Jakob's Law: why an "original" landing page converts less

There's a category of landing pages that fail for a counterintuitive reason: they're too original. A vertical left-hand menu, a logo centered at the bottom, a buy button that doesn't look like a button, a form that slides in from an unexpected panel — every decision is defensible on its own, and together they sink the conversion rate. The principle behind that failure has a name: Jakob's Law, coined by Jakob Nielsen, co-founder of the Nielsen Norman Group. This guide explains exactly what it says, what research actually measures behind the intuition, which conventions French-speaking visitors expect in 2026, and where originality remains a profitable investment instead.

Jakob's Law, in one sentence

The statement fits on one line: users spend most of their time on sites other than yours, so they expect yours to work like all the others. That's not an aesthetic principle, it's an observation about how time is distributed. A visitor may spend forty seconds on your page; they spent the previous twenty years on thousands of other pages. That massive exposure has built a default mental model — where the logo sits, what a link looks like, what a cart icon does — and your forty-second page has no chance of reprogramming it. It can only conform to it, or contradict it and pay the price.

The price, precisely, is learning time. A page that puts its menu where people expect it is free to understand: the visitor reuses a skill they already have. A page that invents its own navigation logic asks the visitor to build a new model before they've even understood the offer. On a site you return to, that effort pays off across dozens of visits. On a landing page there's only one visit — and the learning cost is never amortized. It's exactly the mechanism described in our article on cognitive load on a landing page: whatever the visitor spends decoding the interface, they don't spend evaluating the offer.

What research measures behind the intuition

Jakob's Law is a practitioner's heuristic, but it rests on solid experimental results. The most direct concerns prototypicality — how closely a page resembles what you'd expect from a page in its category. A study by Alexandre Tuch, Eva Presslaber, Markus Stöcklin, Klaus Opwis and Javier Bargas-Avila published in 2012 in the International Journal of Human-Computer Studies had participants rate screenshots of real websites varying in visual complexity and prototypicality, at ever-shorter exposure times. The result: both factors shape aesthetic judgment within 50 milliseconds, and the effect still holds at 17 milliseconds. In other words, the visitor decides whether your page "looks like a real page" before reading a single word — and a low-prototypicality page starts with a handicap acquired before any reading.

That timing is nothing exceptional: a study by Gitte Lindgaard, Gary Fernandes, Cathy Dudek and Judith Brown published in 2006 in Behaviour & Information Technology had already shown, across three successive experiments, that the visual appeal of a homepage is judged consistently within 50 milliseconds. The practical consequence is brutal: your page has no time to explain its singularity. It gets filed as "familiar" or "weird" before your value proposition has been read — something our 5-second test confirms from another angle.

That leaves the question of where. That's the subject of a study by Sandra Roth, Peter Schmutz, Stefan Pauwels, Javier Bargas-Avila and Klaus Opwis published in 2010 in Interacting with Computers, which asked 516 participants to place the most frequent interface objects (logo, navigation, search, main content, login) on blank grids, for three page types: online shops, news portals and company websites. The authors found strikingly convergent mental models: logo top-left, navigation at the top, main content in the center — with some secondary links having migrated to a rich footer. These expectations aren't individual tastes; they're positions shared by the vast majority of users. An element placed elsewhere isn't "bold", it's just searched for longer.

The conventions a French-speaking visitor expects in 2026

Conventions are neither universal nor eternal: they depend on market, language and era. Here are the ones a French-speaking visitor applies by default today, and what most often breaks them.

Two notes on that table. First, the right-hand column doesn't describe lapses of taste but search costs: each row adds a fraction of a second and a micro-frustration. Second, the most underrated row is payment methods: redrawing the Visa or CB logos to match your brand palette destroys the very thing that made them useful, namely instant recognition — a point covered in our article on trust badges and secure-payment signals. For button labels, our roundup of CTA examples that convert offers proven wordings rather than invented ones.

Where originality still pays

Jakob's Law is not a license for blandness, and that's the most common misreading. It applies to the page's mechanics, not its content. The useful rule: be conventional where the visitor acts, distinctive where they evaluate you.

• Brand visuals — typography, palette, illustration, photography: nothing forces you to look like your competitor, and everything pushes you not to. Visual recognition is an asset; interface mechanics are not.
• The hero message — this is where originality pays the most, because it's the only place the visitor is actively looking for a difference. A promise phrased like everyone else's gives no reason to stay; see our anatomy of a landing page that converts.
• Proof — numbers, customer cases, a recorded demo, before/after: this is your own material, impossible to copy. The format can stay classic, the substance has to be irreplaceable.
• The demo format — an interactive simulation, a calculator, a configurator: an originality of content that does pay off, as long as it opens and closes like any other component.
• Editorial tone — the way you write an objection, a guarantee or an FAQ has no reason to be standardized; at equal offers, that's often where perceived difference is decided.

Useful convention or slavish copying: where the line falls

"Doing what everyone does" and "copying your competitor" are two different things, and conflating them produces identity-free pages that don't convert any better for it. A convention is a behavior shared across a whole ecosystem, learned elsewhere and transferable; slavish copying reproduces one particular competitor's particular choices — their section structure, their colors, their arguments. The first reuses something the visitor already owns, the second only reuses someone else's work, usually without knowing its reasons or its results. Three questions settle it.

1. Did the visitor learn it somewhere other than at this competitor? If the element shows up on dozens of sites across different sectors (logo position, burger icon, cart top-right), it's a convention. If it only shows up at this competitor, it's their style, not an expectation.
2. Would breaking it cost time, or only resemblance? Moving the menu costs search time. Choosing a different palette costs nothing — except to those hoping to borrow someone else's recognition.
3. Could I explain why this choice exists? A convention has a functional reason you can state out loud. A copy is usually justified by "that's how they do it", which isn't a reason but an abdication.

That line has direct consequences for structural decisions. Removing the navigation from a landing page, for instance, doesn't violate Jakob's Law: it's a convention of the landing-page format in its own right, one visitors encounter everywhere — our guide to the navigation menu on a landing page explains why. Likewise, respecting the Gestalt principles or the reading order described in our article on Z-pattern vs. F-pattern reading isn't imitation: these are perceptual regularities that predate any competitor.

The most expensive mistakes

• The button that doesn't look like a button — thin outline, transparent fill, text the same size as body copy: the visitor doesn't read it as clickable, and no A/B test rescues that.
• Hijacked scrolling — forced horizontal scroll, scrolljacking, sections that swap places: this is the most deeply ingrained convention there is, and therefore the costliest to break.
• Invented icons — a house-style pictogram for the cart, the account or the menu forces guessing, where the standard icon is read effortlessly.
• A CTA pushed below the fold with no cue — not because a CTA must always sit at the top, but because a page with no visible action point in the first screen contradicts expectation; see our article on the fold.
• An unreadable "signature" typeface — a display font across full paragraphs adds a decoding cost to every line; our guide to landing page readability quantifies the trade-off.
• Stacking novelties all at once — one isolated break with convention is tolerated; five at the same time turn the page into a riddle, and compound with the decision cost described by Hick's Law.

Getting the page live this week

The fastest way to apply Jakob's Law is to not reinvent the skeleton and to concentrate your effort on message and proof. LanderKit templates start from exactly that premise: conventional structures, components that behave as expected, and all the room for originality left to the content. The coach & consultant template follows the booking pattern visitors already know, form to the right of the hero included; the single-product e-commerce template reuses the expected commerce cues — tax-inclusive prices, recognizable payment logos, guarantees next to the button. The 10 LanderKit templates (€89 each, €229 for the full pack) ship as Next.js source code: the mechanics already match expectations, so all that's left is to put in what nobody else can write for you.