DEV Community: Adam Marsden

🦄 You’re shipping vague buttons

Adam Marsden — Wed, 15 Apr 2026 13:47:00 +0000

Hey again 👋

This week's picks all sit in the gap, where the thing technically works but the message doesn't, and the fix is usually a better label, a sharper hypothesis, or being honest about what your product is actually competing with.

Enjoy 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.
This week’s email edition includes Useful Extras and one extra deep-dive pick (missing from this public post).
Subscribe: https://unicornclub.dev/

Sponsored by 20i

Peak Performance WordPress Hosting, No Compromises

Leave single-server hosting in 2015. Choose autoscaling Managed WordPress Hosting built for traffic surges, complex sites & demanding PHP workloads - without missing a beat.

Try 20i® now →

The three picks I'd send to a teammate:

Build: Your UI might look fine and still sound broken

Shape: A fast MVP still misses if you learned the wrong thing

Ship: Solving the problem isn't enough when people can leave

Top 3 this week 👇

Build: The invisible layer of UX most designers ignore ↗

A polished UI can sound like a messy one.

Why: If the role, name, or state is missing, a screen reader turns your nice interface into a stream of vague buttons and links, not really the experience you want for folk.

Adopt: Keep coming back to "role, name, state" and you'll start to catch a lot before handoff.

Shape: Minimum Viable Product (MVP): Definition ↗

A fast MVP is still a miss if you tested the wrong thing.

Why: The useful split here is value first, implementation second. Otherwise you're testing code when you really you were meant to test demand.

Adopt: Write your hypothesis first. If the risk is understanding then prototype it. If it's behaviour, ship the smallest real thing.

Ship: Commercial vs Internal Products ↗

Shipping gets harder when people have an alternative.

Why: Internal tools get more slack and commercial products have to beat whatever people already use, even if the fallback is clunky and bad.

Adopt: Good reminder to ask what you're asking someone to switch from, not just whether the feature works.

Dive into more

Build: Personal website redesign project post: Getting started with the HTML only build ↗ Temporary UI, permanent structure

Full email edition includes: one extra deep-dive pick.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Build: Vertical Slice Architecture in Node.js: One Folder Per Use Case ↗ Organise by use case

Ship: Designers will never have influence without understanding how organizations learn ↗ Influence usually follows decision flow

Full email edition includes: Useful Extras.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Support the newsletter

📢 Book a Sponsorship

Thanks for reading

Adam from Unicorn Club

Want the full email edition (including Useful Extras and one extra deep-dive pick) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

Follow me on X or BlueSky

Connect on LinkedIn

🦄 Craft skill won't make you influential. Neither will AI.

Adam Marsden — Wed, 08 Apr 2026 13:29:00 +0000

Hey, welcome to this weeks newsletter 👋

Most of us are confident about things we shouldn't be. Knowing what to hand off and what to keep is a skill that doesn't get easier just because AI can do more of the work...

Here are my three picks: CSS containment, design principles with a point of view, and why designers struggle to build influence.

Enjoy 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.
This week’s email edition includes Useful Extras and the full Dive into more section (missing from this public post).
Subscribe: https://unicornclub.dev/

Sponsored by 20i

Peak Performance WordPress Hosting, No Compromises

Leave single-server hosting in 2015. Choose autoscaling Managed WordPress Hosting built for traffic surges, complex sites & demanding PHP workloads - without missing a beat.

Try 20i® now →

The three picks I'd send to a teammate:

Build: CSS containment - the layout trick Harry Roberts says most sites are missing

Shape: Design principles deserve a point of view, not a committee

Ship: The real reason designers struggle to build influence at work

Top 3 this week 👇

Build: What Is CSS Containment and How Can I Use It? ↗

Harry Roberts on the one CSS property that changes how browsers calculate layout.

Why: Containment lets the browser skip recalculating the rest of the page when only a widget changes. Most sites don't use it. Find one in your product and test it.

Adopt: Find a self-updating widget, add contain: layout, measure the result.

Shape: A Practical Guide To Design Principles ↗

AI makes skipping the process look appealing. The thing is...

Why: You usually know when a principle is real because you've watched it go wrong. Good principles come from knowing what you've got wrong before, not from articles about what other teams got wrong.

Adopt: Choose your principles deliberately, not because someone declared process is dead.

Ship: Why designers struggle to build influence at work ↗

Craft skill is not the blocker. Influence is.

Why: Better craft doesn't usually change how much influence you have. What moves the needle: showing up before the brief gets written.

Adopt: Shift from executing deliverables to shaping decisions earlier in projects.

Full email edition includes: the full Dive into more section.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Full email edition includes: Useful Extras.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Want the full email edition (including Useful Extras and the full Dive into more section) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

Support the newsletter

📢 Book a Sponsorship

Thanks for reading

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

🦄 4.7% more subscriptions from one checkout change

Adam Marsden — Wed, 01 Apr 2026 13:52:00 +0000

Hey again 👋

Some things get set and never revisited. Site search, checkout currency defaults, whether the AI is the designer or just the builder.

Here are my three picks this week: the search paradox sending users to Google, what actually separates genUI from vibe coding, and what Stripe found across 1.5M subscription checkouts.

Enjoy 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.

This week’s email edition includes Useful Extras and a deeper Dive into more (missing from this public post).

Subscribe: https://unicornclub.dev/

Sponsored by 20i

Peak Performance WordPress Hosting, No Compromises

Leave single-server hosting in 2015. Choose autoscaling Managed WordPress Hosting built for traffic surges, complex sites & demanding PHP workloads - without missing a beat.

Try 20i® now →

The three picks I'd send to a teammate:

Build: Check your site search against words users actually use, not your own vocabulary

Shape: Ask whether the AI decided what to build or you did before evaluating the output

Ship: See what Stripe found when they tested local currency across 1.5M subscription checkouts

Top 3 this week 👇

Build: The Site-Search Paradox: Why The Big Box Always Wins ↗

Users search with their words. Your site only knows yours.

Why: 50% of users hit search immediately, then leave when it returns nothing for their vocabulary rather than yours.

Adopt: Search your site for a synonym of your top product category and count how many results come back empty.

Shape: GenUI vs. Vibe Coding: Who's Designing? ↗

GenUI is the AI's design call. Vibe coding executes yours.

Why: Who initiated the design determines what quality means: the AI's judgment call, or how well it executed yours.

Adopt: Ask whether the AI decided what to build or you did. The failure modes are completely different.

Ship: Testing the Impact of Adaptive Pricing Across 1.5M Subscription Checkout Sessions ↗

Stripe tested local currency on 1.5M checkouts. Conversion up 4.7%.

Why: In 2025, 80% of subscriptions still charged in the seller's currency. Local pricing lifted LTV per session by 5.4%.

Adopt: Check whether your checkout defaults to your base currency for international visitors.

Full email edition includes: Dive into more (5 extra links).

Get it in your inbox.

Subscribe: https://unicornclub.dev/

Full email edition includes: Useful Extras (2 extra links).

Get it in your inbox.

Subscribe: https://unicornclub.dev/

Support the newsletter

📢 Book a Sponsorship

Thanks for reading

Want the full email edition (including Useful Extras and a deeper Dive into more) and next week’s issue in your inbox?

Subscribe: https://unicornclub.dev/

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

🧐 Fast code isn't the bottleneck anymore. Judgment is.

Adam Marsden — Wed, 18 Mar 2026 20:56:00 +0000

Hey 👋

Bottlenecks in a lot of projects right now isn't the code anymore...

Here are my three picks on what to do with that: service architecture before the screen, how to read a stall, and what to write down before you close.

Enjoy 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.
This week’s email edition includes Useful Extras and Dive into more (missing from this public post).
Subscribe: https://unicornclub.dev/

Sponsored by 20i

Peak Performance WordPress Hosting, No Compromises

Leave single-server hosting in 2015. Choose autoscaling Managed WordPress Hosting built for traffic surges, complex sites & demanding PHP workloads - without missing a beat.

Try 20i® now →

The three picks I'd send to a teammate:

Build: Mapping APIs vs. Creating Screens before any UI work starts.

Shape: When a project slows, check whether it's build or alignment first.

Ship: Log "Trade-off" and a "Watch" before closing every project.

Top 3 this week 👇

Build: Re-writing Tapestry for AI workflows ↗

Change the architecture before you change the UI.

Why: Rebuilt a personalised CRM services-first when it turned out nobody actually wanted another app to log into.

Adopt: Map what's an API and what needs a screen before touching the UI on an AI-integrated build.

Shape: Software is a coordination problem. AI can't help you with that. ↗

Faster execution in the wrong direction is still the wrong direction.

Why: AI speeds up execution. But if the team isn't pointing the same direction, that speed just burns more runway.

Adopt: When a project slows down, check whether it's a build problem or an alignment problem with your team first.

Ship: Judgment and creativity are all you need. ↗

Cheap execution shifts the bottleneck to your decisions.

Why: When agents do the implementing, your team's time shifts to design decisions and review, which is where quality lives now.

Adopt: Two lines I add to every project close: Trade-off: [what I gave up] / Watch: [what this enables or breaks].

Full email edition includes: Dive into more.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Support the newsletter

If this was useful, here are two small ways to help it travel:

📢 Book a Sponsorship

Thanks for reading

Want the full email edition (including Useful Extras and Dive into more) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

⚖️ Design decisions need 5x more justification

Adam Marsden — Wed, 11 Mar 2026 12:30:00 +0000

Hey 👋

The dangerous moment in page sign-off is when your HTML breaks silently. The browser won't read it. You won't know until after ship, and by then the cost is credibility.

This week: a free test that catches it before sign-off, a name for the structural bias that makes design decisions take 5x longer to approve, and a two-line gate that stops underpowered tests from shipping.

Enjoy 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.
This week’s email edition includes Useful Extras and the full Dive into more section (missing from this public post).
Subscribe: https://unicornclub.dev/

Effortless feedback. Happier clients. Better websites.

Getting client feedback on websites is made ridiculously easy with BugHerd.

Clients simply point, click and comment
BugHerd automatically grabs a screenshot & saves technical details
A task is created for your team to track

Goodbye endless emails. Get a client feedback process that just works - get BugHerd!

Start a free trial →

Three things worth your attention this week:

Build: Test browser read-aloud before sign-off and treat silence as broken HTML.

Shape: Name what counts as good enough evidence before your next review.

Ship: Add "Practical threshold:" and "Act if:" before running any test.

Top 3 this week 👇

Build: Your Browser Can Already Speak a Page ↗

Treat browser silence at sign-off as a broken HTML report, not a quirk.

Why: A page that won't speak at sign-off has broken HTML structure and accessibility failures ready to ship.

Adopt: Run browser read-aloud before you sign off a page. If it won't speak, something in your markup needs fixing.

Shape: The Justification Tax ↗

Name your evidence threshold before design review starts, or lose the room.

Why: It costs you the ideas designers stop pitching because the approval bar is too high to clear.

Adopt: Before your next design review, add one row to your spec: "Evidence we'd accept." Fill it in before the session starts.

Ship: Statistical Significance Isn't the Same as Practical Significance ↗

Statistically significant doesn't mean practically important.

Why: Statistical significance means you ruled out luck, not that the effect is large enough to act on.

Adopt: Before running any test, add two lines to your task description: "Practical threshold: / Act if:"

Full email edition includes: the full Dive into more section.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Full email edition includes: Useful Extras.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Support the newsletter

If this was useful, here are two small ways to help it travel:

📢 Book a Sponsorship

Thanks for reading

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

Want the full email edition (including the full Dive into more section and Useful Extras) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

🔶 CSS shapes have been a hack. This fixes it.

Adam Marsden — Wed, 04 Mar 2026 16:28:00 +0000

Hey 👋

You can spot a CSS tooltip by what breaks. The focus ring disappears, the shadow detaches, the clip-path leaves a gap.

This week we have a CSS property that rewrites the box geometry directly, a one-question test for whether your design system contribution model has become overhead, and a postmortem structure that learns from success as much as failure.

Enjoy 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.
This week’s email edition includes Useful Extras and a Dive into more item: The Technical Decision Framework (missing from this public post).
Subscribe: https://unicornclub.dev/

AI Marketing OS For Founders

Crowdstake creates custom landing pages, writes copy, and grows your waitlist while you work. Launch fast, validate ideas, and get your first customers—no marketing experience needed.

Start building →

Three things worth your attention this week:

Build: Test border-shape next time you reach for clip-path.

Shape: Check whether your design sys contributions start with introductions.

Ship: Log each research recommendation with "Goal:" and "Shipped:" to track what reaches usersRun a postmortem after your next project with "What we knew:" and "What changed after launch:".

Top 3 this week 👇

Build: border-shape: the future of the non-rectangular web ↗

Tooltips and custom shapes without the clip-path workarounds.

Why: clip-path hacks that power most CSS tooltips kill focus styles and detach box-shadows from the border, but border-shape rewrites the box geometry so the background, focus ring, and shadow all follow the new shape automatically.

Adopt: Next time you reach for clip-path to build a tooltip or non-standard shape, test border-shape in Chrome Canary 146+ with the experimental web platform features flag enabled.

Shape: Design system contributions work better when everyone knows your name ↗

Most contribution models fail because the team got too big.

Why: In large organisations, DS contribution models scale process to match risk, which turns shared effort into a formal loop nobody trusts enough to complete.

Adopt: If your contributions start with introductions, the team has outgrown informal trust, so scope them to a named working group or replace the model with a lighter intake process.

Ship: Project Postmortems for UX Teams: Learning from Success and Failure ↗

Postmortems work for success as much as failure.

Why: Most UX teams only run postmortems after failures, which means they miss what worked: the decisions that held, the defaults that stuck, the approaches worth repeating.

Adopt: After your next project ships, run a 60-minute session regardless of outcome and log each key decision with two labels: "What we knew:" and "What changed after launch:"

What would make this newsletter more useful to you? ↗

Takes 30 seconds. Honest answers shape the content and direction of the newsletter.

Dive into more

Build: Understanding CSS corner-shape and the Power of the Superellipse ↗ — FrontendMasters on the maths behind corner-shape, how it differs from border-shape, and when you'd reach for each. This is the companion read to this week's Build pick.

Full email edition includes: The Technical Decision Framework.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Build: Sprites on the Web ↗ — Josh W. Comeau on sprite animation: when spritesheets beat CSS animation, with interactive demos and code you can lift for a loading or idle state.

Shape: The Hidden Users of Your Design System and How to Support Them ↗ — Developers testing in isolation, QA engineers, PMs making scope calls. Name them in your contribution guidelines and the onboarding section gets a lot more honest.

Shape: Software interoperability ↗ — Justin Jackson on why open software compounds over time. Worth reading before you commit to locking your team into a proprietary tool.

Ship: The engineeringification of everything ↗ — PostHog on reversibility thinking and measurement culture spreading from engineering into product and design. If your team hasn't named this pattern yet, they're probably already living it.

Useful Extras

Full email edition includes: Useful Extras.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Support the newsletter

If this was useful, here are two small ways to help it travel:

📢 Book a Sponsorship

Thanks for reading

Want the full email edition (including Useful Extras and a Dive into more item: The Technical Decision Framework) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

🛠️ 200 Figma variables. Two seconds.

Adam Marsden — Wed, 25 Feb 2026 14:34:00 +0000

Hey 👋

This week - a tool for offloading design-system busywork without losing design decisions, a validation layer that catches malformed components before they escape, and a tracking method that shows exactly where research recommendations die.

Enjoy 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.
This week’s email edition includes Useful Extras and AI Coding Summit (online event details + promo code) (missing from this public post).
Subscribe: https://unicornclub.dev/

How Pixelstorm speeds up website feedback with BugHerd

See how Melbourne-based digital agency, Pixelstorm, resolves website issues faster and launches with confidence.

No login required for clients
Clients simply point, click and comment
BugHerd automatically grabs a screenshot & saves technical details
A task is created for your team to track

Keep your clients happy - make it easy for them to provide website feedback with BugHerd.

Think about these:

Build: Try an MCP tool for design-system library work before building manually.

Shape: Add a validation check that rejects malformed component usage before it ships.

Ship: Log each research recommendation with "Goal:" and "Shipped:" to track what reaches users.

Top 3 this week 👇

Build: The MCP Tool That's Changing How I Use Figma ↗︎

Offload the library mechanics and keep the design thinking.

Why: MCP tools like Figma Console can automate repetitive design-system tasks (200+ variables in seconds), freeing time for the experience work that matters.

Adopt: Before the next library build, try an MCP tool for variable creation and component mapping.

Shape: Your design system has opinions. They're just not being enforced ↗︎

Documentation says modal needs actions. Production shipped without them.

Why: Without enforcement, design systems become hope: rules written down, trusted to be followed, discovered broken in production when a user stares at a dead end.

Adopt: Add a validation check that rejects malformed component usage before it leaves the design file or codebase.

Ship: Tracking Adoption of Research Recommendations: The Recommendation-Adoption Score ↗︎

Treat recommendations like inventory. Track where value gets lost.

Why: Research recommendations represent time, energy, and money, but without structured tracking, adoption gets overstated and breakage stays hidden until credibility takes a hit.

Adopt: For your next study, log each recommendation with "Goal:" and "Shipped:" labels so you can track what actually reached users.

Full email edition includes: AI Coding Summit (online event details + promo code).
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Dive into more

Shape: How design systems offer creative safety for product teams ↗︎ — Sameness is healthy when things that are the same look and feel the same; unhealthy when things that are different look the same. Use this lens in your next component review.

Shape: The Cost of Consistency: Avoiding Design System Bottlenecks ↗︎ — A rigid system turns a 5-minute padding change into a multi-week migration. Watch for when the "maintenance phase" cost rivals development cost.

Build: LLM-generated skills work, if you generate them afterwards ↗︎ — Self-generated skills before a task do not help. After the task, ask the model to write up what it learned for reuse next time.

Shape: The Technical Decision Framework ↗︎ — A framework for the choices that shape your codebase. Stop second-guessing your architecture, stack, and scope choices. Get the framework engineers use to make decisions they don't regret.

Full email edition includes: Useful Extras.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Support the newsletter

📢 Book a Sponsorship

Thanks for reading

Want the full email edition (including Useful Extras and AI Coding Summit (online event details + promo code)) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

🫣 The hidden danger of shipping fast

Adam Marsden — Wed, 18 Feb 2026 11:40:00 +0000

Hey 👋

The riskiest moments are when something “looks done” but nobody can tell if it will survive contact with real users.

This week is about making early value obvious, choosing foundations that carry forward, and shipping in a way users can actually adopt.

Jump in 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.
This week’s email edition includes Useful Extras and AI Coding Summit promo details (missing from this public post).
Subscribe: https://unicornclub.dev/

Get clear website feedback & launch faster with BugHerd

Move faster through website reviews, with less back-and-forth.

Clients simply point, click & comment
BugHerd automatically grabs a screenshot & saves technical details
A task is created for your team to track

With BugHerd, you’ll deliver websites faster and keep your clients happy.

Start free trial →

Think about these:

Build: Make the first minute feel safe, not clever.

Shape: Decide which work is disposable, then design for it.

Ship: Slow the noise, not the shipping, with release notes.

Top 3 this week 👇

Build: Building Digital Trust: An Empathy-Centred UX Framework For Mental Health Apps ↗︎

“Design for low capacity moments, not ideal attention and energy.”

Why: It turns trust into concrete UI defaults like validating onboarding language, low-stimulus screens, and opt-in sensory feedback, so vulnerable users do not bounce or feel judged.

Adopt: In your next onboarding or “help me now” flow, swap one demand for one relief using “It’s okay to…” plus an explicit optional path.

Shape: The hidden cost of AI prototypes that are made to die ↗︎

“A prototype that cannot extend is future rework in disguise.”

Why: It names the expensive failure mode: teams validate on output that is not portable or inspectable, then pay later in translation, rebuilds, and slowed iteration.

Adopt: Before approving a prototype path, add three lines to the brief: “Lifespan:”, “Portability:”, “Extend without rewrite: yes/no”.

Ship: The hidden danger of shipping fast ↗︎

“You can outship users’ ability to notice and adopt.”

Why: When release volume exceeds user attention, “done” becomes an invisible backlog, time-to-value stretches, and quality degrades through partial or misunderstood adoption.

Adopt: Publish each release note with “Trade-off:” and “Watch:” so teams stay selective about what gets loud and what stays quietly excellent.

Full email edition includes: AI Coding Summit promo details.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Dive into more

Build: The State of Buttons ↗︎ — When a control has both state and action, make the container carry state (shape, width, icon) so users do not have to guess.

Ship: On screwing up ↗︎ — After a bad deploy, skip the self-defence and report fast with a plain factual summary so the incident response is not working blind.

Shape: The Three Juggling Acts (Strategic, Lazy, and Survival) ↗︎ — If “keeping options open” is starting to feel like drift, use the five questions to force a pruning decision before you hit survival mode.

Build: HTML Tips - The hidden Gems ↗︎ — A grab-bag of overlooked HTML features, worth scanning when you are tightening forms, structure, or accessibility without adding more JavaScript.

Full email edition includes: Useful Extras.
Get it in your inbox.
Subscribe: https://unicornclub.dev/

Support the newsletter

📢 Book a Sponsorship

Thanks for reading

Want the full email edition (including Useful Extras and AI Coding Summit promo details) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

🔇 You’re breaking screen readers

Adam Marsden — Wed, 11 Feb 2026 20:31:00 +0000

Hey there 👋

We’ve made a couple of tweaks to the newsletter going forward. You’ll get to value faster because the very best of the week is now right up top. If you want to go deeper, the extra finds are waiting for you near the end.

This week we’re looking at the interface glue that decides whether “AI” feels useful or just noisy: accessible motion, a clear owner for design system exceptions, and a simple RISK check in release notes when model behaviour can shift under your feet.

Jump in 🦄 - Adam at Unicorn Club.

Get Unicorn Club in your inbox.

This week’s email edition includes Useful Extras and AI Coding Summit promo details (missing from this public post).

Subscribe: https://unicornclub.dev/

Sponsored by 20i

Peak Performance WordPress Hosting, No Compromises

Leave single-server hosting in 2015. Choose autoscaling Managed WordPress Hosting built for traffic surges, complex sites & demanding PHP workloads - without missing a beat.

Try 20i® now →

If you only steal three ideas this week, steal these:

Build: Block letter-splitting animations unless screen readers pass.

Shape: Give one person design-system veto power, with clear intake criteria.

Ship: Add a “Relevant, Inclusive, Safe, Kind” (RISK) eval line to release notes when model behaviour can shift.

Top 3 this week 👇

Build: You Know What? Just Don’t Split Words into Letters ↗︎

Stop shipping animated type that screen readers spell out.

Why: Splitting words into per-letter elements can make assistive tech read only fragments, and ARIA “fixes” fail (for example aria-label is prohibited on div with the generic role).

Adopt: Default to no per-letter DOM splitting. If you must, include screen reader results for two browser/reader pairings in the PR. No proof, no merge.

Shape: Your Design System Needs an Enforcer ↗︎

Consistency dies from a thousand “small” component exceptions.

Why: Local tweaks compound into dozens of variations (Laura Klein’s carousel example), which raises maintenance cost and turns learned behaviour into a cognitive burden.

Adopt: Put an enforcer in the design system contribution path with authority to approve exceptions. Intake question: “Would this help three or more teams?”.

Ship: The new UX Toolkit: data, context, and evals ↗︎

Non-deterministic output needs a release gate, not vibes.

Why: In rollout, large models change over time and without evals we cannot know what is failing or whether a new version is safer or worse for your product.

Adopt: Add a RISK eval line to release notes for model changes. Note whether it was human review, code-based evals, or “LLM as judge”.

Full email edition includes: AI Coding Summit promo details.

Get it in your inbox.

Subscribe: https://unicornclub.dev/

Dive into more

Build: AI’s text-trap: Moving towards a more interactive future ↗︎ — Assistants should render components, not paragraphs. Pick one workflow step and make it structured UI.

Shape: Surfaces, capabilities, and extensions ↗︎ — Separates UI surfaces from underlying capabilities and extensions. A useful lens for avoiding “blank canvas” platforms.

Ship: Refactoring internal documentation in Notion ↗︎ — Shows a practical doc cleanup loop when search and Notion AI make stale pages harmful.

Ship: TBM 405: Hope, Context, and Control ↗︎ — Legibility versus local context, and how AI can push either direction. Worth checking where your reporting strips the context people need to make trade-offs.

Shape: The circular logic of our metrics ↗︎ — Calls out “intuitive” as shorthand for “familiar” and how salience-driven metrics make products “yell”.

Full email edition includes: Useful Extras.

Get it in your inbox.

Subscribe: https://unicornclub.dev/

Support the newsletter

📢 Book a Sponsorship

Thanks for reading

Want the full email edition (including Useful Extras and AI Coding Summit promo details) and next week’s issue in your inbox?

Subscribe: https://unicornclub.dev/

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

🦄 Staging said “yes”. User tickets said no...

Adam Marsden — Wed, 04 Feb 2026 15:24:00 +0000

Hey again 👋

Design review is getting weird again. We’re arguing about labels and “process”, while staging looks green and support tickets tell a different story.

Get Unicorn Club in your inbox.
This week’s email edition includes the copy/paste templates (missing from this public post) plus AI Coding Summit (Online, Feb 26–27, 2026).
Subscribe: https://unicornclub.dev/

This week is about making work legible: collapse a two-step choice into one set of radio buttons, write a one-page “direction brief” for agent-built tasks, and add a quick incentive check before you turn a metric into a target.

Dig in! 🦄 - Adam at Unicorn Club.

Sponsored by 20i

Peak Performance WordPress Hosting, No Compromises

Leave single-server hosting in 2015. Choose autoscaling Managed WordPress Hosting built for traffic surges, complex sites & demanding PHP workloads - without missing a beat.

Try 20i® now →

🏗️ Build

Make better interfaces.

If you’re struggling to write the content, you probably have an interaction problem

This bites when a checkout design review gets stuck on radio button labels for delivery versus collection: put all options in one radio group and the content suddenly writes itself.

Why it matters: The trap is polishing labels to paper over a clunky step, which bloats copy and still confuses people, so redesign the radio group to remove the extra decision.
Try this: In your checkout flow this week, collapse a two-step choice into one radio group.

🧩 Shape

Shared foundations across teams.

What the vibe engineering workflow tells us about the future of UX roles

‘Vibe engineering’ only stays safe if you break work into tiny tasks and write down constraints for errors and empty lists.

Why it matters: What catches teams out is treating a coding agent like a magic sprint, which leaves edge cases and navigation states unhandled, so write a task breakdown and review plan first.
Adopt this week: For one feature, write a one-page spec with constraints, edge states, and verification steps.

Full email edition includes: AI Coding Summit (Online, Feb 26–27, 2026).
Get it in your inbox.
Subscribe: https://unicornclub.dev/

The creator of Clawd: "I ship code I don't read"

The thing that changes in "reviews" is you start reviewing the prompt and the test run that covers loading and error states, not just the diff, when code lands fast.

Why it matters: Teams often merge agent-written changes like normal commits, then pay with a verification bottleneck and brittle releases, so capture prompts and test evidence alongside the diff.
Adopt this week: Add a “Prompt + verification” block to your pull request template, and require prompt text plus the test command used before merge.

Copy/paste template available in the email edition: Prompt + verification block.
A ready-to-drop PR section so agent-written changes ship with prompts and test evidence, not vibes.
Subscribe: https://unicornclub.dev/

P.S. This week’s sponsor is 20i

WordPress hosting built for traffic spikes and staying fast.

Try it for $1 →

🚀 Ship

Release, measure, iterate.

How Product Discovery changes with AI

Worth it for reframing deployment approval: treat a production prototype of an onboarding screen as research, because desirability still needs humans even if feasibility is cheap.

Why it matters: Without a production prototype, teams trust staging feedback, pay with late pivots and support tickets, and this pushes you to validate desirability with real usage early.
First step: Ship one screen behind a flag in your release config. If activation stays flat or tickets rise over 7 days, turn the flag off and revert.

Copy/paste template available in the email edition: Production prototype check-in.
A quick log to capture what you shipped, what you learned, and the decision you need.
Subscribe: https://unicornclub.dev/

The Cobra Effect: When Good Incentives Go Bad

Goodhart’s Law is the warning for weekly dashboard review, when a signup form conversion target becomes the goal and teams start optimising the interface for the metric, not comprehension.

Why it matters: Metrics turn toxic when they become targets, costing you warped behaviour and worse interfaces, so pressure-test second-order consequences before you tie bonuses, roadmaps, or praise to a number.
Adopt this week: For one metric, add a second-order section to its analytics doc and link it in the experiment brief for dashboard review.

Copy/paste template available in the email edition: Second-order consequences check.
Two prompts to sanity-check what behaviour you’ll trigger before you turn a metric into a target.
Subscribe: https://unicornclub.dev/

Support the newsletter

📢 Book a Sponsorship

Want the full email edition (including the templates and AI Coding Summit (Online, Feb 26–27, 2026)) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

Thanks for reading

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn

🦄 Users aren’t stuck, your errors are

Adam Marsden — Wed, 28 Jan 2026 14:45:00 +0000

Hey 👋

Design review and QA keep getting stuck on the wrong thing: someone’s arguing over “pixel perfect” while the pricing page taxonomy is confusing and the form errors are doing a bad job of helping anyone recover.

Get Unicorn Club in your inbox.
This week’s email edition includes the copy/paste templates (missing from this public post) plus the AI Coding Summit event pick (with details and promo code).
Subscribe: https://unicornclub.dev/

This week, swap pixel-chasing for intent rules, tighten your diagnostics copy, and add one lightweight release loop that forces real user signal to show up fast.

Enjoy this week 🦄 - Adam at Unicorn Club.

Sponsored by 20i

Peak Performance WordPress Hosting, No Compromises

Leave single-server hosting in 2015. Choose autoscaling Managed WordPress Hosting built for traffic surges, complex sites & demanding PHP workloads - without missing a beat.

Try 20i® now →

🏗️ Build

Make better interfaces.

The Product-Minded Engineer: The importance of good errors and warnings

The thing that changes in QA is whether an error state helps someone recover, or just tells them they’re wrong. This excerpt treats diagnostics as a primary interface and pushes you to design them around real scenarios.

Why it matters: Teams ship tidy screens and shrug at error copy. That creates support tickets and abandoned tasks. This gives a scenario-first way to write actionable diagnostics early.
Adopt this week: Add this error-message template to your component README.

Copy/paste template available in the email edition: Add this error-message template to your component README.
A simple structure for writing error messages that explain the problem and the next step.
Subscribe: https://unicornclub.dev/

Rethinking “Pixel Perfect” Web Design

This reframes “quality” as preserving intent across viewports, localisation, and accessibility settings, using rules and tokens rather than offsets.

Why it matters: Chasing the last 2px pushes teams into magic numbers, which slows refactors and breaks with localisation, so shifting to intent and tokens keeps layouts robust.
Try this: Use this “Intent, not pixels” block. Use it to resolve design review feedback without magic numbers.

Copy/paste template available in the email edition: Use this “Intent, not pixels” block.
A quick review prompt that clarifies what must stay fixed versus what can flex across devices and settings.
Subscribe: https://unicornclub.dev/

🧩 Shape

Shared foundations across teams.

Why users get lost in your product

If users get lost in onboarding, it’s often because your product hierarchy is muddled. This makes taxonomy feel like real infrastructure by tying navigation labels, pricing pages, and analytics naming to the same underlying structure.

Why it matters: Unclear taxonomy makes teams ship new features with new names every time, which adds clarity debt and hurts self-serve upgrades, so this gives a simple hierarchy and naming defaults.
Adopt this week: Inventory your product nouns into three levels, then publish the hierarchy and naming rules in your internal product handbook (60 mins).

Full email edition includes: the AI Coding Summit event pick (with details and promo code).
Get it in your inbox.
Subscribe: https://unicornclub.dev/

User Panels 101

Recruitment drags in planning meetings when you need five user calls next week, and this frames a user panel as infrastructure, not a list. The useful detail is the governance step that prevents duplicate outreach and privacy mishaps.

Why it matters: Starting recruitment from scratch each time burns weeks and budget, and it skews who you hear from, so a panel adds continuity and faster scheduling.
Try this: Draft a one-page panel charter with segments, consent wording, and access rules, then file it in your research operations docs (60 mins).

P.S. This week’s sponsor is 20i

WordPress hosting built for traffic spikes and staying fast.

Try it for $1 →

🚀 Ship

Release, measure, iterate.

Feedback loops are the new unfair advantage

Fast feedback loops win in release week because they turn real usage into next-week changes, not next-quarter debates. The emphasis is on shipping small, watching closely, and responding quickly.

Why it matters: Without tight loops, teams ship confidently then wait weeks for signal, which locks in the wrong interface decisions, so this pushes for shorter ship and learn cycles.
Adopt this week: Post a feedback loop scan after each release using this five-line template, and include one decision ask before you start the next build (15 mins).

Copy/paste template available in the email edition: Post a feedback loop scan after each release using this five-line template.
A short post-release scan that captures what shipped, what you learned, and the next decision to make.
Subscribe: https://unicornclub.dev/

Support the newsletter

📢 Book a Sponsorship

Thanks for reading

Adam from Unicorn Club

Want the full email edition (including the templates and the AI Coding Summit event pick (with details and promo code)) and next week’s issue in your inbox?
Subscribe: https://unicornclub.dev/

Follow me on X or BlueSky

Connect on LinkedIn

🏎️ Speed is cheap. Trust gaps cost months

Adam Marsden — Wed, 21 Jan 2026 15:11:00 +0000

Hey 👋

Speed is cheap now you can ship a decent-looking interface quickly. Problem is you then spend months paying for confusion, trust gaps, and rework.

Get Unicorn Club in your inbox.

This week’s email edition includes the copy/paste templates (missing from this public post) plus the “weekly lanes doc” idea and why it works.

Subscribe: https://unicornclub.dev/

This week pick the right interface surface for each AI intent, run a lightweight audit that produces fixable observations, and tighten your release loop with risk-based QA.

Enjoy this week 🦄 - Adam at Unicorn Club.

Sponsored by 20i

Peak Performance WordPress Hosting, No Compromises

Experience next-gen Managed WordPress hosting with 20i®:

Genuine autoscaling that instantly adapts to traffic spikes
Turbocharged high-frequency CPUs for exceptional clock speeds
Unlimited global CDN pre-caching for lightning-fast load times

Try 20i® for $1 →

🏗️ Build

Make better interfaces.

Beyond chat: 8 core user intents driving AI interaction

Stop defaulting to a AI chat box in design review. Map each AI feature to a user intent and a UI surface like a review queue, canvas, or digest. It helps you design transparency, control, and failure states before you start building.

Why it matters: Treating every AI feature as chat is the trap, this framework forces intent and a metric you can actually validate.
Try this: Write an intent card for one AI feature (30 mins), then paste it into the design doc and the pull request description before review.

Copy/paste template available in the email edition: Try this: intent card.

A fill-in card to map an AI feature to intent, surface, success metric, guardrails, and failure states.

Subscribe: https://unicornclub.dev/

Everything I know about running UX Audits

This bites when support tickets climb and a redesign gets proposed by instinct in design review, because it lays out a UX audit that turns evidence into prioritised fixes. Use it on one flow like checkout to capture problems, evidence, and a recommendation an engineer can ship.

Why it matters: Without scope and objective, audits become a grab-bag of nitpicks. This process keeps you anchored to key performance indicators, complaints, and testable recommendations.
Adopt this week: Audit one critical flow (60 mins) and attach a one-page “problem → evidence → recommendation” summary.

🧩 Shape

Shared foundations across teams.

Design Systems for Software Engineers

The thing that changes in your system is you treat shared components as contracts: states, keyboard focus, loading, and analytics events are part of the definition, not follow-up work. It’s a grounded tour of design system engineering from design files to a code library, including how to catch visual drift early.

Why it matters: Most teams standardise visuals but ignore interaction states, which causes drift and slow fixes across the product, and this guide shows how to encode behaviour, tests, and ownership.
Adopt this week: Add a component contract section to one shared component (45 mins) and commit it to your documentation.

Copy/paste template available in the email edition: Adopt this week: component contract section.

A quick contract checklist so shared components include states, accessibility, analytics, and regression coverage.

Subscribe: https://unicornclub.dev/

Your problem framing is sabotaging your strategy

Steal this for planning workshops where everyone jumps to a feature, and force a shared problem statement that describes the behaviour change, not the technology, before anyone draws the UI. It keeps work tickets from reading like button-click instructions and producing exactly that experience.

Why it matters: If you only ship solutions, you optimise for clicks and busywork and the interface turns into a checklist, and this pushes teams to define the real customer problem together first.
Try this: Replace one solution-first ticket with a problem-design brief (30 mins) and paste it into the ticket description before your next design review.

Copy/paste template available in the email edition: Try this: problem-design brief.

A short brief that forces behaviour change, signals, and scope before anyone starts designing a solution.

Subscribe: https://unicornclub.dev/

P.S. This week’s sponsor is 20i

WordPress hosting built for traffic spikes and staying fast.

Try it for $1 →

🚀 Ship

Release, measure, iterate.

Workshopping ideas for our future in Quality Engineering

Quality engineering is less about more test cases, and more about whole-team habits that show up in QA: shared language, hard questions, and fast feedback loops.

Why it matters: What catches teams out is assuming quality is a final gate, which pushes bugs into late QA and incidents, and these ideas pull risk and learning earlier into everyday delivery.
Try this: Run a risk brainstorm on one release-critical screen and capture the top five risks.

Solving Problems the Hard Way

Full email edition includes: the “weekly lanes doc” idea and why it works.

Get it in your inbox.

Subscribe: https://unicornclub.dev/

Thanks for reading

Adam from Unicorn Club

Follow me on X or BlueSky

Connect on LinkedIn