Skip to content
Martin Bangho
Menu

How I work with GEO and SEO evidence and measurement.

Here I explain the evidence I use in GEO and SEO work, how I check findings and how I evaluate changes.

Discuss your brief

GEO and SEO: a shared foundation, different evaluation

SEO helps search engines discover, process and display relevant pages. GEO focuses on how a brand and its content are represented and interpreted in generative answers. It is not simply a new name for SEO and does not replace its technical or content foundations.

For AI, we examine answers, sources and factual accuracy on an agreed set of questions. For SEO, we use indexation, query and organic traffic data. Changes in AI answers alone do not prove revenue growth; rankings and citations cannot be guaranteed.

GEO (Generative Engine Optimization), AEO (Answer Engine Optimization), LLMO and AI SEO describe overlapping approaches, not four separate products. Here the job is external discoverability and accurate representation of your business, not internal AI automation or producing a quota of articles. This work suits businesses able to verify facts and implement changes; it is not a fit if the requirement is a guaranteed recommendation or a fixed number of citations.

Shared foundation, different measures

Accessible website · content · verified facts
SEO
Indexation · queries · organic traffic
GEO
Answer · source · factual accuracy

The specific system selects sources; a website change does not control selection.

Illustrative methodology model

From available evidence to delivery priorities

Audit and evidence

I first check whether your existing audit or strategy is still useful and current. I build on what we can use, fill the gaps and propose a new analysis when evidence is missing. The review covers your website and current activities: technical health, content, consistent brand information, authorship, links and mentions, and available AI visibility data. I use competitor comparisons to identify relevant gaps rather than apply a generic checklist.

Priorities and ownership

Findings lead to agreement on priorities, ownership and measurement. For example, a discrepancy between a business profile and the website needs a source of truth, an owner for the correction and a check afterwards. This illustrates the method, not a client result. Scope follows the actual situation, rather than a fixed package.

Audit inputs and outputs

For the audit, bring the website, target markets and languages, customer questions, competitors and any authorised read-only exports. I assess availability to crawlers, entity clarity, evidence and whether important passages answer the question in context. The output is a prioritised roadmap with owners, dependencies and a proposed measurement protocol. Existing raw AI answers can be analysed; a protocol alone is not a measured baseline.

Evidence into a priority map

AvailabilityEntity clarityVerifiable evidenceAnswer usefulness
Higher impact / lower effortReview first
Higher impact / higher effortPlan an initiative
Lower impact / lower effortAssess benefit
Lower impact / higher effortConsider deferring

For a finding: what · why · owner · verification. Missing evidence is not zero.

Illustrative methodology model

Delivery with or without an in-house team

My involvement

With an in-house team, I take ownership of agreed initiatives and work alongside marketing, content and engineering.

Without an in-house team

Without an in-house team, I develop the strategy and begin the agreed work.

Agreed delivery scope

We define my delivery scope and suppliers’ responsibilities beforehand.

Your facts and source material

Your team confirms business facts, subject expertise, rights to source material and access.

Delivery owners and capacity limits

Development, content production and any PR distribution need named owners; unlimited delivery capacity or secured media coverage are not included.

Agree roles before delivery

With an internal teamAgreed initiatives alongside marketing, content and engineering

Without an internal teamMy delivery and specialist work within agreed scope

Business facts and rights
Confirmed by your team
Implementing a change
An owner agreed beforehand
Review and acceptance
Agreed evidence and responsibility
Illustrative methodology model

Methods chosen for the problem, not separate packages

  • Technical SEO: accessibility to crawlers, indexation, speed and structured data
  • Content: topics and queries, content gaps, site structure, metadata and internal links
  • Authority: relevant links, digital PR and credible mentions, distinguishing paid placements from independent editorial coverage
  • Evaluation: agreed metrics, data limitations and next priorities

Four lenses on one page

A page answering a question

A direct answer in context

Documented source and authorship

Relevant related content

Availability
Can content be fetched and processed?
Content
Does it answer with evidence?
Authority
What does an independent source support?
Evaluation
How do we verify the specific change?
Illustrative methodology model

Make business information consistent and verifiable

Fact sheet and discrepancy register

The review brings names, services, contact details, About information, authorship and source evidence into a fact sheet and discrepancy register.

Source, date and owner

Important facts need a source, date and an owner responsible for updates. Website text, relevant external profiles and structured data should describe the same entity.

Independent evidence and rights

An owned profile is not independent corroboration. Original evidence, credible mentions and relevant digital PR support trust, within verified scope and with clarified rights.

No platform guarantee

Neither a Wikipedia page nor a Knowledge Panel is guaranteed. My role, source evidence and its limitations are described in the Gen Digital case study. The published SEO overview is not evidence of increased AI visibility.

Fact, provenance and ownership

Brand / entity
Owned websiteProfile and authorIndependent evidence
Fact
Source · date · update owner
Discrepancy
What differs · where to correct · how to verify

Solid connection: owned representation. Dashed: independent provenance, not automatic corroboration.

Illustrative methodology model

Six answer surfaces, six separate observations

Observed surfaces

ChatGPT, Gemini Apps, Google AI Overviews, Google AI Mode, Perplexity and Microsoft Copilot are assessed separately.

Questions and sources

We define relevant questions and observed sources for each selected surface rather than pooling them into one ranking or assuming coverage of every product.

Recorded conditions

Answers depend on the product, retrieval mode, available sources, language, location and observation time.

Limits of control

Improving access, clarity and evidence does not control whether a platform retrieves, cites or recommends a page.

Source list versus citation

A link in a source list alone does not prove an inline citation.

Training and guarantees

Search visibility is not proof of model training, and citations and positions cannot be guaranteed.

Separate surfaces, separate observations

ChatGPT
Mode · language · location · time
Answer / sources: record separately
Gemini Apps
Mode · language · location · time
Answer / sources: record separately
Google AI Overviews
Mode · language · location · time
Answer / sources: record separately
Google AI Mode
Mode · language · location · time
Answer / sources: record separately
Perplexity
Mode · language · location · time
Answer / sources: record separately
Microsoft Copilot
Mode · language · location · time
Answer / sources: record separately

Citation ≠ mention ≠ recommendation. An API is not the consumer interface.

Illustrative methodology model

Measure answers, not a universal AI ranking

Question panel

We agree a question panel around discovery, comparisons, branded facts and buying decisions.

Recorded conditions

The protocol records the exact prompt, surface, model or version where available, language, location, search mode, date and repetitions.

Traceable sources

Raw answers and their sources must remain traceable; missing observations stay unknown.

Separate measures

Brand mentions, linked citations, recommendations and factual errors are different measures.

Denominators

Each share needs an explicit denominator and repeatable collection conditions.

SEO and business metrics

We evaluate SEO indexation, organic queries and traffic separately, then connect available AI referrals and enquiries to business goals without claiming that answer visibility proves revenue.

Collection scope

Live collection, monitoring frequency and any service commitment depend on a verified collection method, permissions, capacity and an agreed budget.

Consumer-interface limitation

An API response is not evidence of what a customer saw in a consumer interface; no automatic baseline or consumer-interface monitoring SLA is included by default.

The same protocol, repeated observations

ProtocolQuestions · surface · mode · language · location
Baseline observation
  • Time and condition version
  • Raw answer and sources
  • Mentions, citations, factual errors
Repeated observation
  • Time and condition version
  • Raw answer and sources
  • Mentions, citations, factual errors
Compare only with stated limitationsAn answer change alone does not establish business impact.
Illustrative methodology model

Local discovery: accurate facts, relevant questions

For location-dependent decisions, the audit checks consistency of business name, address and phone details, Google Business Profile information and genuinely local customer scenarios. This is a scoped review, not a promise to manage a location network or manufacture near-identical city pages.

Consistent facts for a real location

Real locationName · address · phone
Website
Name: verify consistency
Address: verify consistency
Phone: verify consistency
Business profile
Name: verify consistency
Address: verify consistency
Phone: verify consistency
Other relevant sources
Name: verify consistency
Address: verify consistency
Phone: verify consistency

Correct discrepancies at the source. Do not create fictitious locations.

Illustrative methodology model

Product data: audit before integration

For ecommerce, the review covers product detail pages, feeds and structured data for consistency, accessibility and evidence. Names, specifications, prices and availability should come from the same current source. Recommendations identify what needs correction and who must implement it. Feed engineering, merchant integrations or checkout inside an assistant require a separately agreed technical scope and suitable delivery resources; they are not included as a ready-made integration.

One product, consistent data

Verified product sourceIdentity · price · availability
Visible pageThe same product and current facts
Structured dataThe same product and current facts
Product feedThe same product and current facts

Mismatch → source of truth → correct representations → recheck. Integration follows an audit and agreement.

Illustrative methodology model

Scope and price follow the brief

Agreement before work starts

We agree scope, budget, handover and review frequency beforehand.

Mentoring is not a minimum

The four-hour starting scope for mentoring is not a minimum for this engagement.

Collaboration conditions

We start by mapping your customers’ journey: how they find you, what information they lack, which feedback has no one to receive it and where their experience could improve. I call this initial review screening. Ideally, we combine it with an audit and use the findings to guide agreed follow-on work. Scope, responsibilities, budget and billing are confirmed in our agreement. Submitting the form does not itself start paid work.

All outputs are handed over to your team in the agreed format. Further sharing requires Martin Bangho’s consent. Rights to third-party tools and components remain subject to their licences; handing over outputs does not transfer those rights. Rankings, AI citations and the timing of business results are not guaranteed.

  • Scope: Audit and agreed follow-on GEO and SEO initiatives, direct delivery or collaboration with your in-house team.
  • Price: scope and budget are confirmed before work starts.
  • Who it is for: Businesses that need to connect AI and search visibility with concrete priorities and delivery.
Back to GEO and SEO collaboration

Let’s discuss your project.

Tell me what your company is working on and where you need help. You do not need a finished brief.

Send me a message

A few lines about what your company needs are enough. You don’t need a finished brief. All fields are required.

By submitting you agree to the processing of personal data for the purpose of handling your enquiry. Privacy policy