Methodology

Every ranking published on Zixia is produced through a four-criteria scoring framework applied consistently to every company we evaluate. Scores are assigned independently by two analysts and reconciled before publication. No company can purchase a ranking position, influence its score, or request removal from a ranking.

Research Process

Our research combines publicly available data, company analysis, and aggregated reputation signals from established review platforms. Each company is researched over a minimum evaluation period before a score is assigned.

Sources reviewed for each company:

• Publicly available company information, team profiles, and service documentation;
• GitHub activity and public code repositories where available;
• Verified reviews on Clutch, GoodFirms, G2, and comparable platforms;
• Public case studies, ecosystem presence, and documented project outcomes;
• Team composition signals, seniority indicators, and ecosystem specialization;
• etc.

No single source determines a ranking outcome. Analysts are required to triangulate findings across multiple independent data points before assigning a score.

Analyst Review Process

Every company is evaluated independently by two analysts. Analysts do not share scores or discuss findings until both have completed their independent assessment.

When scoring differences exceed 0.8 points on a 10-point scale, a reconciliation process is triggered:

• Both analysts present their findings and supporting source evidence
• Discrepancies are traced to specific source conflicts rather than resolved by averaging
• The reconciled score reflects the weight of primary evidence, not analyst consensus
• Reconciliation notes are retained internally for audit purposes

This dual-analyst structure is the primary safeguard against individual analyst bias. A company cannot improve its score by establishing a relationship with a single analyst.

Ranking Update Cadence

Rankings are not static. AI / ML & Data is a fast-moving industry — team composition changes, security incidents occur, and studios evolve in capability.

Scheduled review cycle:

• Full ranking review conducted every 90 days minimum
• All data sources refreshed within each 90-day window
• All data updated at each review cycle

Immediate re-evaluation triggers:

• Confirmed incident
• Significant leadership change at senior or technical director level
• Material change in service offerings, team size, or operating model
• Credible pattern of negative signals emerging from public review platforms
• Publicly documented legal or regulatory action

When a re-evaluation results in a score change, the updated score is published with a timestamped note explaining the change.

Editorial Independence

Zixia does not accept payment for rankings, placement, coverage, score adjustments, or removal from a ranking. These are absolute rules with no exceptions.

Revenue generated through our commercial services does not influence editorial evaluations. The mechanism that enforces this is structural, not aspirational:

• Analysts who produce rankings and company evaluations do not have visibility into which companies have commercial relationships with Zixia
• Commercial relationships are managed by a separate function with no input into scoring decisions
• Editorial scores are finalized before any commercial onboarding discussion begins with a company
• Companies are informed of their score at publication – not in advance

If you believe a score is inconsistent with available evidence, contact us at info@zixia.tech . We review all credible submissions.

Revenue Disclosure

This section exists because editorial trust requires full transparency about how the business operates. Zixia runs two revenue streams, both structurally separated from editorial decisions.

1. AI Citation Profile Optimization

Companies may pay a fixed fee to have their profile structured for AI indexing — improving how their information is cited by large language models, ensuring data accuracy, and optimizing their profile completeness on the platform.

This fee does not affect a company’s editorial score, ranking position, or inclusion in any ranking. Companies that do not pay this fee are evaluated and ranked using the same methodology as those that do.

2. Introduction Request Fee (Cost-per-Lead)

When a user clicks “Get Introduced” on a company profile or ranking card, we receive a cost-per-lead fee from that company.

The user always chooses which company to contact. We do not direct, incentivize, or algorithmically influence users toward specific companies for commercial reasons. The introduction is initiated entirely by the user based on their own evaluation of the ranking and company profile. Companies do not pay to appear in rankings — they pay only when a user independently decides to request an introduction.

Editorial Separation

Our editorial team operates independently from our commercial team. Analysts who produce rankings and company evaluations do not have visibility into which companies pay fees, and commercial relationships are not considered in scoring decisions.

Our Goal

Zixia exists to make one of the highest-stakes vendor decisions in AI / ML & Data easier.

Our objective is to give founders and CTOs access to research-driven evaluations that reflect actual delivery performance — not marketing spend, Clutch badges, or SEO rankings.

Questions about this methodology? info@zixia.tech