The continuous aggregation of diverse content into a single scannable flow, transforming scattered updates from multiple sources into an ongoing stream that sustains engagement over time. Unlike activity logs that record system events, feeds aggregate content types — posts, updates, articles, media — for consumption and quick engagement, prioritising recency.
Ordering strategies
- Chronological: reverse-chronological ordering where newest items always appear first. Simple and predictable – users know where to look for updates.
- Algorithmic: Sequence determined by relevance algorithms. Orders content based on predicted interest rather than strict time order.
Variants
- Personal: Aggregates content relevant to individual user – posts from connections, subscribed sources, or topics of interest. Single unified feed across multiple sources.
- Curated: Publisher-defined collections organised by topic, source type, or editorial selection. Users browse multiple feeds (e.g., “Technology”, “Design”, “News”) alongside personal feed.
- Custom: User-created filters based on topics, tags, or lists of specific sources. Advanced functionality for power users wanting precise control over content feeds.
Item anatomy
Each feed item typically includes:
- Content: full text for micro-updates, or title/teaser with thumbnail for longer content
- Source: actor (person, organisation, system) who created or triggered the item
- Timestamp: relative time for recent items (“2 hours ago”), switching to absolute timestamps as content ages
- Actions: lightweight engagement options (like, star, bookmark) alongside deeper responses (comment, share)
Related components
- Cards – common container format for feed items
- Messaging / Comment thread
Resources & references
To-do
- Infinite scrolling — loads additional content as the user reaches the end of current items.
Consequences
- updates from everywhere arrive in one scannable flow, so noticing what changed costs one look instead of several rounds
- the ordering is a claim about what matters: chronological is predictable and treats everything as equal, algorithmic is relevant and unaccountable
- a feed has no end, so there is no state in which the actor is finished with it — which is both why it sustains attention and why it is hard to put down
Related patterns
Serves
- Assistance — assistance's perceiving step — keeping recent events visible so the actor can notice what matters
- Delegation — the feed gives the user awareness of background agent activity during delegation's execution stage
Complements
- Activity log — distinct from feeds; logs are histories of actions taken by an entity, whilst feeds aggregate content of interest to users from diverse sources
- Collaboration — uses feeds to surface updates and changes in shared workspaces
- Notification — alerts users to new feed activity requiring attention
- Conversation — comment threads enable conversations around feed items