The tendency for people to remember items at the beginning (primacy) and end (recency) of a list better than those in the middle. In navigation and menu design, this guides the placement of the most important items at the first and last positions — not buried in the middle where attention is lowest.
Common contexts
- Placing a primary CTA at the last position in a mobile bottom navigation bar rather than in the centre where it receives the least attention
- Reordering a pricing page feature comparison list to put the most differentiating features first and last, not in the middle
- Structuring a survey's response options so that the most common answer isn't lost in the middle of a five-item scale
Use when
Apply the serial position effect when designing any ordered list, navigation menu, or set of options where one or two items are significantly more important than the others — and when you cannot reduce the total number of items.
Avoid when
Don't rely on serial position as a substitute for reducing list length — if a navigation has ten items, moving the important ones to the ends doesn't solve the core problem that ten items exceeds working memory capacity regardless of their order.
The middle of a list is a burial ground — if a navigation item is important enough to include, it's important enough to position at the first or last slot, not sandwiched between items that will receive more recall by default.
Real-world examples
- Spotify places the most popular tracks at positions 1 and 2 in algorithmically generated playlists (primacy effect) and uses the final track as a strong closer, leveraging both serial position effects to maximise listen completion.
- Navigation menus across gov.uk, Amazon, and Apple all place the most critical links at the first and last positions, applying serial position effect to maximise the findability of their highest-traffic destinations.
- Nielsen Norman Group eye-tracking studies show that users reading lists of search results have highest recall for items 1–2 and the last item, directly validating the serial position effect in digital information scanning.