Learning graphic design from scratch

What does a graphic designer actually do?

They decide. Graphic design means arranging text, image and empty space on a surface so that one specific message reaches one specific person. The exported file is the result of the work, not the work itself; the work is the set of decisions behind it.

Three decisions repeat in every project and the rest are details. First: what gets removed. A beginner keeps everything and then hunts for a layout that will fit it all. Second: what takes first rank, meaning that if the reader sees only one thing, what should that thing be. Third: what repeats, because repetition is what turns a handful of scattered pages into a brand.

Every real job also arrives with four constraints, and a good design is one that has come to terms with all four. The surface: how big it is and from what distance it will be seen. The reader: who they are and in what state they look at it, walking past on a street or sitting at a desk. Production: whether it will be printed or stay on a screen, and what the budget is. And the brand: the thing that already exists and that your design must not contradict.

Our position here is blunt: a client does not pay for the file, they pay for the decision, and the file is the receipt for it. Someone who only knows the software applies whatever change they are told to apply; a designer says this change breaks the message, and brings a reason. That second sentence is what sets the price of the work.

How many branches does this field split into?

Five branches, and in practice they mix far less than job adverts suggest. Brand identity: the logo, the palette, the type and the small book that tells everyone else how to use them. Print: catalogues, packaging, business cards, where colour mode, bleed and resolution start to matter and where, unlike the web, a mistake cannot be taken back. Digital and social: posts, banners and ads, at small sizes and on screens no two of which show colour the same way. Motion: the same design, plus time. And user interface: designing something that will be operated, not only looked at.

What gets said less often is that nobody is good at all five, and nobody needs to be. The branch is what decides which tool you work in, what format you export and where the job goes wrong. A print designer who hands over a file in RGB has handed over dull colours on paper, and at that point nothing can be done. A social designer who leaves fine detail in loses the design on a phone screen. These are not lapses of taste, they are lapses of branch.

Two of the five have their own paths on this site and that is the right place to read them: motion and video and the principles of interface design. The rest of this path stands on the first three.

Why does most of the work stay below the waterline?

Because the only thing a client sees is the final file, and that is the smallest piece of the job. Underneath it sit five layers, none of which show up in a portfolio and any of which can sink the project.

First the brief: what problem this design is meant to solve. A sentence like "I want a poster" is not a brief; "I want people walking past the shop to realise we are open on Saturdays too" is a brief. Second the constraints: size, viewing distance, print budget, where it will be mounted. Third the versions that were rejected, which in practice take most of the project time and are shown nowhere. Fourth the production spec: colour mode, bleed, the smallest size at which the logo is still legible. And fifth the handover: source files, the fonts and the right to use them, and who is able to edit this later.

The fifth layer causes more trouble than the rest and is talked about least. A logo is not a file; it is a family of files at different sizes and in different states, plus the rule for where each one is used. If your handover is a PNG, then six months later the client wants a shopfront sign, comes back, and you draw it again from nothing. And if the font you used carries no commercial licence, it is not your problem that ends there, it is the client problem that starts there.

What should I learn first?

Five rungs, in this order, and the order matters more than the speed.

The first rung is looking, the cheapest of them and the most ignored. Collect twenty designs that are doing their job well and write one sentence for each about why it works. The sentence has to be specific: "because only one thing is big and everything else is genuinely small" is a sentence, "because it is clean" is not. That one exercise, over a few weeks, is what makes it possible for you to talk about a design at all.

The second rung is the principles: hierarchy, contrast, alignment, whitespace and repetition. Learn them not as laws but as explanations, explanations of how the eye groups things. The third rung is colour and type, the two places where amateur work gives itself away most, and happily both of them take rules.

The fourth rung is one tool, only one, until you know its shortcuts without looking. And the fifth is a real job: for a real person, with a real constraint and feedback you will not enjoy. Practice without a client has a ceiling, because you write the brief and then you also approve it.

What can I practise with, at no cost?

Three free, open source applications cover all of the practice work, and the principles in this path are identical in all three. Inkscape for vectors, meaning logos, icons and anything to be printed large. GIMP for pixel images. And Penpot for page and interface layout, which is open source and can be self hosted. Figma also has a free plan and it is enough to start with.

Type does not have to be bought either, in order to practise. Google Fonts for Latin, and for Persian, Vazirmatn, published under the SIL Open Font License and carrying its own Persian digits. That second detail matters more than it looks: a font without Persian digits borrows them from another font, and at that moment their shape and height stop matching the rest of the sentence.

And there is one exercise that genuinely works: pick a good design, look at it properly, close it, and rebuild it from nothing without the original in front of you. Then put the two side by side. The gap you see is your list of lessons, and unlike watching a tutorial you discovered this list yourself, which is why it stays.

One blunt limit belongs here too: these free tools have no ceiling for learning, and they do have one for professional handover. A print ready file with the right colour profile, and exchanging files with a team working in something else, are the two places where sooner or later you have to make a decision.

How long until my eye starts working?

We will not give a number, because any number we gave would be made up. What can be said is that the work splits into two pieces moving at very different speeds, and courses sell the fast piece while implying it is all of it.

The fast piece is teachable: the vocabulary of the field takes a few weeks, one tool takes a month or two, and production knowledge settles after a handful of real jobs. The slow piece is taste, and taste is really memory: the memory of things you have seen and thought about. Memory cannot be downloaded, and that is precisely why no course shortens it.

So if somebody sells you a course that makes you a professional designer in a stated period, what they are really selling is the fast piece. There is nothing wrong with that, as long as you know where the rest has to come from: from looking, from rebuilding, and from taking on work. The only genuine shortcut is making that looking deliberate and regular instead of waiting for it to happen on its own.