
The Scenario: A B2B Lead-Gen Site With Beautiful Forms and Invisible Content
Picture a Johannesburg B2B services firm that has just rebuilt its site. The design is genuinely better: a three-step qualification form in a slide-over panel, case studies behind tabs, pricing copy in an accordion, and a client-side router that makes navigation feel instant. Sales likes the form. Marketing likes the way it looks.
Then organic enquiries flatten. Nothing is broken; the copy still exists on screen. But much of what was written to rank - the qualifying detail in tab panes, the long-form case study narrative, the pricing table - no longer exists in the HTML a crawler receives first. This is the seam we work in: the quiet trade a UX decision makes against indexability, one of the many web design decisions that leak rankings and revenue. That same seam runs through Shopify and WooCommerce filter states, SaaS documentation portals, and any interface built around JavaScript.
Why UX and SEO Are Reviewed Together, Not Sequentially
Most teams review design in isolation, then hand it to an SEO for sign-off. By then, interaction patterns are baked into the component library. The cost is structural: a modal holding your best lead-generation copy, a tab set hiding specification detail, a consent banner blocking the very content a crawler needs.
Reviewing both audiences in one pass is cheaper than retrofitting. A component that fails a screen reader usually also fails a crawler - and a fix for one frequently resolves both.
The Two-Audience Design Review: A Component Scorecard
Before code, score each component twice. It takes an afternoon and heads off most of the problems below.
| Component | For the person | For the crawler |
|---|---|---|
| Lead form | Clear labels, error recovery, mobile keyboard | Accessible in initial HTML; required fields marked |
| Tab set | Predictable focus order, keyboard operable | All panes in rendered HTML, not only active tab |
| Accordion | Progressive disclosure, readable headings | Full copy present regardless of collapsed state |
| Modal panel | Focus trap, dismiss easily | Key copy duplicated in crawlable page body |
| Consent banner | Fair, equal-choice options | Does not obstruct content; analytics honour consent |
What the Person Needs From a Component
Predictability, legible labels, keyboard and screen-reader operability, sensible defaults, and no surprise - especially on slow connections where an instant-feeling router can stutter.
What the Crawler Needs From the Same Component
Semantic HTML, descriptive link text, and the copy genuinely present in the rendered document - not only after an interaction the bot will never perform.
The Rendering Question: What JavaScript-Heavy UX Removes From the HTML
Google renders JavaScript, but rendering is deferred and sequenced: crawl, then queue for rendering, then index. Content and links must appear in the rendered HTML to count. Client-side routing is a common leak that undermines your website structure SEO - navigation built as click handlers rather than anchors can leave links uncrawlable.
Treat server-rendered content as the baseline. If key copy and links only exist after user interaction, assume a machine may never see them. Seen properly, this is simply part of SEO and web design working as one discipline.
The Interaction Pattern Table: SEO Cost Versus UX Benefit
Not every pattern is a risk. Score each against the cost of lost visibility.
| Pattern | UX benefit | SEO cost | Mitigation |
|---|---|---|---|
| Tabs hiding long copy | Tidier page, faster scan | Low if HTML present | Headings inside each pane |
| Infinite scroll | Fluid browsing | High - paginated URLs vanish | Preserve crawlable pagination |
| Client-side router | Instant feel | High - links not anchors | Server routes and real anchors |
| Modal lead form | Reduced friction | Medium - copy invisible | Duplicate key copy on page |
Accessibility as SEO: Why WCAG Work Is Crawl Work
The overlap between WCAG 2.2 and crawlability is well documented, if not absolute. Semantic structure, labels, and landmarks serve both. When accessibility is descoped late, indexability usually slips with it. Budget for the overlap early rather than treating accessibility as remediation.
Semantic HTML, Landmarks and Link Text as Shared Deliverables
One heading hierarchy, one set of landmarks, descriptive link text. 'Read the case study' beats 'click here' for both audiences.
The South African Context: Device Mix, Data Cost and Inclusive Design
Mixed device quality and data costs make progressive enhancement commercially sensible. A lean, server-rendered page is kinder to users on constrained connections and easier for Google's mobile-first indexing to evaluate. Design for the median device, not the newest handset in the room.
Design System to Developer Handoff: Tokens That Carry Meaning
Encode meaning in the handoff: heading levels mapped to tokens, required field labels, link-text guidance. If a component token carries no semantic instruction, a developer will improvise - and improvisation is where indexability leaks.
component: text.link
text: descriptive, never 'here'
element: not JavaScript click
note: crawler must follow itPage Experience Signals, Briefly: INP, CLS and Layout Decisions
Core Web Vitals (LCP, INP, CLS; INP replaced FID in 2024) are a ranking input, not a standalone system. Stability matters: late-loading banners and shifting modals hurt users and measurements. We cover thresholds in our page-speed work; here, keep layout decisions stable before launch.
How to Measure Whether UX Changes Helped or Hurt
Measurement is a process, not a dashboard screenshot - consistent with Prebo Digital's custom reporting approach and Strategy → Build → Test → Scale → Report model.
Comparing Rendered DOM With Crawlable HTML
Pull both versions of key templates. Google's URL Inspection 'Tested page' rendering shows what Googlebot executed; compare against your source HTML for missing copy and links. Then watch Search Console Core Web Vitals and Enhancements reports for movement.
Funnel Fidelity and Reporting Discipline
Validate GA4 with consent-mode segments, reconcile platform-reported conversions with server-side data, and review drop-off per step. Run VWO tests deliberately - one interaction change at a time - so you can attribute the outcome. For commerce teams, hold the line on profitability and MER; for SaaS, watch CAC and LTV rather than raw traffic.




