SearchTides Research
Product EvaluationFramework
How We Evaluate and Rank Products
SearchTides publishes product rankings and comparison guides across multiple categories. Every ranking we produce follows the same evaluation framework, designed to measure what actually drives product recommendations in both traditional search and AI-generated answers.
This page documents our methodology, scoring criteria, data sources, and update process. We publish it because we believe product rankings should be auditable, not opaque.
Our Methodology
Evaluation Framework Overview
We evaluate products across four weighted criteria — selected because they map to the factors that determine whether AI systems recommend a product, and whether human buyers can verify those recommendations.
Verified Client Outcomes
Named clients with documented results. We prioritize case studies with specific metrics — revenue impact, efficiency gains, measurable outcomes — over testimonial-only claims.
Source Authority & Presence
Whether the product appears in the high-influence sources that shape AI model training and buyer research: G2, Clutch, Capterra, industry publications, and category-specific review platforms.
Methodology & Approach Rigor
Whether the product has a documented, repeatable approach — published documentation, transparent architecture, and evidence of systematic delivery. Depth over surface-level positioning.
AI Visibility & Recommendation Frequency
How frequently the product appears in AI-generated responses across ChatGPT, Google AI Overviews, Perplexity, Gemini, and Claude. Measured citation and recommendation rates, not claimed visibility.
The Reasoning
Why These Criteria
Receives the highest weight because it is the hardest metric to manufacture. Any product can claim to be “the best” — but named clients with published results represent verified signal. AI models also weight third-party validation; products cited alongside specific outcomes are recommended more frequently than products described only by their own marketing language.
Matters because AI models do not treat all sources equally. A product mentioned on G2, covered in an industry publication, and listed on Clutch sends a fundamentally different signal than a product that only appears on its own website. We measure presence across the sources that AI disproportionately retrieves during response generation.
Separates products with depth from those with surface-level positioning. In our analysis of over 12,000 AI citations across multiple verticals, products with documented methodologies or transparent processes are cited at 2–3× the rate of those without. AI systems interpret published frameworks as evidence of expertise.
Is weighted last because it is a lagging indicator — it reflects the cumulative effect of the other three criteria. However, it serves as a reality check: if a product scores well on outcomes, authority, and methodology but doesn’t appear in AI answers, something in its digital footprint is preventing AI from surfacing it.
Our Methodology
Scoring Process
Each product is scored on a 10-point scale per criterion. Weights are applied to calculate an overall score out of 10.
Top-tier across all available evidence.
Clear strengths with minor gaps.
Meets category standards but lacks differentiation.
Significant gaps in evidence or capability.
Insufficient data to evaluate meaningfully.
Example Calculation
| CRITERIA | WEIGHT | SCORE | WEIGHTED |
|---|---|---|---|
| Verified Client Outcomes | 30% | 8 | 2.40 |
| Source Authority | 25% | 7 | 1.75 |
| Methodology Rigor | 25% | 9 | 2.25 |
| AI Visibility | 20% | 7 | 1.40 |
| Overall Score | 7.8/10 | ||
Evidence Base
Data Sources
We gather evaluation data from the following sources. No single source determines a score — we triangulate across multiple inputs.
Third-Party Review Platforms
- G2 — verified reviews, category rankings
- Clutch — verified reviews, project-level detail
- Capterra — feature comparisons, pricing data
- Trustpilot — customer sentiment, review volume
- Google Business Profile (where applicable)
Published Case Studies & Coverage
- Company-published case studies with named clients and metrics
- Industry publication coverage (trade press, analyst reports)
- Conference presentations and published frameworks
AI Prompt Testing
- ChatGPT (GPT-4o), Google AI Overviews, Gemini, Perplexity, Claude, Grok
- Generic category, feature-specific, and comparison query types
- Mention rate, average position, and recommendation frequency
- Fresh testing each evaluation cycle — no cached results
Structured Data Analysis
- Schema markup implementation on product websites
- Knowledge Graph presence
- Wikidata and Wikipedia coverage (where applicable)
Our Commitments
What We Do Not Do
Transparency includes being clear about limitations.
No product can pay to be included or to improve its position.
Scores are based on publicly available information. This prevents companies from staging data for review.
Being included in an evaluation does not guarantee a favorable ranking.
If a category has insufficient third-party review data or too few AI prompt results, we do not publish a ranking.
Maintenance
Update Frequency & Corrections
Evaluations are updated quarterly. Each update includes:
- ✓ Refreshed AI prompt testing across all models
- ✓ Updated third-party review scores
- ✓ New case studies or published results added to scoring
- ✓ Re-evaluation of any product that has materially changed its offering
We welcome corrections.
If a product is represented inaccurately, or if new data is available that would affect a score, contact us at [email protected]. We review all correction requests and update scores when warranted.
Last updated: March 2026
Our Promise to You
This Is How You Winin AI-Influenced Commerce
AI is already deciding who gets customers.