Deleting words does not lower comprehension cost

What it is

The task-level cost of concise copy evaluates the total effort required to understand correctly, decide, act, and recover, rather than optimizing word count. Removing repetition and irrelevant detail can reduce effort. Removing an object, condition, relationship, or next-step cue merely transfers cost from characters on screen to inference, hesitation, rework, and help seeking. Concision is the outcome of supporting a task with sufficient information, not the shortest possible sentence.

Why it happens

Interface copy sits inside controls, visual hierarchy, and prior action. A reader must integrate the words with the active object, system state, available actions, and consequences to form an actionable judgment. A short phrase that depends on instructions scrolled out of view, an ambiguous pronoun, or an unstated condition may scan quickly but force backtracking, guessing, or trial and error. A somewhat longer self-contained message may lead to the right action sooner. At the other extreme, repeated facts, politeness filler, and prematurely expanded secondary detail compete with the decisive cues. The useful optimization target is the path from seeing copy to correct action and necessary recovery, not the isolated string.

Studying it

Compare concise and fuller versions while holding core propositions, available actions, visual placement, and interaction path as constant as possible. Measure paraphrase accuracy, decision accuracy, time to first correct action, hesitation or backtracking, help seeking, errors, and recovery time. Post-task cue attribution can reveal which words supported the judgment and which were decorative. Separate first-time from repeated use because familiarity makes short labels more predictable. Screen size, typography, language proficiency, domain knowledge, and entry context can also confound a word-count effect. If a rewrite removes decision information as well as words, report the cost of information removal rather than a pure sentence-length effect.

Where it stops holding

This is not a claim that longer is clearer. Familiar low-risk actions, spatially stable navigation, and states with one plausible interpretation can use very short labels. Recognizable icons paired with accessible names may reduce visible copy. A professional term can be both shorter and more precise for experts while increasing cost for novices. Statutory, contractual, or safety information cannot be removed merely because a usability preference favors brevity. When the principal problem is layout, visual priority, or action architecture, further sentence editing will not eliminate task cost.

Applying it

• State the user judgment the message must support: what happened, what must be decided, and what comes next. Mark each word's informational, disambiguating, connective, or tonal function before removing words that contribute none.
• Review copy inside the actual control and state. Check whether a reader must return to an earlier page, open help, or try the action to learn its object, condition, or result; do not approve strings by character count in a spreadsheet.
• Separate first-use or infrequent, practiced, and consequential tasks. Stable low-risk labels can lean on context; unfamiliar or decision-dense moments retain self-contained cues and move secondary detail into progressive disclosure.
• Test concise and complete versions on the same task. Deletion lowers cost only when decisive information survives, errors do not rise, and end-to-end time or recovery effort falls.

Related

• Same group: T1.03.2 Omitting the subject creates unclear reference · T1.03.3 High-stakes situations favor completeness over brevity
• Adjacent: T1.07.1 Readability formulas are only rough proxies · T2.10.3 Too much microcopy dilutes the critical few
• Search terms: task comprehension cost · concise copy · paraphrase task