First you work out who comes here and why, what they need to get done in two minutes, what their path to the goal looks like. Good looks follow from that, they are not the task. Half of those conversations ended right there, with the client going off to find an artist. Early on I sometimes gave in, out of inexperience, and those projects turned out poorly. The clients who listened stayed happy with a result that held up over the long run.
Since then the scale of my work has grown a thousandfold. I designed large B2B systems, worked in fintech, and a year ago I took on a project at national scale: how data moves between the producer, the ministry, and the cabinet, how a person in a village gets a service that never used to reach them. The approach stayed the same. What changed is the cost of a mistake. It is no longer measured in conversion points but in how many people get the help they need in time, and how many spend weeks on something that should take minutes.
The longer I work, the clearer one thing becomes: a screen and a country are built the same way. The scale changes, the number of players changes, the weight of the consequences changes. The principle holds. That is what this piece is about.
Drawing and designing
The whole point sits in the difference between those two words. To draw is to start from a picture: here is the screen, let it look nice. To design is to start from a goal and its constraints: here is the person, here is what they need to get, here is what stands in the way, here is what we build the path out of. There is an artistic side to design too, but it is one part of the work, not the work itself. The color of a facade is part of a house, not a substitute for the foundation and the load-bearing walls.
I trained as a mathematician, so it is easier for me to think about this through a fractal. A fractal is a structure built by the same law at every level of zoom. Zoom in on a branch and it repeats the shape of the whole tree. Zoom in again and the twig repeats the branch. The systems people make are built the same way. A national system is made of services, a service of scenarios, a scenario of individual screens and forms. At every level the same question: who comes here, what do they need, what stands in the way, what do we build the path out of.
This is the thing that holds my whole practice together. Once you learn to design something small and carry it through to the end, you start to see the same law in the large. A single form a person fills in and the link between whole agencies rest on one question: what does the person coming for a result need, and what stands in their way. The difference is in size, not in principle.
The measure of quality: attention to the person on the other end
There is a simple way to measure the quality of any system: how many details in it are thought through for the person who uses it. Not for the owner, not for the report, but for the one who comes for a result. The more such details, the better the system treats its user, and the better it works.
I once described this as levels of attention, from the lowest to the highest. At the lowest, a system is built from the inside out: around what suits the people who build and commission it. The person on the other end is not considered at all. The higher levels each account for that person more fully: first you simply take an interest in what would be easier for them, then you study who they are and what gets in their way, then you build the whole system around their path. There is a ceiling too, beyond which attention to detail turns into overengineering, but large government systems rarely reach it. More often they get stuck at the lower levels.
Here is a real example of the lowest level at national scale. In Chile a person with forty years of driving behind them cannot get a license until they pass a school exam for eight grades of education. The reason is not that anyone doubted their ability to drive. It is that one agency cannot confirm their schooling, obtained in another country, with another agency. The data exists, but there is no bridge between the two states' systems, so the person is sent to sit an eighth-grade exam just to be let near the driving test. The system was built from the inside out. The person passing through it was not considered.
That attention is what gets lost most often as a system grows. The larger the structure, the further the designer sits from the actual person, and the easier it becomes to build everything from the inside out.
Everything affects everything
You cannot touch one branch of a tree without touching the rest. Cut off a lower branch and sunlight reaches leaves it never reached before, and the whole tree responds. Systems are the same. One careless step at the start of the path blocks everything that comes after it.
I saw this up close in El Salvador's agriculture. There is a government service meant to carry knowledge and support out to producers. At a meeting about my project, with people from a UN agency in the field, a figure came up: that service reaches about twenty-five thousand people out of more than four hundred thousand. The other ninety-plus percent stay out of contact with it. The data is there, and plenty of it: on soil, on weather, on the farms themselves. It just sits in a form that does a farmer in the field no good at all. The same question goes unanswered: what does the person on the other end need, and does it reach them.
That is the cost of a careless start at the scale of a country. On a website the same gap would cost a point of conversion. Here it leaves hundreds of thousands of people without the service the whole thing was built for. Different scale, same law: the quality of the whole is decided in the first, earliest steps of design, not the last. When the leaves start to dry, watering the leaves is pointless. The cause is in the root, and that is what you water.
What does change
A large system is not simply a large website. The principle carries over, but some things change with scale in a way the designer has to know.
When you make a product, it is usually clear who owns it and what number success is measured by. In a system that stitches several agencies together, that clarity is gone: each has its own budget, its own tasks, its own boss, and no one owns what happens at the seam between them. This changes the work itself. Half of the design here goes not into the interface but into agreeing who answers for which piece and how the pieces meet.
The second difference is harder still. On a product a team can rewrite a bad module in a week. In a government system that same module is often fixed in a law or a regulation, and to change it you change the rule, not the code. The speed here is set by procedure, not by engineering.
So what carries over is the thinking, not the tools or the tricks. The programs I worked in twenty years ago are long forgotten. The way of thinking stayed and works at any scale: where to begin, what counts as the first step, how to measure quality by attention to the person. A tool lasts a few years. The principle rests on how systems are built in the first place: it describes how any complex structure grows and holds together. That is the foundation.
What is left behind
A well-designed system has one property that only shows with time. People keep using it long after the designer has left. A national system outlives the team that built it and the government that commissioned it. This is what separates design from drawing a picture: you lay down what will work for people you will never meet.
There is a line I have kept in front of me for years:
"No one deserves more admiration than the one who plants trees in whose shade he will never sit."
Twenty years ago I planted small shrubs: websites that thousands of people passed through. Now I design a forest the size of a country. But I build by the same law: you choose the spot, think about the soil, the wind, the climate and the light, and lay down what will grow and bear fruit long after you.
