Our Approach

How an Independent Audit Works

A structured, evidence-led review of how search and acquisition effort, people and budget connect to business outcomes. Every conclusion is traceable to its evidence and labelled by confidence.

The Framework at a Glance

Detailed audit workflow

Nine stages in three phases: agree the questions and safeguards, collect and test the evidence, then turn findings into ranked priorities and a roadmap.

  1. Discovery and success criteria The questions the review must answer.
  2. Access request and safeguards Read-only, least-privilege access.
  3. Evidence collection Plans, work records, contracts and data.
  4. Interviews and deliverable review Context, then delivered vs planned.
  5. Cost and resource mapping Time, fees and tools by workstream.
  6. Cross-checking data sources Search Console, analytics, ads and CRM.
  7. Findings with confidence labels and limitations Verified, Supported or Hypothesis.
  8. Impact and effort priorities Ranked by likely business impact.
  9. Management readout and roadmap Owners, next steps and sequence.

The same sequence applies whether we are reviewing an in-house SEO team, an agency relationship, a lead generation programme or paid acquisition. Scope and depth are agreed at the start.

Stage by Stage

1. Discovery and success criteria

We start by agreeing what the review must answer. Typical questions: is SEO spend proportionate to its contribution, is the agency delivering what was contracted, are priorities aligned with revenue? We document the business goals, the period under review and what “good” should look like, so findings are judged against agreed criteria rather than opinion.

2. Access request and safeguards

We request only the access needed to answer the agreed questions, preferably read-only, and confirm confidentiality terms before any data is shared. See the safeguards section below.

3. Evidence collection

We gather the material that shows what was planned, what was done and what changed: strategies, roadmaps, task trackers, reports, contracts and statements of work, and platform data.

4. Interviews and deliverable review

Short structured conversations with the people responsible clarify context that documents miss. We then compare deliverables against what was planned or contracted, looking at quality, relevance and whether the work was implemented.

5. Cost and resource mapping

We map internal time, agency fees, tools and other spend against the main workstreams. The aim is to see where resources go, not to produce a statutory financial audit.

6. Cross-checking data sources

Search Console, analytics, ad platforms and CRM data are compared with each other and with reported results. Discrepancies are often as informative as the numbers themselves.

7. Findings with confidence labels and limitations

Each finding states the evidence behind it, a confidence label and any limitation, for example missing history, sampling or attribution gaps.

8. Impact and effort priorities

Recommendations are ranked by likely business impact and the effort or cost to act, so quick, valuable changes are visible alongside longer-term work.

Prioritization matrix: impact × effort

Findings sorted into four groups. Quick wins come first, major projects are planned and resourced, fill-ins are batched, and low-impact, high-effort work is deprioritized.

Example visualization · Illustrative sample

Business impact →

Quick wins — high impact, low effort

  • Fix indexation of key commercial pages
  • Assign owners to unowned backlog items

Major projects — high impact, high effort

  • Rebuild internal linking by topic
  • Restructure roles and reporting lines

Fill-ins — low impact, low effort

  • Tidy metadata on older pages
  • Merge duplicate dashboards

Deprioritize — low impact, high effort

  • Rewrite low-traffic archive content
  • Build custom reports nobody uses

Effort →

Example items only. Real findings are placed using the evidence and confidence label behind each one.

9. Management readout and roadmap

We present findings to leadership in plain language, then provide a written report and a sequenced roadmap with suggested owners and next steps.

Typical Evidence Sources

No single source proves performance on its own. We use several and are explicit about what each can and cannot show.

Evidence sources behind each finding

Every finding draws on four groups of sources: platform data, commercial records, work records and interviews. A finding supported by more than one group earns a higher confidence label.

Finding

Evidence, confidence label, limitations

Platform data

  • Google Search Console
  • Web analytics
  • Ad platform data

Commercial records

  • CRM records
  • Contracts and invoices

Work records

  • Task trackers and roadmaps
  • Reports and deliverables

People

  • Interviews with the people responsible
  • Corroborated against data or documents
SourceWhat it can showWhat it cannot prove on its own
Google Search ConsoleSearch impressions, clicks, queries and indexing signals for GoogleRevenue, conversions, or performance in other search engines
Web analyticsSessions, behaviour and conversions as configuredAccurate attribution if tagging, consent or channel rules are incomplete
Ad platform dataSpend, delivery and platform-reported conversionsIncremental impact, or lead quality after the form submission
CRM recordsLead progression, qualification and closed revenue by sourceCorrect source attribution if capture fields are unreliable
Task trackers and roadmapsWhat was planned, prioritised and completedWhether the completed work had any business effect
Reports and deliverablesWhat was communicated to stakeholdersThat the reported figures are complete or correctly defined
Contracts and invoicesWhat was agreed and paid forWhether the agreed scope was delivered to the expected standard
InterviewsContext, constraints and decision historyFacts that are not corroborated by data or documents

Confidence Labels

Every finding carries one of these labels, so readers can tell established facts from informed judgement.

LabelMeaning
VerifiedConfirmed directly in data or documents we reviewed, with no material contradicting evidence
SupportedConsistent with several sources, but at least one link in the chain is inferred or incomplete
HypothesisA plausible explanation that the available evidence cannot yet confirm; includes what would be needed to test it

Confidence labels describe the strength of the evidence, not the size of the problem. A hypothesis can still be worth testing if its potential impact is high.

Safeguards

An audit reviews processes, outputs and spend. It is not covert employee surveillance, and it is conducted openly with the knowledge of the people involved.

  • Data minimisation: we request and retain only what is needed to answer the agreed questions.
  • Confidentiality: NDAs or confidentiality terms are put in place as agreed before access is granted.
  • Least privilege: read-only access wherever platforms allow it, removed at the end of the engagement.
  • No individual employee data unless it is strictly relevant to the agreed scope and authorised by the client.
  • No credentials by email or forms: access is granted through platform user management, not shared passwords.
  • AI with human review: where AI helps (summarizing data, clustering queries, drafting report text), a person checks every output against the source before it becomes a finding, and client data is only processed with tools and terms you approve.

What an Audit Does Not Do

Being clear about limits keeps the findings credible and useful.

  • It does not certify performance or provide statutory financial assurance.
  • It does not assign blame to individuals; it assesses processes, decisions, outputs and spend.
  • It does not guarantee that following the roadmap will produce a specific ranking, lead volume or return.
  • It does not replace your own judgement on trade-offs that depend on strategy, budget or risk appetite.

Sample Deliverable Fields (Blank Template)

Template only. This shows the structure of a findings register; it contains no client data, findings or results.

FindingEvidenceConfidenceImpactEffortOwnerNext step
———————
———————
———————

The final report also includes the agreed questions, scope and period, sources reviewed, limitations, and a sequenced roadmap.

Start With the Questions You Need Answered

Tell us what you want to verify. We will propose a scope that fits your situation before any access is requested.