SingRank

Method

Clear evidence behind the work we recommend.

We review your available search, website and enquiry data, investigate the problems and explain what to prioritise. Measurements, estimates and assumptions are labelled so you can understand the basis of each recommendation.

How we interpret the evidence

Use relevant sources We use connected first-party data and verified research relevant to the question. We check the reporting period and coverage before interpreting a result.

Distinguish unknown from zero Missing data does not prove that nothing happened. If a source is unavailable or incomplete, we explain the gap.

Separate findings from assumptions A direct observation, a supported explanation and a hypothesis have different levels of confidence. We make that distinction clear.

Our approach to links

We do not buy backlinks for ourselves or clients. Our work focuses on useful content, clear website structure and relevant internal links.

We do not invent backlink counts or assume that more content alone can solve every competitive challenge. Recommendations depend on the evidence available.

What we actually measure

  • Google Search Console — Available search impressions, clicks, queries and page performance for a connected property. Reporting coverage and dates matter.
  • Microsoft Clarity — Recorded website behaviour can help identify usability issues. It does not reveal every visitor’s intentions.
  • Crawler logs — Available server or platform logs can help investigate crawler requests and access failures. A request does not prove an AI recommendation.
  • AI referral tracking — Identifiable referral information can show visits from AI tools. Some sources are not shared and remain unknown.
  • Contact actions and enquiries — Supported contact events and recorded enquiries are reviewed with their available page and channel context. They are distinct from completed sales.
  • Content inventory — Indexed content helps us compare proposed topics with existing pages and identify overlap or gaps. We check coverage before treating an inventory as complete.

How scores help us prioritise

Scores help organise a review. They are decision aids rather than guarantees, and we check the underlying evidence before acting.

Changes in traffic

Compare the same metric across comparable periods

We check whether a change is sustained, whether enough data is available and whether tracking or site conditions changed.

Content freshness

Review age alongside accuracy and performance

A page may need an update when information changes or its usefulness declines. Age alone is not a reason to rewrite it.

Search opportunity

Demand + relevance + competition + potential enquiries

We prioritise searches that fit the business and have useful evidence behind them. Traffic potential is considered alongside buying intent.

Recovery priorities

Investigate the cause before choosing the fix

Technical problems, content changes and competition call for different responses. We explain the evidence before recommending work.

Competing pages

Compare search intent and page overlap

Two similar pages are not automatically a problem. We examine what each page does before deciding whether to strengthen, link or consolidate them.

Confidence in a finding

Confirmed / Likely / Hypothesis

Confirmed findings have direct evidence. Likely explanations have corroborating signals. A hypothesis needs further checks.

How reporting stays useful

Data sources refresh on different schedules. We review the available reporting dates rather than describing every figure as real-time.

  • Search and website data — Reviewed according to the collection schedule and source availability.
  • Content priorities — Revisited as new performance data and relevant research become available.
  • Technical and market checks — Run on their applicable schedules or when a specific problem needs investigation.

An example of investigating crawler access

Our recorded 12 August 2026 review found failed AI crawler requests on this website. Comparing crawler requests with server responses gave us a specific access problem to investigate. A ranking chart alone would not explain that failure.

Know what we recommend and why

You can ask where a figure came from, what period it covers and what it can actually tell you. Our recommendations should make the next decision clearer.

The engineering behind all this → · The pixel that does the logging → · GEO — AI search, measured →

Get a free visibility check