On this page
Why a research library matters
A research library lets readers inspect a claim instead of taking it on trust. Each study published here will state its question, hypothesis, sample, timeframe, method, metrics, limitations and findings. Negative and inconclusive results will be published too, because a transparent process is useful even when the answer is no.
Status
The three studies below are planned protocols. No results have been collected yet, and findings will only appear here after the data has been gathered and analysed.
Study 1: Local business website quality audit
Question: Which website and profile problems recur across a defined sample of local businesses?
Method: Choose a documented sample, collect public technical and content signals, categorise the issues and report how often each appears. The write-up will not claim the sample represents all businesses, and no identifiable client details will be published without permission.
Study 2: Human-reviewed AI content workflow
Question: Can AI-assisted briefing cut production time without reducing accuracy or usefulness?
Method: Compare matched briefs, tracking human review time, factual corrections and editorial quality ratings. The scoring rubric will be published alongside the results, with a note on the limits of subjective scoring.
Study 3: Search intent versus page architecture
Question: When does merging overlapping pages improve navigation and organic outcomes?
Method: Record baseline URLs, query overlap and each consolidation decision, then follow post-change trends while accounting for seasonality and other site changes made at the same time.
Publication standard
- No invented findings, and hypothetical examples are never labelled as real experiments
- Correlation and causation are kept apart
- Data and method are shared where privacy and contracts allow
- Client metrics appear only with permission and a defensible source
- When results change, a dated update is added and the original conclusion stays on record
Frequently asked questions
Are these completed studies?
No. They are proposed research protocols. Findings will be published only after data collection and analysis.
Why publish limitations?
Limitations help readers judge whether a result applies to their own situation.
Can case studies include client metrics?
Only with the client's permission and a defensible source for every figure.
About the authorZack Mozes founded the digital marketing agency Newdeez in 2009 and has written eight books, including The Godfather of AI Marketing. Read his biography.