Miller's Law

Key Points

• •Working memory holds roughly 4 ± 1 items at a time (updated from the original 7 ± 2 estimate)
• •Chunking—grouping related items—lets users process more by treating groups as single units
• •The "magic number 7" misapplication: visible menus are scanned, not memorized, so 7-item limits on nav are unfounded
• •Working memory decays rapidly without rehearsal (roughly 20-30 seconds)
• •Cognitive load increases with each item held simultaneously in mind
• •Phone numbers, credit card numbers, and IP addresses all use chunking to aid memory
• •Miller's paper focused on chunking, not the number seven—he was surprised by its misinterpretation
• •Cognitive load theory emerged from Miller's insight that our working memory has inherent limits
• •[FIGURE 3-1 — Working memory buffer] Shows: Visual diagram of working memory as a limited buffer with finite "slots" for storing information. Teaches: Working memory has specific capacity—when exceeded, we lose information to accommodate new input, leading to task difficulty and frustration. Reproduce as: Interactive buffer visualization showing items entering/leaving memory slots.
• •[FIGURE 3-2 — Phone number chunking] Shows: Side-by-side comparison of phone number as raw digits (4805550123) vs chunked format (480-555-0123). Teaches: Chunking transforms a 10-digit string into 3 manageable chunks, dramatically improving processability and recall. Reproduce as: Interactive formatter showing raw vs chunked numbers.
• •[FIGURE 3-3 — Wall of text problem] Shows: Dense paragraph from Wikipedia with no hierarchy, formatting, or visual breaks. Teaches: Unstructured content increases cognitive load—users struggle to scan, process, and find information. Reproduce as: Before example in a content formatting comparison.
• •[FIGURE 3-4 — Wall of text improved] Shows: Same content with headings, whitespace, reduced line length, underlined links, and highlighted keywords. Teaches: Chunking through hierarchy and formatting reduces cognitive load and improves scanability. Reproduce as: After example showing hierarchy, spacing, and formatting applied.
• •[FIGURE 3-5 — Bloomberg chunking] Shows: Bloomberg financial interface with dense data organized into visually distinct groups. Teaches: Even complex, data-heavy interfaces become usable when chunking creates clear visual groupings. Reproduce as: Annotated screenshot showing chunk boundaries.
• •[FIGURE 3-6 — E-commerce chunking] Shows: Product grid with filters grouped into categories on an e-commerce site. Teaches: Chunking groups products and filters into discernible sections, helping users process large catalogs. Reproduce as: E-commerce layout with grouped filters and product grid.
• •[FIGURE 3-7 — Google Docs toolbar] Shows: Google Docs toolbar with functions separated into discernible button groups. Teaches: Chunking in toolbars groups related functions, making complex software approachable. Reproduce as: Toolbar mockup showing grouped vs ungrouped icons.
• •[FIGURE 3-8 — Nike navigation] Shows: Nike.com navigation with more than 7 items that remains scannable. Teaches: Visible navigation doesn't need 7-item limits—chunking and hierarchy make even large menus scannable. Reproduce as: Navigation example disproving the "7 items max" myth.

Guidelines

• ✓Break long content into meaningful chunks (e.g., group form fields by category)
• ✓Use progressive disclosure to limit simultaneous information load
• ✗Don't require users to hold information from one page to use on another
• ✓Keep multi-step processes to 3-5 visible steps when users must track progress mentally
• ✓Format numeric data with spacing or separators (e.g., 4242 4242 4242 4242)
• ✗Don't misapply: visible options in menus don't need a 7-item limit because users scan, not memorize

Formula

Working memory capacity: 4 ± 1 items (updated from original 7 ± 2)

Related Topics