Overview
ScaleGrowth’s standard recommendation engine is designed to work flexibly with sparse datasets – generating matches through a mix of profile, preference, and behavioral signals. However, for enterprise or scale-tier customers who need stricter control over recommendation logic, we offer a custom feature called a Hard Filter.
This guide covers:
- What Hard Filters are and how they work
- When to use them (and when not to)
- Key requirements and best practices
- How to request a Hard Filter setup
🎯 What Is a Hard Filter?
A Hard Filter is a customized recommendation rule that strictly limits what content or members are shown based on a specific field or set of fields.
Think of it like placing a permanent saved search filter inside the recommendation engine itself.
Only content or members that match on the specified field(s) will be eligible to appear in a member’s recommendations.
Example Use Case:
Let’s say you want to only recommend content that matches a member’s selected topics of interest.
You can configure a Hard Filter to ensure:
- The system checks that the content’s tags must match the member’s preferences
- Any content that doesn’t match the selected tags is excluded entirely
⚙️ How It Works
- A field (e.g. Preference: Tag Set 1) is selected as the filtering criteria
- The system filters out all non-matching content or members before running its normal tiered matching logic
- Only results that meet the field condition(s) are passed into the ranking algorithm
This is stricter than standard matching, where the system may still show Tier 2 or Tier 3 matches even if there's no perfect field-level match.
🚧 Important Considerations Before Enabling
1. It’s a Customization
Hard Filters are not a standard toggle – they require a custom setup by our data science and product teams. This is available to customers our Enterprise plans.
2. You Must Have Enough Content
If your content (or member base) is not thoroughly tagged or is too limited in quantity, applying a Hard Filter may result in zero recommendations.
3. The Filter Field Should Be Required
To maximize match potential:
- The field used as the filter (e.g. Preference: Tag Set 1) should be required during onboarding
- All relevant content or member profiles should be consistently tagged with values from that field
4. Clean, Consistent Data Is Essential
Because the system is restricted from recommending “fuzzy” matches outside the filter, data quality becomes critical.
✅ When to Use Hard Filters
Hard Filters are ideal when:
- You have a large, well-tagged dataset
- You want to ensure recommendations only reflect member preferences
- You are running a use case that requires high relevance and specificity, such as:
- Regulated spaces (e.g. healthcare, legal, certified experts only)
- Enterprise matching programs
- Structured mentorship or coaching platforms
⚠️ When NOT to Use Hard Filters
Avoid Hard Filters if:
- You are still onboarding members or content
- Tagging is inconsistent or incomplete
- You want broader discovery or exploratory matching
- Your platform has fewer than ~500 content items or active member profiles
Over-restricting recommendations in early stages may result in a poor experience with empty or repetitive suggestions.
📩 How to Request a Hard Filter Setup
To enable a Hard Filter, please contact your Customer Success Manager or submit a request via the Success Center.
Our team will:
- Review your field structure and tagging coverage
- Advise on whether a Hard Filter is appropriate
- Configure the matching engine with the selected filters
- Test results and provide optimization guidance
💡 Pro Tip: Combine with Saved Searches
If you're not ready for a Hard Filter but want high-intent discovery, consider using:
- Saved Searches + Personalized Digests
- These give members control over what they see – without modifying the recommendation engine globally.
Learn more: Using Saved Searches and Personalized Digests for Member-Centered Discovery
Related Articles
- How ScaleGrowth Designed Its Recommendation Engine for Member Platforms
- Understanding Profile vs. Preferences Fields for Matching & Recommendations
- Using Saved Searches and Personalized Digests for Member-Centered Discovery
