Amara runs a dog-training school on the edge of Leeds. Her homepage opens with a lovely photo of a spaniel and the words “Welcome to Pawsitive”. Then comes the story of how she founded the school in 2014. Then her qualifications. Then a long gallery. Then, near the bottom, a contact form.
Daniel lands on it at 10 p.m. His rescue dog has just eaten a sofa cushion and lunged at a cyclist. He has three questions, in this order: can you fix this, how does it work, and what does it cost?
The page answers a question he did not ask (when did you start?), then one he might ask later (are you qualified?), and never answers the third. He opens the next result.
Nothing on Amara's page is wrong. It is just in the wrong order.
Amara and Daniel are imagined. The research and laws below are real.
The page spine
Every visitor arrives with a queue of questions. The first ones decide whether they stay at all. The later ones decide whether they act. A good landing page answers them in roughly the order they come up.
For most small businesses selling a service or a product, that queue looks remarkably similar. Here it is as a spine: the sections down the page on the left, the question each one answers on the right.
Read the right-hand column alone and it sounds like a real conversation. That is the test. If you swapped two sections, would the conversation still make sense? “What will it cost?” before “Can you actually do it?” is a reasonable question, but most people are not ready to judge a price until they believe the thing works.
The hero gets its own chapter. This one is about everything after it.
Do this: write the visitor's question above each section of your page, in pencil or in a comment. Any section whose question you cannot name is a candidate for cutting or moving.
AIDA is older than its legend
The most famous ordering rule in marketing is AIDA: Attention, Interest, Desire, Action. You will often read that it was invented by the American advertising man E. St. Elmo Lewis in 1898.
The evidence for that date is thinner than its fame suggests. The attribution traces back to the psychologist Edward K. Strong, who wrote in “Theories of Selling” in 1925 that Lewis had used the slogan “attract attention, maintain interest, create desire” in an 1898 course and later added a fourth step about action. Strong did not point to a document from 1898.
In a 2024 paper in the Japan Marketing History Review, Akinori Iwamoto went back to the primary sources. He concluded that there is no evidence Lewis invented the prototype of AIDA in or around 1898, and argued that the salesmanship writers Arthur Frederick Sheldon and Frank H. Dukesmith did more to shape the formula, with Sheldon the stronger claim to being its originator.
Two other formulas show up in every copywriting course. PAS, Problem, Agitate, Solve, opens with the reader's pain, makes it vivid, then presents the fix. Its origin is unclear; it is passed around as practitioner wisdom without a traceable first source. And the objection order, which we will come to next, simply arranges sections around the doubts that stop people buying.
What are these frameworks worth? They are models of a sequence of mental states, drawn from selling experience. They are not experimental findings, and nobody has shown that following AIDA letter by letter beats a well-made page that does not. Their real value is as a reminder: attention is not the goal, it is the entry fee. A page that grabs attention and then stops has done a quarter of the job.
Notice too that the agitation in PAS has an ethical floor. Describing someone's problem accurately is helpful. Inflating their fear to sell the fix is the kind of pressure the honest urgency chapter warns about.
Do this: use AIDA as a checklist, not a template. After your hero, ask whether the next section builds interest in their problem, and whether anything before the ask builds real desire, meaning a picture of life after buying.
Order by objection
The most useful version of section order is also the least glamorous. Write down every reason a reasonable person might not buy. Then make sure each one is answered before the place on the page where it would stop them.
Your best source is not a framework. It is your inbox. The questions people ask in emails, on calls and in messages before buying are the objections your page failed to answer. If three people this month asked “Do you work with reactive dogs?”, that question belongs high on Amara's page, not in the FAQ.
| What the visitor thinks | Where to answer it | How |
|---|---|---|
| “My dog is a special case.” | Problem, right after the hero | Name the common cases: pulling, lunging, chewing, anxiety when left alone. |
| “Classes won't work for a reactive dog.” | Offer | Say plainly that reactive dogs start with one-to-one sessions. |
| “How do I know you're any good?” | Proof | One detailed case of a dog like theirs, then reviews and qualifications. See Showing your work. |
| “It'll take months and I work shifts.” | Process | Show the steps, typical number of sessions, evening slots. |
| “It's probably expensive.” | Price | Show the real price, what it includes, and any package saving. |
| “What if it doesn't work?” | FAQ or next to the price | State your actual policy, honestly. No invented guarantees. |
Notice what this does to the “About us” section. Amara's founding story and qualifications are not deleted. They move into Proof, where they answer a question the visitor is actually asking at that moment: “Can she really do it?”
The biggest objections deserve sections. The small, specific ones belong in the FAQ. If an FAQ entry is asked by most buyers, promote it into the body of the page.
Do this: list the last ten questions prospects asked you before buying. Tick each one your page answers, and note where. Any unticked question is your next section or FAQ entry.
People scroll for a reason
An old piece of advice says everything important must be “above the fold”. It is half right.
In a 2018 eyetracking study, the Nielsen Norman Group analyzed more than 130,000 eye fixations from 120 participants on a 1920 × 1080 screen. Therese Fessenden reported that people spent 57% of their page-viewing time above the fold, down from 80% in NN/g's 2010 study. The first two screenfuls took 74% of viewing time, and the first three took 81%.
Two lessons sit in those numbers. Attention still falls off sharply as people scroll, so the order of sections matters: what comes first gets the most looking. But a lot of attention now goes below the first screen, so a page does not have to cram everything into it. Fessenden's own summary is that people will scroll when they have a reason to.
The reason is the spine. Each section should end in a way that makes the next one feel necessary. A problem section that names Daniel's sofa cushion makes him want to see the offer. A proof section that shows a dog like his makes him want to see the price.
NN/g also warns about “false floors”: a band of whitespace or a full-width image that looks like the end of the page, so people stop. Let a heading or a bit of the next section peek above the bottom of the screen.
Want to know if your first screen earns the scroll? Run the three-second test on it.
Do this: open your page on a laptop and on a phone. At every screen boundary, check that something visibly continues below. If a section ends in a full stop and empty space, add a bridge.
One job per section
Why not put the price, the process and three testimonials into one tidy block? Because each extra thing a visitor must hold in mind competes for the same limited capacity.
Working memory is limited. In his 1988 paper, Sweller argued, with a computational model and experiments, that conventional problem solving uses so much processing capacity that little is left for learning the underlying patterns (schemas). Later work turned this into design advice for instruction: remove load that does not serve the goal.
Sweller's study was about learning to solve problems, not about web pages, so treat this as an analogy. But the analogy fits a common landing page failure: asking a visitor to solve a problem while reading. A pricing block that requires comparing four tiers, two add-ons and an annual discount, placed in the middle of a testimonial carousel, forces exactly that.
The practical translation is simple. Give each section one job and one question. Let its heading state the answer, so a scanning visitor can follow the argument from headings alone. Keep unrelated decisions in separate places, and do the hard comparisons for the visitor where you can. More on choice sets is in Pricing and anchoring.
One more trap: sections that look like advertising. Brightly boxed promotions and banner-shaped blocks are often skipped even when they hold useful information. The banner blindness tip explains why, and how to make important sections look like content.
Do this: read only the headings of your page, top to bottom. If they do not tell the story on their own, rewrite them as answers (“Most dogs need four to six sessions”) rather than labels (“Our process”).
Endings and the last ask
People who reach the bottom of a landing page are the most interested people you will get. Too many pages greet them with a footer and nothing else.
When people try to recall a list, they remember items from the start (primacy) and the end (recency) better than items from the middle. Murdock's experiments used lists of 10 to 40 common words, presented at one or two seconds per word, and found both effects at every list length.
A web page is not a word list, and nobody is asked to recall your sections in order. So do not read Murdock's curve as proof that the last section “converts better”. The cautious lesson is that beginnings and endings are natural moments of attention. At the end of a page, the visitor has just finished evaluating you. That is exactly when they need the next step in front of them, without scrolling back up to the hero to find it.
So repeat the main call to action at the end, using the same wording as at the top. Same action, same label. A different label at the bottom (“Get started” up top, “Contact us” below) makes people wonder whether these are two different things. The Calls to action chapter covers wording.
You may also have read that pages should leave “open loops” because of the Zeigarnik effect, the idea that we remember unfinished tasks better than finished ones. Be careful with that one.
Working in Kurt Lewin's Berlin lab, Zeigarnik reported that people recalled interrupted tasks better than completed ones. A 2025 meta-analysis by Ghibellini and Meier found no general memory advantage for unfinished tasks. It did find a reliable tendency to resume interrupted tasks, the related Ovsiankina effect.
The meta-analysis suggests the original memory effect may have depended on conditions that were common in the 1920s lab, such as the experimenter's authority. So “leave them hanging and they'll remember you” rests on weak ground. The resumption finding is more useful: people tend to want to finish what they have started. A short, visible process (“three steps, you are on step one”) respects that tendency honestly. A teaser headline that withholds the answer to get a scroll mostly annoys.
Do this: end your page with a short closing section: one sentence that restates the outcome, the same call to action as the hero, and a line on what happens after clicking (“We reply within one working day”).
Long page or short page?
There is no correct length. There is a correct amount of answering, and it depends on how much the visitor already knows and how much is at stake.
Copywriters often use a model from Eugene Schwartz's 1966 book Breakthrough Advertising. Schwartz described prospects at different stages of awareness, from people who already know and want your product and just need the offer, to people who do not yet recognize their problem at all. In his framing, the stage decides how much work the headline and copy have to do. It is a practitioner heuristic from a famous copywriter, not a tested finding, but it explains a lot of the long-versus-short argument.
Look at how the spine behaves across the grid. On a short page, the problem section might be one line and proof a single star rating. On a long page, each section might be a full screen. The order barely changes. What changes is how much each question needs answering.
A practical warning: long is not the same as padded. If you add sections that answer nobody's question, you make the useful ones harder to find. Length should come from real objections, not from a belief that longer pages sell more.
Do this: place your offer on the grid. If you sell something expensive and unfamiliar on a page with three short sections, you probably need more proof and process. If you sell something simple to people who already want it, cut to the price and the button sooner.
What the law wants visible
For businesses selling to consumers, some answers on the spine are not optional, and some have to appear at a specific moment. These rules matter most for pages where people can buy or book directly.
Before a consumer is bound by an online contract, the trader must give clear information including the main characteristics, the trader's identity and contact details, and the total price inclusive of taxes, plus any delivery or other charges (Art. 6(1)(a)–(e)). Directly before the consumer places an order that obliges them to pay, the main characteristics, total price, contract duration and any minimum duration must be shown clearly and prominently, and the button must say “order with obligation to pay” or something equally unambiguous (Art. 8(2)). If not, the consumer is not bound. Delivery restrictions and accepted payment methods must be shown at the latest at the start of the ordering process (Art. 8(3)).
Source: Directive 2011/83/EU, consolidated text. National laws implement it, for example Germany's “button solution” in BGB §312j. General information, not legal advice.
Even if nobody buys on your page, an advert that shows a product and its price can count as an “invitation to purchase” under the Unfair Commercial Practices Directive. Then the price inclusive of taxes, the main characteristics, your identity and address, and any right of withdrawal count as material information (Art. 7(4)). Leaving them out, or hiding them “in an unclear, unintelligible, ambiguous or untimely manner” (Art. 7(2)), can be a misleading omission.
For goods offered to consumers, the selling price means the final price including VAT and all other taxes (Art. 2(a)), and it must be “unambiguous, easily identifiable and clearly legible” (Art. 4(1)). Since 28 May 2022, any announced price reduction must show the prior price, meaning the lowest price the trader applied in at least the 30 days before the reduction (Art. 6a). Member States may exempt products supplied in the course of providing a service (Art. 3(2)); for services, the UCPD's price rules above still apply.
Source: Directive 98/6/EC, consolidated text. General information, not legal advice.
For the spine, this has a clear consequence. A price section that says “from €49” when almost everyone pays €89 plus a booking fee is not just weak copy, it may be misleading. And a price that appears for the first time at the final step is a problem the law is specifically designed to catch.
The FAQ is advertising too
FAQ answers are commercial claims like any other part of the page. “Can I cancel anytime?” answered with “Of course!” must match your actual terms. An accordion is a fine place for detail, but essential cost or contract information buried in a collapsed answer is the kind of untimely presentation Art. 7(2) of the UCPD describes.
A note on search. For years, sites marked up FAQ sections with structured data to get expandable answers in Google results. Google limited those FAQ rich results to well-known government and health sites in 2023, and its Search Central changelog records that the feature stopped appearing in Google Search from 7 May 2026. Write your FAQ for the people on the page, not for a snippet.
Do this: check that your price section shows the full price including taxes and fees, and that no FAQ answer promises more than your terms allow.
Reorder your page tomorrow
Try it before you touch your own page. The sections below are in the order many small-business pages use. Move them until the conversation makes sense.
- Heroheadline, outcome, main button
- Pricetotal price, what's included
- Proofone case study, reviews
- Final askrepeat the main button
- FAQremaining worries
- Problemtheir situation, their words
- Processsteps and timing
- Offerwhat they get
With JavaScript on, each row gets up and down buttons and the checks below update as you move sections.
When you are done, apply the same thinking to your real page.
Label every section with its question
Write the visitor's question above each block. Cut or merge sections whose question you cannot name.
Move proof above price
Belief first, cost second. If your testimonials sit below the price, swap them.
Turn the “About” block into proof
Your story and qualifications answer “Can they do it?”. Put them where that question comes up.
Mine your inbox for objections
The last ten pre-sale questions are your FAQ, and the most common ones deserve sections.
Rewrite headings as answers
A scanning visitor should get the argument from headings alone.
Bridge every screen boundary
No false floors. Let the next section show above the bottom edge.
End with the same ask
Repeat the hero's call to action, word for word, plus what happens next.
Check the legal must-haves
Full price including taxes and fees, honest reductions, clear contract terms before the order button.
Amara's page did not need new content. It needed the dog problem near the top, her qualifications next to a story about a reactive spaniel, her prices in plain sight, and the booking button at the end where Daniel, finally convinced, would be looking for it.
What the research supports
NN/g's eyetracking data show that attention concentrates near the top but that people do scroll when there is a reason. AIDA, PAS, the objection order and Schwartz's awareness stages are practitioner models with long histories, not experimentally tested page structures. The serial position effect and cognitive load theory are well established in memory and learning research; applying them to section order is reasonable analogy, not direct evidence. The famous Zeigarnik memory effect did not hold up in a 2025 meta-analysis, though the tendency to resume unfinished tasks did. No study shows one universal section order that wins for every business.
Sources
- Strong, E. K., Jr. (1925). Theories of selling. Journal of Applied Psychology, 9(1), 75–86. Bibliographic record; the Lewis attribution is described via Iwamoto and secondary accounts.
- Iwamoto, A. (2024). The origin of AIDA: Who invented and formulated the AIDA model? Japan Marketing History Review, 3(2), 150–166. Abstract consulted for the conclusions on Lewis, Sheldon and Dukesmith.
- Fessenden, T. (2018), Nielsen Norman Group. Scrolling and Attention. Eyetracking method, viewing-time shares and the false-floor recommendation.
- Sweller, J. (1988). Cognitive Load During Problem Solving: Effects on Learning. Cognitive Science, 12(2), 257–285. Abstract consulted.
- Murdock, B. B., Jr. (1962). The serial position effect of free recall. Journal of Experimental Psychology, 64(5), 482–488. Design (list lengths and rates) checked against a published reanalysis of the original dataset.
- Zeigarnik, B. (1927). Das Behalten erledigter und unerledigter Handlungen. In K. Lewin (Ed.), Untersuchungen zur Handlungs- und Affektpsychologie III. Psychologische Forschung, 9, 1–85. Bibliographic record.
- Ghibellini, R. & Meier, B. (2025). Interruption, recall and resumption: a meta-analysis of the Zeigarnik and Ovsiankina effects. Humanities and Social Sciences Communications, 12, 962. Abstract consulted.
- Schwartz, E. M. (1966). Breakthrough Advertising. Described from published summaries; treated as a practitioner heuristic.
- Directive 2011/83/EU (Consumer Rights Directive), consolidated text. Articles 6(1) and 8(1)–(4) consulted.
- Directive 2005/29/EC (Unfair Commercial Practices Directive), consolidated text. Articles 6 and 7(1), 7(2), 7(4) consulted.
- Directive 98/6/EC (Price Indication Directive), consolidated text. Articles 1–4 and 6a consulted.
- Google Search Central. Latest Google Search Documentation Updates. Entries of 8 May 2026 (FAQ rich result deprecated from 7 May 2026) and 15 June 2026 (documentation removed); the 2023 restriction to government and health sites as recorded there.
Sources checked 18 October 2026. Legal summaries are general information, not legal advice; check the rules in your own country, and whether you sell to consumers or businesses. Amara, Daniel and the dog-training school are fictional; the figures are original illustrations, with Figure 3 drawing on NN/g's published percentages.