Design is not decoration and it is not taste. A design job starts with a problem — someone needs an audience to notice something, understand it, and act — and it ends with an artifact that solves it. What separates a designer from someone pushing shapes around is a process: a repeatable sequence you can follow when inspiration does not arrive, and a set of principles you can point at when a client asks why the layout feels wrong.
The design process
- Define the problem. Write down what the piece must achieve, for whom, and by when. 'Make a poster' is not a problem statement; 'get Grade 9s to show up to the club fair on 22 September' is.
- Research. Gather the constraints — dimensions, budget, brand colours, deadline — and look at how similar problems have been solved. Collect references before you draw anything.
- Generate ideas. Thumbnails: small, fast, ugly sketches, at least ten of them. The point is quantity. You cannot judge an idea against nothing.
- Choose and develop. Pick two or three thumbnails and build them up into rough comps at real proportions.
- Produce a prototype or proof. A full-resolution mock-up of the chosen direction, in the real software, at the real size.
- Test and get feedback. Show it to someone in the actual audience. Watch what they look at first and what they misread.
- Revise and produce. Act on the feedback, then export to the correct format and specification for the destination.
- Evaluate. Did it achieve what step one said it had to? What would you do differently next time?
The four principles that fix most layouts
| Principle | What it means | What it fixes |
|---|---|---|
| Contrast | If two elements are not the same, make them very different — in size, weight, colour or typeface | A flat page where nothing draws the eye first |
| Repetition | Repeat a visual element — a colour, a rule, a heading style — throughout the piece | A design that feels like unrelated fragments |
| Alignment | Every element lines up with something else. Nothing is placed arbitrarily | A messy, amateur look you cannot quite name |
| Proximity | Group related items close together and push unrelated ones apart | A reader who cannot tell which caption belongs to which image |
These four are usually taught together as a checklist, and that is exactly how to use them. When a layout is not working, run down the list: is there a clear first thing to look at? Does anything repeat? Is every edge aligned to something? Are related things actually touching? One of the four is almost always the culprit.
Hierarchy and white space
Visual hierarchy is the order in which a viewer takes things in, and you control it. Size, weight, colour, position and isolation all pull the eye. Decide deliberately what should be seen first, second and third — a poster with three things all shouting has no hierarchy, and the viewer reads none of them.
- Balance — visual weight distributed across the page. Symmetrical balance is formal and stable; asymmetrical balance is more dynamic and harder to do well.
- Emphasis — one clear focal point. If everything is emphasized, nothing is.
- Unity — the piece reads as one thing, usually achieved through repetition and a limited palette.
- Movement — the path the eye travels. Lines, gaze direction and diagonals all steer it.
Working as a team, and giving useful feedback
Most real media work happens in teams, and most of the friction is about communication rather than skill. Critique is a skill in its own right: describe what you actually see before you judge it ('my eye goes to the date first' beats 'I don't like it'), tie every comment to the brief rather than to your own taste, and separate the work from the person who made it. When you receive critique, write it down and ask questions before you defend anything.
- "I don't like the colours"
- "Make it pop"
- "It's fine I guess"
- Rewriting the design for them
- "The orange and red are close in value, so the headline doesn't separate"
- "The first thing I read is the sponsor logo, not the date"
- "The brief said Grade 9s — this type feels aimed at adults"
- Naming the problem and leaving the solution to the designer