Research Methodology
SEO is an evolving field where official documentation, observed behavior,
third-party data, industry research, and speculation often exist side by side.
AF Search uses a practical research framework to distinguish stronger evidence
from interpretation and to provide clearer context around search-related topics.
How AF Search Evaluates SEO Information
Not all SEO information carries the same level of reliability.
We generally evaluate information using several layers of evidence rather than
treating every claim, ranking study, metric, or anecdote as equally reliable.
| Evidence Type | Typical Reliability | How We Use It |
|---|---|---|
| Official documentation | High | Used as the primary reference for documented platform behavior |
| Direct technical observation | High to Medium | Used to understand how systems behave in practice |
| Repeated testing or controlled experiments | Medium to High | Used when methodology and limitations are reasonably clear |
| Third-party datasets | Medium | Used for estimates, comparisons, and broader patterns |
| Industry observations | Medium to Low | Used as supporting context rather than universal proof |
| Individual anecdotes | Low | Used cautiously and not treated as general evidence |
| Rumors or unverified claims | Very Low | Generally avoided unless clearly discussed as speculation |
1. Official Documentation Comes First
When a search engine, analytics platform, browser, or web standard provides
official documentation, AF Search generally treats that material as the best
starting point.
Examples may include documentation related to:
- Crawling
- Indexing
- Structured data
- Robots directives
- Sitemaps
- Search appearance
- Analytics implementation
- Web performance
Official documentation does not always answer every practical SEO question,
but it helps establish what is explicitly documented before interpretation begins.
2. We Separate Documentation From Observation
Something can be observable without being officially documented.
For example, website owners may repeatedly observe changes in crawling,
rankings, indexing, or search result appearance without receiving a complete
explanation from the platform involved.
In these cases, AF Search aims to make the distinction clear.
Documented Information
Information directly supported by official documentation, public statements,
or clearly defined platform behavior.
Observed Behavior
Behavior that can be measured or repeatedly observed but may not have a
complete official explanation.
Interpretation
A reasonable explanation based on available evidence, but not necessarily
confirmed by the platform itself.
Speculation
A possible explanation that lacks enough evidence to be treated as established.
3. Correlation Is Not Automatically Causation
SEO studies frequently compare ranking pages against different characteristics.
A study might show that highly ranked pages often have more backlinks,
longer content, faster performance, stronger brands, or other shared traits.
That does not automatically prove that the measured factor caused the ranking.
For example, successful websites may naturally attract more links because they
are already visible and widely known.
The relationship may therefore work in more than one direction.
AF Search treats correlation studies as useful signals rather than automatic
proof of ranking causation.
4. Third-Party SEO Metrics Are Estimates
SEO platforms create proprietary metrics to help users compare websites,
keywords, links, and search visibility.
These metrics can be extremely useful, but they are not the same as direct
search engine data.
Examples include:
- Domain authority scores
- Domain rating metrics
- Trust or citation metrics
- Estimated organic traffic
- Keyword difficulty scores
- Estimated search volume
- Link quality scores
Different tools use different crawlers, databases, formulas, sampling methods,
and update schedules.
As a result, two tools can produce different values for the same website or
keyword without either value necessarily being incorrect.
5. We Prefer Multiple Signals Over a Single Metric
SEO decisions are usually stronger when multiple sources of information point
in the same direction.
For example, evaluating a page may involve:
- Search Console data
- Website analytics
- Technical inspection
- Search result analysis
- Content quality and relevance
- Internal linking
- Backlink data
- Competitive comparison
A single metric rarely provides enough context to explain overall search
performance.
6. SEO Tests Have Limits
SEO experiments can be useful, but search environments are difficult to fully
control.
During a test, many variables may change at the same time.
- Competitors may update their pages
- Search demand may change
- Search engines may update ranking systems
- New backlinks may appear
- Pages may be crawled at different times
- User behavior may shift
This makes it difficult to treat a single SEO test as universal proof.
AF Search gives more weight to findings that can be reproduced, observed across
multiple cases, or supported by other forms of evidence.
7. Case Studies Are Context, Not Universal Rules
A strategy that produces strong results for one website may perform differently
on another.
Websites vary in:
- Age
- Authority
- Industry
- Competition
- Technical condition
- Content quality
- Link profile
- Audience behavior
Case studies can provide useful examples, but they should not automatically be
interpreted as guaranteed outcomes.
8. Search Results Can Vary
Search results are not always identical for every user or every moment.
Results may vary depending on factors such as:
- Location
- Language
- Device
- Query wording
- Search context
- Index changes
- Search system updates
For this reason, a single manually checked search result should not always be
treated as a complete measure of SEO performance.
9. Search Engine Systems Change
SEO information has a shelf life.
Search engines continuously change crawling systems, ranking systems,
interfaces, search features, policies, and documentation.
Advice that was accurate several years ago may no longer reflect current
behavior.
AF Search may update existing content when significant changes make older
information incomplete or misleading.
10. AI Search Requires the Same Evidence Standards
AI-powered search introduces new terminology and new claims about visibility,
citations, answer engines, and content discovery.
Because this area is developing quickly, strong conclusions can be difficult
to make.
AF Search approaches AI search topics using the same basic framework:
- Begin with documented information where available
- Observe measurable behavior
- Compare results across multiple cases
- Separate evidence from interpretation
- Avoid presenting emerging theories as established facts
How We Describe Confidence
When appropriate, AF Search may use different language depending on how strong
the available evidence appears to be.
| Language | General Meaning |
|---|---|
| Is / Does | Used when information is well established or documented |
| Can | Used when an outcome is possible but not guaranteed |
| May | Used when evidence suggests a relationship but uncertainty remains |
| Appears to | Used for repeated observations that are not fully confirmed |
| May suggest | Used for weaker or emerging evidence |
| Speculation | Used when a claim does not have sufficient supporting evidence |
Our Typical Research Process
The exact process depends on the topic, but AF Search generally follows a
workflow similar to this:
- Define the specific SEO question
- Check relevant official documentation
- Review available technical or search performance data
- Compare multiple credible sources where useful
- Review practical observations or experiments
- Identify important limitations and conflicting evidence
- Separate documented facts from interpretation
- Present the conclusion with an appropriate level of certainty
What AF Search Tries to Avoid
SEO becomes less useful when uncertainty is hidden behind confident language.
We try to avoid:
- Presenting rumors as facts
- Claiming to know undocumented ranking formulas
- Treating third-party metrics as direct Google metrics
- Assuming correlation proves causation
- Generalizing from one website or case study
- Promising guaranteed rankings
- Using outdated SEO advice without context
- Presenting estimates as exact measurements
Evidence Matters More Than Certainty
SEO rarely offers perfect information.
Search engines are complex systems, some ranking processes are not publicly
documented in detail, and the web changes constantly.
The goal of AF Search is therefore not to pretend that every search question
has a definitive answer.
Our aim is to identify what is known, what can be observed, what is reasonably
supported, and where uncertainty remains.