Methodology
How WindspeedSEO turns evidence into prioritized SEO work
This page describes how WindspeedSEO decides what matters, in enough detail that you can disagree with it. Prioritization you cannot inspect is just an opinion with a number attached.
Evidence rules
Every finding WindspeedSEO raises must trace to an observable source. In practice that means one of five things:
- First-party Google data. Search Console for search performance, Analytics 4 for behavior. Treated as the primary record.
- Platform-managed performance data. PageSpeed for lab measurement, CrUX for real-user field data where Google reports it.
- Public observation. What a search engine can see when it visits your site: status codes, canonicals, robots directives, sitemaps, markup, internal links.
- Bounded external intelligence. Keyword, SERP, competitor and backlink data from paid providers, reported as third-party estimates rather than as facts.
- Repeatable AI-visibility measurement. A fixed query set run on a schedule, reported as frequency across runs.
Anything that cannot be traced to one of those is context or opinion, and is labeled as such.
Source ordering, and why it affects your price
Sources are used in a deliberate order: stored and cached facts first, then free first-party Google data, then deterministic computation, then inexpensive model calls, then paid SEO intelligence, and premium reasoning only where it genuinely changes the answer.
This is a cost discipline with a customer-visible consequence. Systems that call paid providers without restraint have to pass that cost on as usage limits or credits. Flat monthly pricing is only sustainable if the cheap sources do most of the work.
How findings are scored
Four axes, applied to every candidate finding.
Change magnitude
How large the movement is relative to that page or query’s own normal variance — not relative to the site as a whole.
Commercial relevance
What the affected pages and queries are worth, drawn from analytics behavior and conversion data where it is configured.
Evidence strength
How many independent signals agree. A decline confirmed by position, clicks and competitor movement outranks one visible in a single metric.
Expected effort
An advisory effort band used for sequencing and workload balancing. It is not a time-tracking claim and is never billed as hours.
Findings that are large, commercially relevant, well-evidenced and tractable rise to the top. Everything else stays visible and stays below.
One ordering rule overrides the score: anything preventing a commercially valuable page from being indexed outranks anything that merely degrades how well it competes. A page that cannot appear at all is not competing.
Verification standard
Deterministic validators run first, always. They answer questions with objective answers: is the change live, does the URL return the expected status, is the canonical correct, is the content indexable, did the redirect chain collapse to a single hop, does the published article exist at the expected address.
Language models may be used to interpret an ambiguous requirement or to explain how to remediate a finding. They are never used to declare that a QA check passed. Where a requirement genuinely cannot be machine-verified, it is marked for manual QA rather than assumed.
Only QA-verified work becomes customer-visible as completed work.
Attribution standard
Observation and causation are reported separately. WindspeedSEO will state what changed, by how much and on which pages. It will claim that a specific change caused a specific result only where the timing and the evidence support it, and it will say when a change is too recent to evaluate.
Seasonality, competitor activity and broad algorithmic movement are named as context — including in months where naming them makes the results look worse.
What this methodology will not do
- Guarantee rankings, traffic or revenue.
- Fabricate or adjust recorded keyword volume or difficulty values.
- Present a third-party provider estimate as a measured fact.
- Report unverified work as complete.
- Present the same generic recommendation regardless of what the evidence shows.
Return to the WindspeedSEO overview, or see the methodology in operation in how WindspeedSEO works.
See the methodology applied to your website
The same evidence rules and prioritization model run on both plans.