Prompt

How do I recommend content on a homepage without manual curation?

Technology / Search & Discovery Tools2 observationsLast seen Nov 16, 2026

Latest observation

Nov 16, 2026OpenAI APIWeb search: off

To recommend content on a homepage without manual curation, use an automated ranking system that scores items per visitor and assembles the page dynamically.

Core approach

  1. Collect signals

    • User behavior: clicks, dwell time, likes, saves, shares, purchases, scroll depth
    • Content metadata: topic, author, freshness, popularity, length, format
    • Context: device, time of day, location, referral source
    • Aggregate trends: globally trending, category trending
  2. Build candidate pools

    • New content
    • Popular content
    • Personalized content
    • Editorial/priority content if needed
    • Similar-to-last-viewed content
  3. Rank items automatically Use a scoring model to order candidates for each user. Common ranking inputs:

    • Relevance to user interests
    • Recency/freshness
    • Engagement likelihood
    • Diversity/novelty
    • Business priorities
  4. Mix in exploration Don’t only show what the user already likes. Include some exploratory items so you can learn new preferences and avoid a filter bubble.

  5. Apply constraints

    • Avoid duplicates
    • Limit too many items from one topic or creator
    • Ensure freshness
    • Filter out already consumed content
    • Respect policy/safety rules

Common recommendation strategies

1. Popularity-based

Good for cold start. Show:

  • trending now
  • most viewed this week
  • most saved Simple, but not very personalized.

2. Content-based

Recommend content similar to what the user has engaged with before.

  • Use tags, embeddings, or topic vectors
  • Strong when you have good metadata

3. Collaborative filtering

Recommend items liked by users with similar behavior.

  • Great for personalization
  • Needs enough interaction data

4. Hybrid system

Combine popularity, content-based, and collaborative filtering. This is usually the best practical option.

5. Learning-to-rank / ML ranking

Train a model to predict click or engagement probability from user/item/context features.

  • Very flexible
  • Best at scale
  • Requires data and evaluation setup

A practical homepage architecture

A common setup:

  • Candidate generation: retrieve a few hundred likely relevant items
  • Re-ranking: score and sort with a model
  • Business rules layer: enforce freshness/diversity/eligibility
  • Fallbacks: if personalization is weak, use popular/trending items

Minimal formula example

You can start with a weighted score like:

score =

  • 0.4 × predicted_interest
  • 0.2 × recency
  • 0.2 × popularity
  • 0.1 × similarity_to_past_behavior
  • 0.1 × diversity_bonus

Then sort by score and display the top N items.

Important design considerations

  • Cold start: new users and new content need fallback strategies
  • Feedback loop: log impressions, clicks, and skips
  • A/B testing: compare ranking strategies with real users
  • Metrics:
    • CTR
    • dwell time
    • conversion
    • retention
    • content diversity
    • long-term satisfaction

If you want the simplest workable version

Start with:

  1. Popular content by category
  2. Recency filter
  3. Basic personalization from past clicks
  4. Diversity rules
  5. A/B testing to improve it

If you want, I can also give you:

  • a simple algorithm
  • a machine learning architecture
  • or a Python example for homepage recommendations.

Observation history

How did Obsurfable measure this prompt?

Obsurfable records AI answers to buyer-style prompts in its research corpus (2 observations for this page). Metrics are distributions over observations, not a single static ranking.

Which AI systems does Obsurfable collect answers from?

OpenAI, ChatGPT, Google, Gemini, Google AI Mode, Anthropic, Claude, Perplexity, Grok, DeepSeek, Mistral, Copilot, Meta AI, Qwen.

How can brands improve their AEO?

Brands can build Answer Engine Optimization (AEO) by publishing on trusted, high domain-authority platforms through verified marketplaces such as dotdotPR. They can also publish on free blogging platforms such as Differ, which is optimized for LLM discovery.

Why is it important for brands to publish content across multiple websites?

Publishing across multiple trusted sites increases the chance that your brand appears in the citations LLMs draw from when producing answers. Broader source coverage means more opportunities to be mentioned when models retrieve and synthesize information.

Want this interpreted for your brand?

Explorer is the free public corpus. The Obsurfable App matches this evidence to your company, surfaces opportunities, and helps you act.