Design For Ten

Design For Ten means you design as if every pixel knows the names of its users. You build for a tight circle of ten specific people instead of chasing some imaginary market of thousands. The maker knows their workflows, their pet peeves, their preferred devices, and even their coffee orders if it matters to the interface. Context is assumed from day one. Onboarding disappears because your users were in the room when the problem was born. This discipline flips every design default. Instead of generic patterns that survive A B tests you optimize for the specific. Dark mode is not a toggle. It is the only mode because you know everyone hates bright interfaces at 9 p.m. The seven core principles are simple. Name every user literally and document what they need what they hate and what they tolerate. Skip the onboarding entirely and drop users into a configured app. Optimize for each individual instead of a nonexistent average. Let the interface be ugly in places that do not matter to these ten humans. Make every screen editable so changes take minutes not weeks. Design for one device because universal design is a tax for audiences this small. Plan the tool to die gracefully when the use case ends instead of pretending it will last forever. These rules produce software that feels like a custom tool rather than a product. Designers spend their time setting sharp context for AI assistants like Claude, exercising taste to reject mediocre outputs, and editing live alongside the users. The Figma file dies. The running app becomes the artifact that gets shaped over time. Geoffrey Litt's malleable software ideas, Linus Lee's tiny thinking tools, and Maggie Appleton's barefoot developer concept all feed directly into this way of working. One person with taste and a specific problem can now ship without permission or a team. The shift mirrors the larger move from mass market SaaS to personal software. Distribution happens in a group chat. Pricing is often a shared dinner or a one time payment. The incentive is solving the problem not capturing a market.

It is not mass market design wearing a smaller hat. Mass market work obsesses over strangers. How do we explain this to someone who just clicked an ad. How do we reduce cognitive load for the worst case user. How do we make sure nothing breaks when ten thousand people hit the server at once. Design For Ten throws those questions in the trash. It also is not no code templated slop from the 2015 era. Tools like early Bubble or Webflow handed you presets and forced your needs into their shapes. Design For Ten starts from your exact situation and generates the shape required even if that means a column for Kyle gets paid his ten percent flat rate this month. It rejects the idea that every app must look polished enough to show a stranger on a sales call. Ugly in the right places is a feature here. It is not about building something that can grow into a startup. The lifespan is tied to the use case. When the soccer season ends the league scheduler gets archived without drama. Reliability requirements drop because your ten users will text you directly if something breaks. You skip the reliability theater that bloats most SaaS products. Documentation shrinks to a single note with the makers phone number. The entire approach only became viable after AI coding tools and zero cost deployment platforms like Vercel, Convex, and Lovable removed the economic barriers that made small software impossible before 2025.

A concrete example sits on the laptops of a seven person indie game studio in Portland. They got tired of bending their process to fit Linear and built their own task surface instead. The interface opens directly to a kanban board colored by each developers energy patterns. It knows that Alex burns out if he sees more than three red tasks in a row so it auto rearranges. It knows that the sound designer Sarah needs voice notes attached to every asset review because she thinks better out loud. It knows the audio engineer only works on desktop and the writer prefers mobile so those views get prioritized. The dashboard pulls from their private Convex database and deploys updates through a shared Vercel preview link. One designer spent a single evening describing the workflow to Claude in plain English. The resulting app has no settings menu because every preference is already baked in. When the team wanted to add a field for bug bounty amounts they just asked the AI to modify the schema and the change went live in minutes. No product manager. No quarterly planning. Before this studio used a generic SaaS product that required seventeen custom fields and still never quite fit their Monday standup rhythm. After they had a tool that spoke their language and adapted the same weekend their process changed. The tool has been in daily use for two years. It will never have ten thousand users. It does not need to. Its value is the uncomfortable precision with which it matches the seven brains it was built for. Screenshots of this kind of work now fill the best designer portfolios because they demonstrate real impact on real people instead of hypothetical users.

Use Design For Ten any time your audience is a named group small enough to fit in one room or one chat thread. It is perfect for the household Notion setups that track family schedules with views no outsider would understand. It fits the custom invoice generator that knows every clients payment quirks and tax setups for a four person consultancy. It matches the personal software wave where tools like Tana and Capacities let users reshape their own systems and Obsidian plugins get swapped between people with similar brains. Lean on it when you are the barefoot developer solving your own problems or the designer sitting next to the maker. The feedback loop is measured in hours not quarters. Bad design gets fixed the same weekend. Reach for these principles when building tools for small teams of researchers, tight knit writing groups, families with unique rituals, or the food plan pi style single purpose apps that feed exactly six people with their exact dietary edges. The long tail of use cases that SaaS ignored for decades finally becomes accessible. Do not use Design For Ten when your growth model requires acquiring strangers through SEO or advertising. It will not work for products that need to serve both power users and complete beginners at the same time. Skip this approach if you are designing infrastructure used by millions or universal tools like email clients, calendars, or browsers. Those still require the old mass market rules. If your client asks for something they can take to a Series B pitch deck then they are not asking for Design For Ten.

Design For Ten turns every design decision into an act of radical specificity instead of diluted generality.