Case Study Club

Usability testing is about to have a comeback

The bottleneck in product used to be the cost of building.

Engineers were expensive. Shipping took months. So we put the filter up front ; research, discovery, design. Figure out what to build before you commit to building it. Designers were the gatekeepers of "is this worth doing."

That cost has collapsed.

Building is cheap now. A team can ship in a week what used to take a quarter. And the moment building got cheap, design started feeling like the bottleneck.

So design is being skipped.

I'm seeing it across teams I work with. Discovery shrinks or disappears. Research gets cut. The pressure to ship pulls everything forward, and the upfront filter ; the thing designers used to own ; gets thinner every sprint.

The filter doesn't disappear though. It just moves.

You can ship without research, but you can't ship without ever finding out if the thing works. The filter is moving from before-build to after-ship. And that means usability testing is about to have a comeback.

Simon Willison made a similar point on Lenny's podcast last month. When anyone can vibe-code a convincing prototype in an afternoon, the differentiator isn't who builds them ; it's who knows which one to bet on. Product judgment. Taste. Usability testing. Same shift from the engineering side.

Two flavors:

Moderated testing for the deep questions ; is this even the right thing, are we solving a real problem, what are we missing. The kind of thing you can't A/B your way to.

Unmoderated testing and A/B for the surface ; which variant wins, which copy converts, where do people get stuck. Cheap, fast, runs while you sleep.

The strategic work didn't disappear. It moved with the build.

If you spent the last few years sharpening discovery and research skills, don't put them down. Point them at what ships. That's where the filter lives now.

— designtools.fyi launches as a filter-based comparison database for design tools, sortable by capability rather than brand. designtools.fyi

— UC San Diego study finds 55.8% of LLM-generated ecommerce components contain at least one deceptive design pattern. UX Collective

— Nikhyl Singhal forecasts 30,000 PM roles eliminated within two years, with 8,000 AI-first roles created. Lenny's Newsletter

— Microsoft Design publishes "Design isn't dying. It's shifting left." Microsoft Design

— Matt Webb publishes "Headless everything for personal AI," arguing personal AI works better speaking to APIs than GUIs. Interconnected

Every Mac since 1983

Sheets.works has charted every Mac since 1984.

The early years are chaos. Apple competes at every price point with overlapping lines. The release cadence looks more like a printer company than a design one.

Then Jobs comes back in 1997 and draws a 2x2 grid on a whiteboard.

Ive's era pushes it further. Fewer products, deeper investment, materials and manufacturing doing the work that twelve overlapping model numbers used to do. Apple Silicon accelerates the same logic.

Forty years, and the strategic shape moves from "compete on every shelf" to "fewer, sharper products." The decisions about what not to build start outweighing the decisions about what to build.

An interactive map of Unicode characters

Charcuteria is an interactive Unicode character browser by David Aerne at Elastiq Studio. Instead of forcing a linear scroll through code points, it treats glyph exploration as a relational problem: shapes, scripts, and symbols cluster by visual and semantic adjacency, so the connections between related characters surface as you move through the standard.

For designers working with type, internationalization, or any symbol system outside Latin, this is a faster way in than the Unicode standard itself.

Written and edited by Jan Haaland. Published weekly from Norway.

Forwarding this to one person is the best way to support Case Study Club. If someone comes to mind, please send it their way.