Version history
Changes to functionality and data presentation.
Evidence and research
The method, system and protocols behind AiVisible findings.
We show how measurement works, how the system develops and how category studies and client projects are published. Every resource states its evidence level.
A finding, a market observation and a project outcome are treated as three different evidence levels.
Credibility standard
What can be verified today
1Metric definitions
2Evidence trail in the system
3Version history
4Research protocols
5Explicit limitations
An inconclusive result is not hidden or replaced with a simulation.
Methodology
We show how measurement works before showing what changed.
Public definitions, question sets, coding rules and limitations make it possible to assess process quality.
Product log
We document system development as part of the research practice.
Evidence trail
Question, answer, model, date and coding.
Limitations
What the system can show and what it cannot confirm.
Research register
Every category study follows the same qualification standard.
The register records the category, dates, models, sample size, key observations and limitations. Market research remains separate from service-effectiveness claims.
Important:A resource enters the public register only after checks for sample, coding and limitation completeness.
Project documentation
Client projects are documented from the initial measurement.
scope and dates
frozen sample core
implementation list
follow-up wave
limitations
business outcome only when available
Publication standard
Every future report should be assessable independently of its conclusion.
Publication standard
Minimum category-research protocol
Active system structure
Objective and research questionrequired
Category selection criteriarequired
Dates and model listrequired
Sample size and denominatorrequired
Question-set versionrequired
Coding definitionsrequired
Underlying resultsrequired
Interpretation limitsrequired
Publication requirement
Observation, interpretation and a claim of effectiveness are three different levels of evidence.
a reviewable method
a result with a denominator
limitations next to the finding
objective and research question
category and brand-selection criteria
dates, models and sample size
coding definitions
results with denominators
observations separated from interpretation
limitations and alternative explanations
What counts as evidence
Not every resource demonstrates service effectiveness.
Protocol
Supports the quality and repeatability of the research process.
Product trail
Shows that a finding has source material and history.
Internal experiment
Demonstrates a mechanism but is not a client result.
Client project
Can support a change claim only with a baseline, scope, follow-up wave and permission.
Next step
Suggest a category worth studying publicly.
We select categories with a clear comparison process and enough brands for a useful sample.
FAQ
Questions before you decide
How do you distinguish a category study from a client project?
A category study describes a market in a defined sample. A client project can support a change only by comparing the baseline, implementation scope, follow-up wave and business conditions.
Does a category study prove the service works?
No. Market research can describe brand presence and answer mechanisms, but it does not prove the effect of a specific implementation.
What must a published project include?
Only approved material is published. It should include the method, baseline, change list, follow-up wave, limitations and a business result only where it can be connected responsibly.
Related resources