My learning 0 of 20 lessons complete
Start lesson 1 →
Back to compare picker

Recommendations vs Related links vs Recently viewed vs Browse by category

Use recommendations when the system ranks suggested items from context, behavior, item similarity, popularity, or model output; related links when editors curate onward destinations from the current page; recently viewed when the list is the user's actual history; and browse by category when users choose from a taxonomy.

Learn the decision

Match the interaction to the person's task

Use recommendations when the system ranks suggested items from context, behavior, item similarity, popularity, or model output; related links when editors curate onward destinations from the current page; recently viewed when the list is the user's actual history; and browse by category when users choose from a taxonomy.

Recommendations

Use it when: Users need discovery help in a large item, product, content, service, or action space.

Make this state clear: Default recommended state with labelled section, item identity, reason, source, and action controls.

Observe the result: A signed-out shopper sees Popular in backpacks rather than fake personalized copy, then gets personalized reasons only after activity exists and consent settings allow it.

Study the full pattern →

Related links

Use it when: The current page has a few genuinely adjacent pages, services, programs, or resources users often need next.

Make this state clear: Default state with a labelled, curated related-links block and descriptive link text.

Observe the result: A user reads claim renewal guidance, chooses Upload evidence for your claim, and sees why that destination is the next useful service page.

Study the full pattern →

Recently viewed

Use it when: Users inspect multiple objects and often need to return to one they recently opened.

Make this state clear: Empty or hidden state before any qualifying item has been viewed.

Observe the result: A user on a shared account removes an unexpected document from the recent list and understands that Favorites are stored separately.

Study the full pattern →

Browse by category

Use it when: Users are exploring a broad catalog, service directory, help center, learning library, product collection, or content repository.

Make this state clear: Top-level category state with clear names, descriptions, and enough examples or counts to choose.

Observe the result: An editor reviews low-click categories, merges overlapping labels, and retags misplaced content after checking search terms and support questions.

Study the full pattern →

Practice the judgment: Imagine this situation: The product mixes ads, recents, saved items, editorial links, and algorithmic suggestions under one vague heading. Write the person's goal, choose one of these patterns, and show the state that helps the person continue. Use the decision rules and source trail below to explain your choice.

Review the underlying principle →

Decision dimensions

Dimension RecommendationsRelated linksRecently viewedBrowse by category
UI or UX UI + UX - Ranked suggestions generated from context, behavior, item similarity, popularity, editorial rules, or recommendation modelsUI + UX - Curated onward links connected to the current contentUI + UX - Automatic list of items the current user opened recentlyUI + UX - Taxonomy-based category navigation for exploring a collection before choosing a destination
UI guidance Render recommendations as a labelled set of suggested items with clear item identity, recommendation reason, source or basis, availability, and a visible way to dismiss or tune at least the current item.Render a short, labelled collection of links whose destinations are closely related to the current content, using destination-specific link text and optional relation labels such as service, guidance, external, or PDF.Render a labelled list or rail of items the current user actually opened, ordered most recent first, with enough identity to recognize each item such as name, type, thumbnail or icon, location, status, and last-viewed time.Render categories as a scannable list or grid with user-facing names, short descriptions, item counts or example contents, and clear parent-child location cues.
UX guidance Use recommendations to reduce discovery effort when the system has evidence that a small set of items, products, services, content, or next actions may be useful in the user's current context.Use related links to support users who have finished or understood the current content and need a relevant next page, adjacent service, reference, or follow-up resource.Use recently viewed to reduce re-finding effort when users compare items, pause work, resume documents, or return to records they inspected during the current or recent sessions.Maintain the taxonomy with evidence from content tagging, search terms, analytics, tree testing, support questions, and editorial review so category labels stay distinct and useful.
Good UI A commerce product page shows Pairs well with this camera, hides out-of-stock accessories, limits the rail to three visible items, and marks manually selected merchant recommendations separately from generated recommendations.A benefits guidance page ends with Related links: Check eligibility for support, Upload evidence for your claim, and Appeal a support decision, each with a concise relation label.A procurement dashboard shows Recently viewed records with title, record type, status, project, last-viewed time, and a remove control for each row.A services page lists Benefits, Housing and local services, Money and tax, and Driving and transport with one-line descriptions and top tasks for each category.
Bad UI A recommendation list keeps showing unavailable, out-of-stock, permission-restricted, or already dismissed items because the UI only validates card shape.A page ends with More information containing Home, Contact us, Apply now, Old 2018 guidance, Help, and an unrelated account settings link.Recent history becomes a surveillance surface on shared accounts or screen shares.Category names overlap and users bounce between siblings.
Good UX A signed-out shopper sees Popular in backpacks rather than fake personalized copy, then gets personalized reasons only after activity exists and consent settings allow it.A user reads claim renewal guidance, chooses Upload evidence for your claim, and sees why that destination is the next useful service page.A user on a shared account removes an unexpected document from the recent list and understands that Favorites are stored separately.An editor reviews low-click categories, merges overlapping labels, and retags misplaced content after checking search terms and support questions.
Bad UX Users assume recommendations are mandatory next steps because the UI mixes them with required workflow tasks.Users follow a generic More information link and land on an unrelated policy collection.Users trust a Recently viewed rail as a recommendation and choose an irrelevant item because sponsored content was mixed into history.Users choose Customers because they need customer support, but the category contains only internal CRM configuration articles.
Best fit Users need discovery help in a large item, product, content, service, or action space.The current page has a few genuinely adjacent pages, services, programs, or resources users often need next.Users inspect multiple objects and often need to return to one they recently opened.Users are exploring a broad catalog, service directory, help center, learning library, product collection, or content repository.
Avoid when The product has too few items for ranking to reduce effort.Links are only loosely associated by topic tags or organizational ownership.The team wants to promote products, recommend next actions, or show related content rather than return personal history.The collection is small enough for a simple list.
Required state Default recommended state with labelled section, item identity, reason, source, and action controls.Default state with a labelled, curated related-links block and descriptive link text.Empty or hidden state before any qualifying item has been viewed.Top-level category state with clear names, descriptions, and enough examples or counts to choose.
Accessibility burden Make recommendation reasons, sponsored labels, availability, and item type available as text.Use descriptive link text that makes sense out of context.Use a heading or labelled region that describes the scope of the list.Use real links for categories that navigate and buttons only for local expansion.
Common misuse Pretending generated suggestions are recently viewed, saved, or required steps.Using related links as a catch-all further-reading dump.Filling recently viewed with recommendations, ads, popular items, or related content.Using the organization chart as the category taxonomy.

Recommendations

UI or UX
UI + UX - Ranked suggestions generated from context, behavior, item similarity, popularity, editorial rules, or recommendation models
UI guidance
Render recommendations as a labelled set of suggested items with clear item identity, recommendation reason, source or basis, availability, and a visible way to dismiss or tune at least the current item.
UX guidance
Use recommendations to reduce discovery effort when the system has evidence that a small set of items, products, services, content, or next actions may be useful in the user's current context.
Good UI
A commerce product page shows Pairs well with this camera, hides out-of-stock accessories, limits the rail to three visible items, and marks manually selected merchant recommendations separately from generated recommendations.
Bad UI
A recommendation list keeps showing unavailable, out-of-stock, permission-restricted, or already dismissed items because the UI only validates card shape.
Good UX
A signed-out shopper sees Popular in backpacks rather than fake personalized copy, then gets personalized reasons only after activity exists and consent settings allow it.
Bad UX
Users assume recommendations are mandatory next steps because the UI mixes them with required workflow tasks.
Best fit
Users need discovery help in a large item, product, content, service, or action space.
Avoid when
The product has too few items for ranking to reduce effort.
Required state
Default recommended state with labelled section, item identity, reason, source, and action controls.
Accessibility burden
Make recommendation reasons, sponsored labels, availability, and item type available as text.
Common misuse
Pretending generated suggestions are recently viewed, saved, or required steps.

Related links

UI or UX
UI + UX - Curated onward links connected to the current content
UI guidance
Render a short, labelled collection of links whose destinations are closely related to the current content, using destination-specific link text and optional relation labels such as service, guidance, external, or PDF.
UX guidance
Use related links to support users who have finished or understood the current content and need a relevant next page, adjacent service, reference, or follow-up resource.
Good UI
A benefits guidance page ends with Related links: Check eligibility for support, Upload evidence for your claim, and Appeal a support decision, each with a concise relation label.
Bad UI
A page ends with More information containing Home, Contact us, Apply now, Old 2018 guidance, Help, and an unrelated account settings link.
Good UX
A user reads claim renewal guidance, chooses Upload evidence for your claim, and sees why that destination is the next useful service page.
Bad UX
Users follow a generic More information link and land on an unrelated policy collection.
Best fit
The current page has a few genuinely adjacent pages, services, programs, or resources users often need next.
Avoid when
Links are only loosely associated by topic tags or organizational ownership.
Required state
Default state with a labelled, curated related-links block and descriptive link text.
Accessibility burden
Use descriptive link text that makes sense out of context.
Common misuse
Using related links as a catch-all further-reading dump.

Recently viewed

UI or UX
UI + UX - Automatic list of items the current user opened recently
UI guidance
Render a labelled list or rail of items the current user actually opened, ordered most recent first, with enough identity to recognize each item such as name, type, thumbnail or icon, location, status, and last-viewed time.
UX guidance
Use recently viewed to reduce re-finding effort when users compare items, pause work, resume documents, or return to records they inspected during the current or recent sessions.
Good UI
A procurement dashboard shows Recently viewed records with title, record type, status, project, last-viewed time, and a remove control for each row.
Bad UI
Recent history becomes a surveillance surface on shared accounts or screen shares.
Good UX
A user on a shared account removes an unexpected document from the recent list and understands that Favorites are stored separately.
Bad UX
Users trust a Recently viewed rail as a recommendation and choose an irrelevant item because sponsored content was mixed into history.
Best fit
Users inspect multiple objects and often need to return to one they recently opened.
Avoid when
The team wants to promote products, recommend next actions, or show related content rather than return personal history.
Required state
Empty or hidden state before any qualifying item has been viewed.
Accessibility burden
Use a heading or labelled region that describes the scope of the list.
Common misuse
Filling recently viewed with recommendations, ads, popular items, or related content.

Browse by category

UI or UX
UI + UX - Taxonomy-based category navigation for exploring a collection before choosing a destination
UI guidance
Render categories as a scannable list or grid with user-facing names, short descriptions, item counts or example contents, and clear parent-child location cues.
UX guidance
Maintain the taxonomy with evidence from content tagging, search terms, analytics, tree testing, support questions, and editorial review so category labels stay distinct and useful.
Good UI
A services page lists Benefits, Housing and local services, Money and tax, and Driving and transport with one-line descriptions and top tasks for each category.
Bad UI
Category names overlap and users bounce between siblings.
Good UX
An editor reviews low-click categories, merges overlapping labels, and retags misplaced content after checking search terms and support questions.
Bad UX
Users choose Customers because they need customer support, but the category contains only internal CRM configuration articles.
Best fit
Users are exploring a broad catalog, service directory, help center, learning library, product collection, or content repository.
Avoid when
The collection is small enough for a simple list.
Required state
Top-level category state with clear names, descriptions, and enough examples or counts to choose.
Accessibility burden
Use real links for categories that navigate and buttons only for local expansion.
Common misuse
Using the organization chart as the category taxonomy.
Decision rules
  • Prefer recommendations when the product predicts or ranks items that may be useful now, such as related products, next services, frequently bought together items, trending content, or personalized suggestions.
  • Prefer related links when a content owner can name a small set of optional onward pages that are relevant from this exact page without using personal behavior or model ranking.
  • Prefer recently viewed when every entry is an item the current user actually opened, ordered by recency, with remove and clear history controls.
  • Prefer browse by category when users need a stable taxonomy of topics, services, products, or content types and want to choose their path rather than accept ranked suggestions.
  • Recommendations should expose the basis for each item with labels such as Because you saved this, Frequently bought together, Trending in this category, or Editorial fallback.
  • Do not label recommendations as recently viewed, saved, or required unless the user actually performed that action or the task requires the item.
  • Separate sponsored, promoted, manual, editorial, and model-generated recommendations so users can judge commercial influence and data source.
  • When personal data is used, provide controls for feedback, dismissal, activity source, history, personalization, and reset; related links and taxonomy pages usually need editorial maintenance controls instead.
  • When recommendation data is weak, use a labelled cold-start fallback such as popular in this category or curated by editors rather than fabricating personalized reasons.
  • Filter out unavailable, permission-restricted, out-of-stock, already-in-cart, duplicate, and repeatedly dismissed items before treating the recommendation set as complete.
Inspect live examples
Failure modes
  • A generated recommendation rail is marked complete because it renders cards, but no item has a reason, source, dismissal path, or availability check.
  • The product mixes ads, recents, saved items, editorial links, and algorithmic suggestions under one vague heading.
  • A recommendation says Based on your activity without giving users control over activity history or personalization.
  • Cold-start users see fake personalized copy instead of a clear popular, trending, or editorial fallback.
  • Related links are replaced with model output that includes stale or weakly related pages no editor has reviewed.
  • Recently viewed history includes predicted items the user never opened, damaging trust and privacy expectations.
  • A taxonomy page is reduced to a recommendation carousel, so users cannot inspect the full category structure.

Saved items