What goes first, what goes last: serial position and page architecture

Hermann Ebbinghaus published his memory curve research in 1885, and the core finding—the serial position effect—has been replicated so many times it is practically furniture in cognitive psychology. Items at the start of a sequence (primacy) and the end (recency) are recalled significantly better than those in the middle. The primacy effect appears because early items receive more rehearsal and deeper encoding into long-term memory. The recency effect appears because late items are still in working memory when recall is tested.

What strikes me is how often this is ignored in page architecture. A typical long-scroll homepage runs: hero → about → features → features → features → social proof → pricing → FAQ → contact. The features in positions three through five are the graveyard. Not because the content is bad, but because memory does not treat the middle kindly.

How I structure pages differently now

• The first full viewport (hero) carries one message and one action. Primacy demands it be the thing I most want the visitor to remember even if they bounce.
• The second section earns credibility—who I am, what I have done, or a brief proof point. Still in the primacy zone.
• Middle sections get internal structure: clear headings, visual variety, interactive elements. If memory is weakest here, each section needs to re-earn attention independently rather than relying on sequence momentum.
• The final section before the footer is a strong, singular call to action. Recency gives this outsized influence on what the visitor carries away. I never waste this slot on an FAQ.

Navigation and the edges of lists

The same effect applies to navigation menus. In a horizontal nav, the first and last items are recalled most reliably. That is why "Home" traditionally sits at the far left and the primary conversion link ("Sign up," "Contact") sits at the far right. It is not a style convention—it is a memory convention. Murdock's research (1962) quantified this: in free recall experiments, the probability of recalling an item plotted against its serial position consistently produced a U-shaped curve, with the middle items recalled at roughly half the rate of items at either end.

Put your weakest content in the middle or take it out. Never put your strongest argument third out of five—it will be forgotten first.

Prospect theory and how you frame each position

Serial position tells you where to place things. Kahneman and Tversky's prospect theory (1979) tells you how to frame them. Their central finding—that losses loom roughly twice as large as equivalent gains—has direct implications for section copy. A hero that says "Stop losing customers to slow load times" lands harder than "Enjoy faster load times" even though the information content is identical. The framing effect means that how you present the same fact changes how people feel about it, and feeling drives action at the visceral level Norman describes.

In practice, I use loss-framed language sparingly and honestly—typically at the top of a page where attention is high and stakes are being established. Mid-page, I shift to gain-framed language: what you get, what improves, what becomes possible. By the final CTA I return to a gentle loss frame: "Do not leave this unresolved." That sequence—loss, gain, loss—mirrors the U-shape of the serial position curve itself: strong start, lighter middle, strong close.

Implementation intentions at the end

Peter Gollwitzer's research on implementation intentions (1999) shows that people are dramatically more likely to follow through on a goal if they form a specific if-then plan: "If it is Monday morning, I will email the developer." His meta-analysis found a medium-to-large effect size (d = 0.65) for if-then planning across nearly a hundred studies. The product implication: your closing CTA should not just say "Get in touch." It should make the next step concrete and temporally grounded: "Book a 20-minute call this week" or "Send us your brief and we will reply within a day." You are helping the visitor form an implementation intention at the moment when recency gives it the best chance of sticking.