Type hierarchy begins with reading order.
Hierarchy is easiest to judge with actual content on a narrow screen and a wide one.
Editorial team ·
Start with a question
Before choosing a type scale, ask what the reader must notice first. The answer might be a page title, an instruction or a warning, depending on the content. Design the order in the document before designing its visual weight. A large decorative word can compete with the useful heading, especially when the screen is narrow. This is a self-initiated layout exercise, not evidence of behaviour observed in a deployed product.
Treat line length as a constraint
Very long lines make it harder to find the start of the next one. An excessively narrow column turns ordinary sentences into a stack of short fragments. Use a readable measure and test it with the longest plausible labels, code snippets and quotations. A type scale that works only with a three-word headline is not yet a resilient system. Allow content to wrap without clipping and give headings enough room before the next section begins.
Build contrast across several cues
Weight, size, spacing and position can establish levels without asking colour to carry all the meaning. Check smaller labels separately: a faint eyebrow may look elegant in a large mock-up but disappear on a phone. Use sufficient text contrast against the actual background under each state. When text sits over art, simplify the background or add a suitable surface rather than assuming every image region is equally readable.
Test the awkward cases
Replace ideal sample text with a long heading, a short paragraph and a dense list. Zoom the browser, increase text size and inspect the reading sequence without CSS. If an icon replaces a word, consider whether an accessible name is needed. Record which constraints were actually tested. This note documents a design hypothesis and the checks it requires before any real deployment, not a claim of measured improvement.
Related work
Continue with the notes index, or explore the overview for the wider context. Apply the question to an original sketch or a small piece of sample content. Check reading order with a keyboard and a narrow viewport before making claims about usability. These conceptual studies present design reasoning, not commissioned work or validated visitor results.