Marketing pages and product UI can run different constraint intensities

What it is

Marketing pages live by campaign: few of them, short-lived, judged on whether they are remembered and clicked. Product UI lives by task: many screens, years of life, judged on completion. One constraint intensity for both makes marketing look like a settings panel, or lets campaign layout hollow out the product’s ability to be repaired. Split constraint intensity lets marketing run loose and product run tight, provided the product side still sits on the same token and component spine. It is not “spend recognition on the splash”; it is admitting the two surfaces have different score functions, so the constraints need not be one set.

Loose is not spineless: shared semantic aliases, type metrics, and accessibility floors still bind both. What loosens is layout skeleton and campaign composition.

Why it happens

Constraint intensity is how far a surface may drift from defaults, and which approvals it must pass. A drift in product UI has a regression surface of every task flow; a drift on a marketing page has a regression surface of that campaign’s few URLs. Put product-grade approval on a campaign page and the campaign misses its slot, or forks uncontrollably in a product repo. Put campaign-grade freedom on settings and forms and the product accumulates local layout that will not follow a theme. The split writes “allowed drift radius” as two numbers: large on marketing, small on product. The spine (semantic color, type steps, focus behavior) is unbreakable in both, so a contrast incident is not repaired by two different methods. Intensity splits; the floor does not.

Where it stops holding

The in-product acquisition screen (sign-up, paywall) is scored as both campaign and task; it does not automatically get the loose marketing tier. It is closer to product, because users hit it again and failures become support tickets. Email, landing pages, and store art are marketing-tier even when they look like product shots. A “campaign banner” embedded on a task path is short-lived but in the way of work: it should use a product-component slot, not paste a whole campaign layout into the page. A small team may run only the product tier; when marketing is outsourced, the floor (contrast, minimum type, mark exclusion) must be in the delivery list, or the agency will follow its own campaign habits and the work will not join the product domain. Legally bound pages (prospectus, privacy) look like marketing long-reads and should still run at product intensity: a wrong word does not cost “not cool enough”.

Applying it

• Tag every URL or in-app route as marketing-tier or product-tier. The tag decides: may it leave the shared grid, may it introduce one-shot illustration composition, does change ride the product release window.
• Both tiers share semantic aliases and the accessibility floor. Marketing-tier may add one-shot layout. Product-tier may not import those one-shots as “atmosphere”.
• Reuse across tiers is an explicit port: a composition proven on marketing enters product only after it is broken into components and tokens, not as a page copy.
• Check: take a campaign page and a settings page. The campaign page should achieve a clearly different composition without rewriting product anatomy; the settings page should be unable to reference the campaign’s one-shot layout classes. Run contrast on both: the same floor. A marketing page that passes visual review and fails contrast is not a split — it is marketing exempted from inspection.

Related

• Same group: R1.07.1 Systems chase reuse; brands chase difference · R1.07.2 Confine difference to a few expression layers · R1.07.3 Customizing everything forfeits the system's return · R1.07.4 Brand recognition is cheaper on a few high-frequency touchpoints · R1.07.6 Rebrand cost depends on whether difference stayed inside the system
• Adjacent: R1.15 Multi-brand and multi-theme · R1.02 Component libraries and variants
• Search terms: split constraint intensity · marketing versus product · constraint intensity