged, knowledgeable testers by collaborating with our Services team. Look broadly!
At this point we also made an effort to refine our tool use. We had been lucky enough to choose tools that worked well for our team, but we had a very difficult time getting our mocks to development in a useful way. InVision’s Design System Manager solution was not yet released, so we looked to Zeplin for communicating mocks and associated data (margins, colors, etc.) to development. Adopting a new tool at this point was actually pretty difficult, even though we had only been set in our ways for a few months! The key was consistency, and once we made an ironclad habit of attaching our Zeplin mocks to our Trello stories so that devs could access them easily, the process stuck.
It quickly became clear that user testing was one of the most valuable, yet underutilized, points of our design process. Up to this point we had been primarily doing observed workflow testing with target personas as well as doing some short surveys and A/B tests internally. These were working for us, but they were simply not enough. Here are a few things we implemented to help mature our user research process:
1. Contact Journal
Writing things down is always a solid step 1
Brenna lead the charge to start codifying our user outreach with a Contact Journal, which was a document that tracked who we talked to, when, and what about. It allowed us to have target users to test for different roles and functions without making the mistake of bothering one or two responsive users constantly. It also allowed us to better communicate with our Implementation, Sales, and Pre-Sales teams which by this time quite busy with onboarding new users. A great problem to have!
Google Analytics is our tool of choice at Active Oversight
This topic fairly speaks for itself. There was some development overhead needed, but the visibility it gave us in Product and Design, and how it has driven our understanding of the way users interact with our experiences, has been necessary to understand our successful and unsuccessful features and shape future work.
Survey - right? Right? Get it?
Coordinating across teams, we were able to start sending periodic feature-specific and Net Promoter Score surveys to our user base. Here my main words of caution would be to not expect more than a 10% response rate, be prepared to field some real client concerns, and be careful with your distribution lists in the same way you are with your user interviews - you don’t want to overload your most responsive and engaged users.
4. Client Ideation
An in-house Client Ideation held by Active Oversight
We extended our user research practice even further by beginning to host client ideation sessions in our office. Working together with our Sales organization we selected groups of 3-8 clients who fell into our target persona and ran Design Studio-style ideations with them in person. This is generally a huge production, particularly the first time, but it is invaluable. When you include users in your process they feel a sense of ownership and delight in seeing their ideas come to life in the product. So many future roadmap items, industry learnings, and important connections were formed during these sessions that I cannot recommend them highly enough. Make calls, buy snacks, and do that client ideation!
A Look Back
Taking in the last nearly two years of design process at Active Oversight, I see that we had a lot of successes early on that allowed us to move quickly and confidently forward. Of course, there were a lot of external factors that supported our success, like a supportive and design-first organization, a good starting position, and consistent access to users as they onboarded. The outcome was a major reduction in rework and wasted work as compared with some of my other professional experiences.
Our greatest successes were the early and robust implementation of deep ideation practices, user research, and client ideation.
It wasn’t all roses, though, unfortunately. We should’ve better understood some foundational UX Research practices before we started live interviews to avoid tainted results early on. We also had tool uptake problems as we began to mature. One of our biggest problems is the growing pains we have gone through with our visual design language: in the beginning, with few functions and no clients, it was easy to create a tight and minimalist visual language. However, as the app expanded in functionality we had to adopt diverse interaction patterns, and sometimes those clashed badly with our other established work and work that predated the design process.
Design gobacks are exhausting
Design is never finished. You will always be iterating upon current design patterns in order to keep up with the market and users and to refine your past work - so while we can’t eliminate design rework, we can hope to reduce it and perform it strategically.
If I had it to do all over again, there are only a few refinements I would make. Truly commit to a new tool, because the developers won’t use it if you don’t. Start with a good understanding of fruitful research practices and begin to craft research plans, even brief ones, as soon as you are writing test scripts so that you get used to assessing test results against your goals and assumptions.
And good luck! Design from 0 to 60 is a bracing and break-neck exercise. Would we have it any other way?