Journal

Building the inline form system
After detailing how I planned to build inline forms in Cushion, I went ahead and built the first one for creating clients inline.
After detailing how I planned to build inline forms in Cushion, I went ahead and built the first one for creating clients inline.
In between big tasks, I distract myself with smaller tasks until I’m ready to finally face it—inline forms.
Chopping away at the new invoice form, I designed and built a new component for breadcrumbs.
Aggravated by people signing up with no real intention to use Cushion leads me to consider a “test mode” where they can try the app without signing up.
While redesigning the invoice form, I share the decision-making behind the design of the new line item footer.
Frustrated with Heroku, I start migrating away from it, beginning with continuous integration.
Upgrading Cushion’s database leads to a week of troubleshooting performance issues.
After receiving an email from Heroku to upgrade Cushion’s database, I upgrade Cushion’s database.
With Cushion reaching the 7-year mark, I reflect on its life and plot a course for its future.
At the fork in the road, I choose the invoice form as the next step in renovating Cushion, including a long-awaited list of improvements.
With a couple big launches behind me, I think about what to tackle next or whether I need to pick a direction just yet.
After launching Cushion’s new invoice page last week, I launch the new account page this week with a ton of other goodies.
With only days before the SCA deadline, I launch the new invoice page with to be SCA-ready while also paving the way for more improvements.
Wrapping up my work on the invoice page, I set up integration testing using Cypress to make sure that everything continues to work end-to-end.
Continuing to make progress on the invoice page, I talk about some of the ways I’m improving the codebase along the way.
A chat with a user leads me to consider a subsidized plan for the New Year to help folks through the pandemic.
I pause work on the account page to make sure Cushion’s invoice payment integration is ready for the SCA deadline.
In needing to redesign the pricing page for the new account section, I decide to rethink the structure of Cushion’s existing plans.
An innocent bug report unfolds into a roller coaster ride of a troubleshooting journey.
As I begin building Cushion’s new account page, I start by scaffolding a static version of it and building the components it requires.
With SCA regulation going into effect at the end of the year, I prioritize migrating to the Stripe customer portal, which means redesigning the account section.
After years and years of incrementally tweaking the design of Cushion, I take a holistic look at the layout system.
After redesigning Cushion’s homepage, I take the next step in migrating to Contentful by rethinking the blog.
To keep my mind busy while dealing with a tough time, I distract myself with Cushion’s new homepage animation.
While building Cushion’s new homepage, I go the extra mile to swap out better-fitting images for different breakpoints.
After migrating Cushion’s changelog to Contentful, I move onto the homepage, where I narrow in on an approach to compose pages entirely in the CMS.
In my initial step towards migrating Cushion’s marketing site to Contentful, I start with the changelog.
I reworked the main nav to include invoicing, which was met with a wave of thanks from users.
After launching Cushion’s new onboarding, I detail the next few items on my to-do list, which for once are not new features.
A weekend turned into a month, but Cushion’s new onboarding is finally live. Instead of focusing on the launch itself, I talk through my favorite part.
To provide a more pleasant codebase for Future Jonnie to inherit, I restructure the onboarding to be composable and easily testable.
I decide to write a few utility methods myself, rather than relying on 3rd party library, and I’m glad I do.
Torn between adding a dependency for a utility method or writing it myself, I talk through the pros and cons of both approaches.
Determined to launch thew onboarding flow, I forgot to write for a week and it shows.
Continuing my work on Cushion’s onboarding flow, I build a new and improved date picker component.
With my recent work on Cushion’s new onboarding flow, I take a meandering walk through the new components I built for it that are all typed and tested.
Stepping back to look at Cushion’s growth over its entire lifetime reveals a crystal clear moment when everything went south.
I start working on the new onboarding flow, but quickly realize it won’t be a weekend project.
As a first stab at onboarding, I design a flow that includes myself asking questions, which hopefully provides a friendly intro to the app.
In an attempt to ease people into the app rather than pushing them out of the airplane, I decide to bring back a proper onboarding flow.
After feeling reluctant to reveal any design progress, I decide to bite the bullet and share a few rough sketches—no matter how uncomfortable it feels.
Continuing the idea of the single click, I talk through a potential path to flattening the nav from multiple layers down to one.
Browsing through Cushion, I realize something strange. As I troubleshoot the anomaly, I discover an embarrassing truth.
I sent my first Cushion newsletter in over a year and a half, which sounds crazy, but nonetheless, it feels great to send them again.
Before diving into analytics, I decide to spend time writing database queries that could validate some of my assumptions using existing data.
After building Cushion so far based purely on assumptions and intuition, I finally start to think about using analytics to help guide me.
A thought-provoking conversation leads me to realize that Cushion is still solid and worth iterating on, instead of restarting from scratch.
After establishing a new guiding principle to keep API requests simple, I describe how I could refactor a complex view that relies on a heavy API request.
I discuss a pain point in the current Cushion involving the number of clicks it takes to get to where you need to go. What if it only took one?