The Marfeel Content Recommender system offers a variety of pre-configured recommender engines to help you deliver content tailored to your audience’s needs. Each recommender weights ranking signals differently to tailor the recommended content accordingly. This guide provides an easy-to-understand explanation of each recommender type, helping you choose the best one for your needs.
| Most Read | Highlights the most popular articles based on overall traffic to attract users to content with proven broad appeal. |
| Latest | Features the most recently published articles on your site to promote Breaking News and ensure users see the newest updates first |
| Trendy | Features articles with a sharp increase in traffic over short periods to lever current momentum to engage users. |
| Similar | Recommends articles closely related to the one being viewed so users keep exploring content they’re already interested in. |
| Highest CTR | Surfaces articles that have performed well in terms of click-through rate to boost visibility of articles users are more likely to click on… |
| Personalized | Tailors recommendations to the user’s interests, preferences and behavior to increase engagement by aligning recommendations with user preferences. |
| Default | General Recommendations for Broad Audiences ensuring all types of content get visibility without over-prioritizing one signal. |
To ensure that users are consistently presented with fresh and engaging content, the recommender system automatically excludes articles that a user has already read from their recommendations. This feature maximizes recirculation by encouraging users to explore new content rather than revisiting what they’ve already seen.
1 . Most Read
The Most Read recommender highlights the most popular articles based on overall traffic. It emphasizes the popularity ranking signal ensuring that articles with the highest number of views over a specified period (e.g., last 24 hours, last week) are shown.
Ideal for
- To showcase your site’s top-performing content that resonates with a broad audience.
- Showcasing evergreen or cornerstone articles that consistently perform well.
- Increasing page views by promoting content already proven popular.
Example Usage
- On your homepage, display a “Most Read Articles” section to attract new visitors with content that others find valuable.
- Display the day’s most read articles on sidebar modules, such as a breaking news story or viral feature.
- In newsletters, include top-performing articles to encourage click-throughs.
2. Latest
Features the most recently published articles on your site. The Latest recommender prioritizes the recency of articles, ensuring that the newest content is presented to users. It uses the publication date as the primary factor, regardless of other signals like popularity or personalization.
Ideal for
- News websites or sections where timely information is crucial like breaking news or other time-sensitive content.
- Keeping frequent visitors up-to-date with the latest content with sections like “Latest News” or “Fresh Updates”.
- Highlighting new content to boost initial visibility.
Example Usage
- A “Latest News” section that automatically updates with new articles as they are published.
- On category or section pages, display the newest articles related to that topic.
- A tech site can use this to display updates on the latest gadgets or industry announcements of any selected topic.
3. Trendy
Highlights articles that are rapidly gaining popularity identifying articles with a sharp increase in traffic over short periods. The Trendy recommender focuses on the trend ranking signal. It identifies articles that have seen a recent spike in traffic, indicating growing interest. It balances recent activity with overall popularity to surface content that’s on the rise.
Ideal for
- To surface articles that are “on the rise” in interest.
- Capturing viral content or topics that are suddenly in demand.
- Engaging users with what’s currently hot or newsworthy.
- Encouraging users to read articles that others are finding interesting right now.
Example Usage
- A “Trending Now” or “What’s Hot” section to showcase articles gaining traction.
- Push notifications or alerts for articles that are rapidly increasing in popularity.
4. Similar
Recommends articles closely related to the one being viewed. The Similar recommender leverages the similarity ranking signal. It analyzes shared topics, sections, authors, and content similarity to suggest articles that are closely related to the current one.
Ideal for
- To keep users engaged with content on the same topic.
- Deepening user exploration within a specific topic or category.
- Increasing session duration by encouraging sequential reading.
Example Usage
- Inserting in-article suggestions that link to similar content.
- For “Related Articles” modules at the end of an article to encourage further reading.
- If a user is reading an article about “healthy recipes,” this recommender might suggest other recipes, articles on nutrition, or cooking tips.
5. Highest CTR
Surfaces articles that have performed well in terms of click-through rate (CTR) within recirculation modules. The Highest CTR recommender emphasizes the Recirculation Modules Click-Through Rate (CTR) ranking signal. It identifies articles that perform exceptionally well when recommended, indicating high user interest when the article is presented.
Ideal for
- To promote content that has proven appeal in your recommendation modules.
- Optimizing recirculation modules for higher overall CTR.
- Highlighting articles that effectively capture user attention.
Example Usage
- Populating sidebar or footer recommendation widgets with high-CTR articles.
- A shopping site might use this to display products or articles with the highest CTR in specific categories.
- Testing different content placements to see which articles consistently get clicks.
6. Personalized
Tailors recommendations to the user’s interests, preferences and behavior. The Personalized recommender uses the personalization ranking signal. It considers the user’s browsing history, favorite topics, authors, and sections to suggest content that aligns with their interests.
Ideal for
- Providing a unique experience for each user.
- Increasing user loyalty by consistently delivering relevant content.
- Encouraging return visits by catering to user preferences.
- For logged-in users or returning visitors.
Example Usage
- A “Recommended for You” section that updates based on the user’s reading habits.
- Email newsletters personalized with articles matching the subscriber’s interests.
- A fashion site might recommend articles about brands, styles, or designers a user frequently views.
7. Default
Provides a balanced recommendation approach, combining all signals without heavy weighting on any particular one.
The Default recommender integrates various signals—threshold , base value , trend , pageview value , and CTR —to provide well-rounded recommendations. It doesn’t overly emphasize any single factor, aiming for a holistic approach.
Ideal for
- General-purpose recommendations when no specific strategy is needed.
- New users or anonymous visitors without personalization data.
- Situations where a balance between popularity, relevance, and freshness is desired.
Example Usage
- For general recommendation modules when no specific strategy is required.
- Ideal for sites with diverse content or audiences.
- Filling recommendation slots on pages where specific targeting isn’t required.
- Serving content to users when personalization data is limited or unavailable.