Overview
Every strain in the High IQ database receives a composite score that determines where it appears in search results, recommendation feeds, and discovery pages. Rather than relying on a single metric like user votes, our scoring system combines multiple data dimensions to surface strains that are genuinely noteworthy — whether they are crowd favorites, chemically distinctive, or emerging breakouts. This page explains each scoring factor, how they combine, and what this means for you as a user.Scoring Factors
High IQ evaluates strains across four primary dimensions:1. Popularity Rank
Each strain is assigned a popularity rank (1 through N) based on aggregate consumer interest signals including search frequency, dispensary availability, and cross-platform mention volume. Rank 1 is the most popular strain in the database.- Ranked strains (1–N) are always surfaced before unranked strains
- Unranked strains (rank 0) still appear in results but are sorted by their composite score and then alphabetically
- Popularity data is refreshed through our competitor monitoring pipeline, which tracks thousands of strains across major cannabis platforms
Popularity rank reflects broad market interest, not personal preference. A strain ranked #1 may not be the best fit for your needs — that is what personalized recommendations are for.
2. Terpene Richness
Strains with detailed terpene profiles score higher than those with limited or missing data. Terpene richness measures:- Number of identified terpenes — A strain listing 8 terpenes with precise concentrations is more informative than one listing only “myrcene-dominant”
- Concentration data — Strains with lab-verified milligram-per-gram (mg/g) values receive a scoring boost over those with only qualitative descriptors
- Profile completeness — The presence of both major terpenes (myrcene, limonene, caryophyllene) and minor terpenes (ocimene, terpinolene, humulene) increases the score
3. Cannabinoid Diversity
Beyond THC percentage, High IQ values strains that report a full cannabinoid panel:| Cannabinoid | What It Indicates |
|---|---|
| THC | Primary psychoactive compound |
| CBD | Non-intoxicating therapeutic compound |
| CBG | ”Mother cannabinoid” precursor |
| CBN | Aged THC derivative, sedative properties |
| THCV | Appetite-suppressing, short-acting THC variant |
| CBC | Anti-inflammatory, non-intoxicating |
4. User Engagement
Engagement signals from High IQ mobile app users contribute to a strain’s score:- Stash additions — How many users have added the strain to their stash
- Ratings and reviews — Direct quality feedback from consumers
- Report appearances — Frequency in AI-generated consumption reports
- Favorites — User bookmark activity
How Scores Combine
The four factors are weighted and combined into a single composite score. The general weighting priority is:Popularity Rank (Highest Weight)
Broad market data provides the most stable baseline signal. Strains with established popularity ranks anchor the top of results.
Terpene Richness
Chemical profile depth is the second most important factor. This rewards data quality and helps differentiate strains that share similar popularity.
Cannabinoid Diversity
Full-panel cannabinoid data provides additional differentiation, especially among strains with similar terpene profiles.
High Spectrum Classification Algorithm
Beyond the composite scoring above, each strain is also classified into one of the 6 High Spectrum Families using a Euclidean distance algorithm based on terpene profile vectors. This classification is separate from the ranking score and determines which High Family a strain belongs to. The algorithm uses a 4-tier fallback strategy to handle varying levels of terpene data quality:Tier 1: Full Terpene Profile (60-100% confidence)
When a strain has 3 or more terpenes with valid concentration data, its terpene profile vector is compared against the centroid (mean terpene ratios) of each of the 6 High Spectrum families using Euclidean distance across 9 key terpenes: myrcene, caryophyllene, limonene, linalool, humulene, pinene, terpinolene, fenchol, and ocimene. The strain is assigned to the family whose centroid it is closest to, provided the confidence meets the 60% threshold. Confidence is derived from the inverse of the distance — closer means higher confidence.
Tier 2: Dominant Terpene + Type (50-70% confidence)
When full terpene data is unavailable but a dominant terpene is known, the algorithm maps the dominant terpene to the most likely cluster. For example, terpinolene-dominant strains map to the Rare Energizer family. Strain type (indica/sativa/hybrid) adjusts the confidence.
Tier 3: Type + Effects (35-50% confidence)
When only strain type and effect descriptions are available, the algorithm uses effect-to-cluster mappings. For example, an indica with “relaxed” and “sleepy” effects maps to the Relaxing Body family.
packages/strain-filters/src/cluster-matcher.ts.
The confidence percentage displayed on strain profiles reflects which tier was used. A strain with a rich terpene profile classified at 85% confidence via Tier 1 is a much stronger classification than one at 40% confidence via Tier 4.
Processing Queue
When new strains enter the database — whether from competitor monitoring, user submissions, or research pipelines — they are added to the unprocessed strain queue. This queue is sorted by popularity to ensure that high-demand strains (OG Kush, Blue Dream, Girl Scout Cookies) are enriched and scored before more obscure entries. The queue processing order:- Ranked strains (popularity rank 1–N) — processed first, highest rank first
- Unranked strains with scores — processed next, highest score first
- Unranked strains without scores — processed last, alphabetically
Search & Discovery Impact
Strain scores affect multiple surfaces across the platform:Search Results
When you search for a strain, results are ordered by relevance combined with composite score. A close name match on a popular, data-rich strain appears above a weak match on an obscure strain.
Similar Strains
The “Similar Strains” feature uses terpene profile similarity weighted by scoring. Higher-scored similar strains appear first.
Recommendations
AI-powered recommendations consider your personal history alongside strain scores, balancing what you might like with what is well-documented and reliable.
Discovery Feeds
Browse and discovery pages prioritize strains that are both popular and chemically well-documented.
Score Transparency
We believe in transparency about how strains are ranked. While the exact weighting formula is proprietary and subject to tuning, the factors described on this page are the complete set of inputs. There is no pay-to-rank system — dispensaries and brands cannot purchase higher placement in strain rankings.Frequently Asked Questions
Why does a strain I know is popular appear low in results?
Why does a strain I know is popular appear low in results?
The strain may have limited terpene or cannabinoid data in our database. As our research pipelines enrich the strain’s profile, its score will increase. You can help by submitting lab data through the label scanner.
How often are scores updated?
How often are scores updated?
Scores are recalculated as new data arrives — through pipeline enrichment, user engagement, and competitor monitoring syncs. Most strains see score updates within a week of new data becoming available.
Can I sort by a single factor?
Can I sort by a single factor?
The strain discovery page supports filtering by type, effects, and terpenes, but the default sort always uses the composite score. This ensures the most reliable and data-rich strains appear first.