Tesler's law on a landing page: complexity never disappears, it changes hands

"We simplified the form, it's just an email now." You hear it on nearly every redesign, and it's half true. The form is shorter, no question. But the budget, the industry, the headcount and the timeline that left the screen didn't stop being necessary: someone will have to collect them later, by phone or by follow-up email. The complexity didn't disappear, it changed owners. That is precisely what Tesler's law describes — and it's one of the most expensive blind spots in landing page optimisation.

Tesler's law, or the conservation of complexity

The law is named after Larry Tesler, an American computer scientist who worked at Xerox PARC in the 1970s — where he contributed to the human-computer interaction work behind cut-copy-paste and modeless interfaces — and later at Apple. It was there, while making the case for a generic application framework to management and to software vendors, that he formulated the law of conservation of complexity: every system contains an irreducible amount of complexity; design cannot eliminate it, only move it. The only real design question becomes: who will carry it — the user or the designer? One clarification the popular versions tend to skip: this is not an experimental result, it's a principle born of practice. Its value isn't predictive, it's diagnostic. It gives you the question to ask of every proposed "simplification": where did the complexity we just removed from the page actually go?

What research does contribute is a measure of what that complexity costs once it lands on the user's side. In a study published in 1988 in the journal Cognitive Science, psychologist John Sweller laid the foundations of cognitive load theory: working memory has a limited capacity, and effort spent on operations unrelated to the task — what he calls extraneous cognitive load — comes directly out of the resources available for the task itself. On a landing page, every decision the visitor makes on your behalf — guessing which category they fall into, reconstructing a price, arbitrating between three poorly differentiated plans — consumes attention that won't go to your offer. Our guide to cognitive load on a landing page covers that mechanism, and our article on Hick's law gives you the stopwatch version.

A shorter form removes nothing, it relocates

The classic case. A B2B landing page asked for seven fields: name, email, phone, company, headcount, budget, timeline. After "simplification," two remain. The submission rate goes up — that's mechanical, and it's the trade-off our article on how many form fields to use quantifies. But nothing that disappeared was decorative: headcount routed the lead to the right rep, budget filtered out off-target enquiries, timeline prioritised callbacks. Those details will now be gathered in a ten-minute first call, multiplied by the number of leads — including the ones that should never have got that far. The complexity was indeed conserved; it simply left the page for the sales team's calendar. That isn't necessarily a bad decision: at low volume and high contract value, a call qualifies better than a dropdown and sells at the same time. The mistake is believing something was removed. A shortened form where nobody downstream picks up what it no longer asks produces the same outcome as a form that's too long: leads that go nowhere, for a less visible reason.

Three savings that aren't savings

The price replaced by "request a quote"

Pulling prices off the page is presented as a simplification: no table to read, a single "Request a quote" button. Yet the pricing complexity hasn't moved an inch — it has been transferred to the visitor in a far costlier form: judging whether your offer is within reach with no reference point at all, handing over their contact details to find out, then waiting. The one who declines that deal doesn't complain, they go to a competitor who publishes prices. Our comparison of visible pricing versus quote-only covers the cases where a quote is still the right call — and the justification always comes from the structure of the offer, never from a gain in simplicity.

The multi-step funnel

Splitting a ten-field form across four short screens works, and our guide to the multi-step form explains why. But be clear about what the split does: it removes no complexity, it spreads it over time. The total amount of information requested is identical; what changes is the effort perceived at the start and the progressive commitment that pushes people to finish. A legitimate shift, but a counterproductive one the moment it's used to hide the real size of the request: a visitor who discovers at step three of six that they'll need a company registration number and three supporting documents walks away feeling misled.

The plan selector and the slot picker

Two components where the transfer is glaring. A selector lining up four SaaS plans, each with fourteen feature rows and a monthly/annual toggle, hands the visitor the modelling work the vendor didn't want to do: guessing which tier matches their usage. The designer can absorb it by asking one question up front ("how many users?"), highlighting the matching plan and pushing the full comparison table further down. Same logic for a meeting: showing an empty three-month calendar asks the visitor to arbitrate on your behalf; offering three concrete slots with a "see other times" link absorbs that arbitration and opens the complexity only to those who need it.

Who should absorb the complexity?

The default answer is simple: the designer, because they pay the cost once, where the user pays it on every visit. Writing the function that infers the country from the IP address takes an hour; a two-hundred-country dropdown costs a few seconds and a marginal drop-off on every single visit, indefinitely. There's a deeper reason not to rely on the user's goodwill: they won't make the extra effort, even when it would benefit them. A study by Wayne Gray and Wai-Tat Fu published in 2004 in Cognitive Science showed that users prefer to rely on approximate information held in memory rather than consult perfect information available on screen, as soon as consulting it demands slightly more perceptual-motor effort — and they do so even when the shortcut raises their error rate. The authors call these "soft constraints": very small differences in cost are enough to steer behaviour, often against what the designer intended. So a visitor won't hunt for your price list inside a PDF and won't compare four plans line by line: they'll take the available shortcut — guess, pick at random, or leave. Any complexity left on their side is unstable; it will be worked around rather than absorbed.

Absorbing it as the designer: what actually works

• Defaults that are chosen, not neutral — the best-selling plan pre-selected, annual billing ticked if that's your recommendation. A good default removes a decision without removing the freedom to make it.
• Pre-filling from the URL — campaign, city, offer or a known email can travel as parameters and drop straight into the form, as our guide to pre-filling a form from the URL details: that's information the visitor doesn't retype when you already held it.
• Automatic detection — time zone, browser language, country, phone number format: all inferable without asking, provided a wrong guess can be corrected.
• Conditional fields — only show the question about number of vehicles to visitors who ticked "company fleet." Our article on the conditional form shows how to stay short for each person while collecting, in aggregate, more qualified information.
• Compute instead of asking — tax-inclusive price from the net one, monthly payment from amount and term, annual saving from the chosen plan. A calculation the page performs is one the visitor doesn't do in their head, or gives up on.
• Defer whatever can wait — billing address, registration number or detailed preferences belong after sign-up, once the user is already committed and the complexity weighs far less.

The through-line never changes: never ask for information you can infer, compute, guess correctly by default or postpone. That is exactly what Tesler's law means applied to conversion — not "keep it simple," which is a wish, but "decide where the complexity goes," which is a trade-off. LanderKit's SaaS waitlist template pushes that trade-off to its limit: one possible action, one visible field, and everything else — traffic source, language, time zone, campaign context — captured without the visitor doing a thing. The 10 LanderKit templates (€89 each, €229 for the full pack) ship as Next.js source code: the complexity already sits on the code's side, where it's only paid for once.