Editorial standards and methodology

Where the data comes from

Most claims on this site fall into one of three categories:

• Primary research we run. Daily prompt sets across ChatGPT, Gemini, Claude, and Perplexity, parsed for brand mentions, citations, and sentiment.
• Third-party studies from sources like SE Ranking, Ahrefs, Capgemini, Yext, Profound, and SparkToro. We cite them by name and link to the primary source where possible.
• Observational pattern recognition from working with customers and watching the AI engines change. Marked clearly as such.

How we mark uncertainty

AI engines change weekly. A stat that was true in March may not hold in May. Where research is from a single study, contested across sources, or based on a small sample, we mark it as directional rather than absolute. Specific numbers are always attributed to the source we got them from, with the publication year.

When we update posts

Each post and guide carries a "Last updated" date in the header. We refresh major guides quarterly. We re-check stat-heavy posts when AI engines ship significant changes (e.g., model upgrades, new citation surfaces) or when a cited source publishes new research.

What we won't do

• Recycle stats across posts without re-checking them against the original source
• Generate research with AI and present it as our own measurement
• Publish anything we haven't tested or wouldn't recommend to our own customers
• Hide author identity or publish anonymously

Corrections and feedback

If you spot something inaccurate, outdated, or unclear, email hello@honeyb.ai. We'll review, correct, and credit the catch where appropriate.