uxantipatterns

Scrolljacking

High severityscrollinteractionaccessibilitymotion

Scrolljacking happens when a page overrides native scroll semantics, pacing, direction, stopping points, or reachability so scrolling no longer behaves like the user's browser, device, or assistive technology normally promises.

It removes user control, raises cognitive load, can trigger motion sensitivity, and turns ordinary navigation into an accessibility responsibility that many custom scroll implementations do not meet.

Users lose their place, miss content, fight the scrollbar, refresh, back out, or abandon the task when the page feels stuck, slow, or broken.

Majorityof NN/g study participants were at least mildly disoriented by scrolljacking

Keep native scrolling as the baseline, layer optional visual enhancement on top, preserve landmarks and return points, and offer explicit continuation controls where endless loading would otherwise take over.

Examples

Clear match7 examples

The anti-pattern is the central failure: the page blocks progress before the user's expectation is satisfied.

Crypto.comClear match
Screenshot placeholderEvidence capture planned
SurfaceFull-page marketing narrative
ExpectationScroll through a page and retain normal control over pace and position.
Premature askScroll input advances a long locked experience instead of ordinary document flow.
EvidenceScrolljacking research report; NN/g case
CaveatLive implementation may have changed since the documented 2023 case.
RelatedMotion overload · Forced narrative · Keyboard trap risk
ReachdeskClear match
Screenshot placeholderEvidence capture planned
SurfaceOfferings page
ExpectationRead and compare offering information at a self-controlled pace.
Premature askDirection and rate are remapped during scrolling.
EvidenceScrolljacking research report; NN/g case
CaveatLive implementation may have changed since the documented 2023 case.
RelatedMotion overload · Cognitive load · Forced narrative
UT AustinClear match
Screenshot placeholderEvidence capture planned
SurfaceUniversity information page
ExpectationFind transfer-credit information quickly.
Premature askWork through an interruptive scroll sequence before continuing the task.
EvidenceScrolljacking research report; NN/g case
CaveatURL points to the likely topical area; the documented capture may not match the current live page.
RelatedTask interruption · Information scent break · Backtracking
SephoraClear match
Screenshot placeholderEvidence capture planned
SurfaceMobile product list
ExpectationBrowse a bounded product list with understandable progress and stopping points.
Premature askContinue through an automatically expanding list with weak pacing and boundary cues.
EvidenceScrolljacking research report; Baymard case
CaveatInfinite list behavior varies by category, device, and current implementation.
RelatedInfinite-scroll trap · Option overload · Refinding failure
NikeClear match
Screenshot placeholderEvidence capture planned
SurfaceProduct listing page
ExpectationBrowse products while retaining access to footer, policy, and support links.
Premature askKeep loading products instead of reaching a stable page end.
EvidenceScrolljacking research report; NN/g infinite-scroll case
CaveatFooter and listing behavior can vary by region, category, and session state.
RelatedInfinite-scroll trap · Footer lockout · Support discovery failure
BMW USAClear match
Screenshot placeholderEvidence capture planned
SurfaceVehicle listing page
ExpectationReturn from a detail page to the same list position.
Premature askRe-find a previous place after the list loses position or context.
EvidenceScrolljacking research report; NN/g infinite-scroll case
CaveatLive implementation may preserve position differently than the documented 2022 case.
RelatedInfinite-scroll trap · Refinding failure · Backtracking
Miami FCClear match
Screenshot placeholderEvidence capture planned
SurfaceContent listing page
ExpectationUnderstand whether the page has ended.
Premature askInterpret a false ending while additional content is still hidden below the fold.
EvidenceScrolljacking research report; NN/g infinite-scroll case
CaveatLive page structure may differ from the documented 2022 case.
RelatedIllusion of completeness · Infinite-scroll trap · Content discovery failure
Partial match3 examples

The pattern is present but mixed with mitigating factors such as partial preview, delayed walls, or product constraints.

Apple product pagePartial match
Screenshot placeholderEvidence capture planned
SurfaceProduct marketing page
ExpectationInspect product information and move forward or back freely.
Premature askScrub through animated scenes before the document progresses normally.
EvidenceScrolljacking research report; NN/g case
CaveatBrand storytelling may add value when short, but the interaction still adds orientation and recovery cost.
RelatedParallax abuse · Motion overload · Forced narrative
TherabodyPartial match
Screenshot placeholderEvidence capture planned
SurfaceProduct marketing page
ExpectationEvaluate product value without losing normal scroll control.
Premature askFollow a visual scroll sequence that alters movement and pacing.
EvidenceScrolljacking research report; NN/g case
CaveatThe effect may be less frustrating than text-heavy hijacks, but it still risks surprise for goal-focused users.
RelatedParallax abuse · Motion overload · Forced narrative
TikTok feedPartial match
Screenshot placeholderEvidence capture planned
SurfaceApp-level feed
ExpectationBrowse a feed while retaining self-control and stopping cues.
Premature askContinue through an endless feed without strong boundaries or stopping points.
EvidenceScrolljacking research report; European Commission and Reuters reporting
CaveatThis is an app-level infinite-scroll case rather than a classic web page scrolljack.
RelatedInfinite-scroll trap · Compulsive use · Boundaryless feed
Boundary case1 examples

The surface resembles the pattern, but context may change the diagnosis or mark a safer edge.

BBC News explainerBoundary case
Screenshot placeholderEvidence capture planned
SurfaceEditorial explainer
ExpectationRead an article while a visual explanation updates.
Premature askUse scrolling to drive a short visual reveal.
EvidenceScrolljacking research report; NN/g case
CaveatThis marks the safer edge: brief, functional scroll-driven explanation can be acceptable when native movement and escape remain clear.
RelatedScrollytelling · Progressive disclosure · Motion overload

Research

Native scroll override harms control and task successNielsen Norman Group, Oslo Metropolitan University

NN/g defines scrolljacking as changing scroll speed and sometimes direction, and reports threats to user control, discoverability, efficiency, and task success. A controlled Oslo Metropolitan University study found lower accuracy and lower satisfaction for a scrolljacking prototype than for an otherwise equivalent native-scroll version.

Scrolljacking increases disorientation and abandonmentNielsen Norman Group case studies

The report summarizes cases where users became disoriented, refreshed, dragged the scrollbar, retraced steps, clicked away, or left the page when long, imprecise, direction-changing, or text-heavy scrolljacking interrupted their task.

Infinite scroll creates landmark, footer, and refinding failuresNielsen Norman Group, Baymard Institute

Infinite lists become harmful when users need boundaries, footer access, comparison, or return-to-place behavior. The report cites Nike footer lockout, BMW list-position loss, Miami FC false endings, and a Sephora mobile list exceeding 527 items.

Motion and keyboard accessibility risks compound the patternW3C, WAI APG, Deque, MDN

Scroll-linked motion can trigger vestibular discomfort and should honor reduced-motion preferences. Custom scrollers, feeds, and intercepted keys require keyboard reachability, logical focus behavior, no traps, useful announcements, and enough semantics for assistive technology.

Safer alternatives preserve native scrolling and explicit continuationNN/g, Baymard Institute, MDN, W3C

Native long pages, anchor navigation, sticky explainers, proximity snap, explicit Load More controls, reduced-motion fallbacks, and passive observation preserve control while still allowing visual enhancement. The practical rule is to enhance scrolling without turning the user into an animation scrubber or endless-feed operator.

Workflow

1Diagnostic assessmentCheck whether scrolling still behaves like native browser and device movement.
  1. Native scroll remains the default; no custom wheel or touch interception on ordinary reading pages
  2. Users can stop at arbitrary positions; no forced mandatory snap on text-heavy content
  3. Critical content is reachable without scrubbing animations or pinned scenes
  4. Footer, contact links, and support content remain reachable
  5. Returning from a detail page restores list position
  6. Every custom scroll container is keyboard reachable and has a logical focus path
  7. Reduced motion disables non-essential parallax, smooth travel, and scrub effects
  8. Mobile behavior is separately tested; desktop effects are not blindly ported
  9. Zoom, reflow, and increased text spacing do not break snap areas or clip content
  10. There is an obvious escape hatch, such as sticky nav, skip links, anchors, or a section index
2Fix planningMap scroll-control failures to native-scroll alternatives and explicit continuation patterns.
Problem signDesign fixKPI movement
Wheel or touch input is interceptedRestore native document scrolling and observe visibility passivelyLower rage-scroll rate and better INP
Story scene locks the viewportUse normal text flow with a sticky supporting visualBetter reading control and lower early exits
Mandatory snap prevents stopping between sectionsUse proximity snap only in bounded non-text-heavy contextsFewer missed sections and better keyboard behavior
Endless list hides footer or boundariesReplace automatic loading with explicit Load MoreHigher footer reach and clearer stopping points
Parallax or scrubbed motion is non-essentialRespect reduced motion and provide a visible motion toggleLower motion-related abandonment and support demand
Custom scroll container is hard to operate by keyboardMake the region focusable and preserve standard key behaviorFewer accessibility defects and keyboard traps
3Experiment planningCompare task progress, control, and accessibility outcomes instead of optimizing spectacle.
ExperimentControlVariantMetric
Hijacked vs native scrollFull-page scrolljack or scrubbed sequenceNative long page with optional sticky enhancementEarly exit after first scroll and task completion
Endless vs Load MoreAutomatic infinite loadingExplicit Load More with count, status, and footer accessFooter reach rate and return-to-list success
Mandatory vs proximity snapMandatory full-screen snap sectionsProximity snap or no snap for text-heavy contentScroll-depth drop-off and rage-scroll rate
Motion-heavy vs reduced-motion fallbackParallax and scrubbed animation for all usersReduced-motion fallback plus optional motion controlSupport-contact rate and engagement quality
Scroll-math vs passive observerScroll handler drives frame-by-frame animationIntersectionObserver updates supporting visualsP75 INP and long-task frequency
4Measurement planningMeasure scroll harm through responsiveness, reachability, refinding, depth, exits, and support signals.

Track first_scroll, scroll_section_enter, snap_boundary_enter, pinned_scene_enter, list_batch_loaded, load_more_click, footer_visible, detail_view, back_to_list, same_position_restored, reduced_motion_enabled, exit, refresh, and scroll_support_contact. Segment by device, input method, reduced-motion preference, keyboard usage, screen size, source, and task intent.