Enterprise SEO — AEO · GEO
Enterprise search was treated as an operating system, not a queue of keywords. The program connected technical health, content planning, local visibility, reputation, structured data, analytics, and field adoption across a regulated insurance business.
The work supported roughly 400 agents across 160 offices while giving marketing, engineering, compliance, analytics, and agency teams a shared way to plan, ship, measure, and improve search demand.
Search became shared infrastructure.
One program connected demand, local trust, publishing, platform quality, and commercial measurement.
Insurance search almost never follows a straight line. Someone might start with a weather question, compare coverage weeks later, look for a nearby agent, read some reviews, start a quote, then finish it on the phone or in an office. All of that is one demand system, even though it crosses different teams, platforms, and reports.
I built one structure to cover that whole system: technical standards to keep pages discoverable, intent maps to shape what content got made, consistent office and agent info across local listings, a real process for handling reviews, and analytics that tied search behavior to actual leads and sales — not just rankings.
Eight connected workstreams
Each discipline had its own craft — the value came from making them work together.
The program was built to hold the whole system.
Every workstream got a backlog, an owner, a quality gate, and a way to measure it — so search worked like normal product and marketing delivery, not a side request channel.
Technical foundation
Crawl paths, indexation, rendering, canonicals, redirects, sitemaps, Core Web Vitals, templates, releases, and continuous monitoring.
Content and information architecture
Topic and intent maps translated demand into durable page families, editorial briefs, internal links, and refresh priorities.
Local presence
Office and agent discovery connected Business Profiles, location pages, citations, service areas, hours, and ownership changes.
Reputation operations
Review requests, monitoring, response standards, complaint escalation, and reporting protected trust without turning replies into canned scripts.
Structured data and entities
Organisation, office, agent, product, article, breadcrumb, and FAQ facts were modeled from visible content and maintained with the page lifecycle.
Answer-engine readiness
Clear answers, named expertise, sourceable facts, strong internal relationships, and maintained pages supported both classic search and AI-assisted discovery.
Analytics and attribution
Search Console, GA4, Business Profile, call, form, quote, and CRM signals were joined into a funnel view that teams could act on.
Governance and adoption
Shared definitions, release checks, reporting rhythms, decision logs, and training let the program work across teams and 160 offices.
A program lead with a cross-functional delivery system.
My role sat between strategy and execution — setting the search architecture, turning it into work teams could actually pick up, and staying close enough to the implementation and the data to catch bad assumptions early.
- Search strategy, opportunity model, and multi-quarter roadmap
- Technical standards, audits, acceptance criteria, and release risk
- Intent architecture, content portfolio, briefs, and internal-link model
- Local and reputation operating model across agents and offices
- Structured-data and answer-readiness standards
- Measurement taxonomy, funnel definitions, dashboards, and executive narrative
- Prioritisation, stakeholder alignment, and cross-team decision making
- Marketing and subject-matter experts shaped useful, accurate content
- Engineering and platform teams implemented templates, performance, schema, and monitoring
- Legal and compliance reviewed regulated claims and public responses
- Analytics and CRM teams connected onsite behavior to downstream outcomes
- Agency operations maintained local facts and supported field adoption
- Leadership resolved priorities, resourcing, and policy decisions
Enterprise search operating system
A single-page map connecting the eight workstreams, their owners, shared inputs, release gates, and business outcomes.
1600×1000px · Figma diagram · anonymised owner labelsThe funnel was mapped before the dashboard
TOFU, MOFU, and BOFU were treated as connected customer states—not reporting labels.
From an unanswered question to a measurable relationship.
Each stage had its own search intent, its own job for the page, its own next action, and its own way to measure success — so traffic that had nowhere useful to go never got mistaken for a win.
- TOFU · DiscoverTop of funnel
Coverage education, seasonal risk, claims guidance, community questions, and plain-language explainers met people before they had chosen a product or provider.
- Non-brand impressions
- New organic users
- Engaged visits
- Topic reach
- MOFU · EvaluateMid funnel
Product pages, comparison guidance, FAQs, calculators, agent profiles, and office pages helped people judge fit, trust, location, and the next practical step.
- Return visits
- Product engagement
- Agent views
- Quote starts
- BOFU · ActBottom of funnel
Quote, contact, appointment, click-to-call, and agent-routing paths reduced ambiguity at the point of intent and preserved source context into downstream systems.
- Form completion
- Qualified leads
- Quote completion
- Policy sales
- Post · ContinuePost-conversion
Service, claims, review, referral, and renewal journeys carried search value beyond acquisition and created better feedback for content and local teams.
- Service success
- Review velocity
- Retention
- Assisted value
A page system organised around real decisions.
Demand → page family → next useful action
I grouped keywords into a small set of durable intent families, each with its own page type — so no single URL had to educate, compare, localize, and convert all at once.
01 Learn
Understand risks, terms, coverage, claims, and everyday insurance decisions.
- Guides
- Glossary
- Claims help
- Seasonal advice
02 Evaluate
Compare options, understand fit, resolve objections, and build confidence.
- Product pillars
- Comparisons
- FAQs
- Calculators
03 Find local help
Locate a trusted office or agent with accurate availability and service information.
- Office pages
- Agent profiles
- County hubs
- Directions
04 Take action
Start a quote, request contact, call an office, or schedule a conversation.
- Quote paths
- Contact forms
- Call actions
- Appointments
05 Get service
Handle a claim, make a payment, find documents, or get policy support.
- Claims
- Payments
- Policy service
- Emergency help
The structure stayed flexible — new query data, onsite behavior, agent questions, or sales feedback could shift priorities without needing a new taxonomy every quarter.
Intent-to-funnel opportunity map
A working view of TOFU, MOFU, BOFU, and service demand mapped to page types, next actions, events, and downstream outcomes.
1600×900px · workshop board or planning sheet · redact query volumesMake the right pages easy to find, understand, and trust
Technical SEO became part of platform quality, release work, and observability.
The foundation was managed in six layers.
Every audit turned into an owned backlog, not a slide deck. I grouped issues by failure mode, business impact, how many templates they touched, and who was actually positioned to fix them once.
- Discovery
Crawl paths and directives
Robots rules, internal links, XML sitemaps, status codes, redirects, and orphan checks kept useful URLs discoverable while controlling low-value crawl paths.
- Selection
Indexation and canonical control
Canonical rules, parameter handling, duplicates, thin states, pagination, and environment controls made the intended URL explicit.
- Rendering
HTML, JavaScript, and templates
Critical text, links, metadata, and structured data remained available in rendered output. Shared templates solved defects at scale.
- Experience
Performance and page quality
Core Web Vitals, image delivery, script cost, mobile behavior, accessibility, and stability were treated as customer and search quality together.
- Meaning
Metadata and structured data
Titles, descriptions, headings, breadcrumbs, entity facts, and eligible schema were generated from the same visible source rather than maintained as a second truth.
- Control
Release checks and monitoring
Pre-release acceptance checks, Search Console trends, template sampling, alerts, and post-release validation caught regressions before they became quarters of lost demand.
The recurring risks were designed out of the workflow.
Enterprise search problems keep coming back if the delivery system doesn’t change. So for each common failure, I built in a control that actually stuck.
- New templates could not rely on a later SEO cleanup. Crawlability, metadata, canonical behavior, structured data, performance, and analytics were included in acceptance criteria before release.Release gate
- Redirects and content retirement were treated as lifecycle work. Owners, replacement destinations, status codes, internal links, and sitemap removal moved together.Lifecycle rule
- Platform-wide defects were fixed at the component or template level whenever possible, avoiding hundreds of fragile page-by-page patches.System fix
- Vendor tags and campaign requests entered through a performance budget, so conversion measurement did not quietly degrade the experience it was meant to improve.Budgeted
Search health and release monitoring
Trend view for crawl failures, index coverage, Core Web Vitals, template regressions, structured-data errors, and remediation ownership.
1440×900px · dashboard screenshot · obscure URLs and internal ticketsTemplate-level technical audit
Representative audit screen showing how issues were grouped by template reach, severity, customer impact, and responsible delivery team.
1440×900px · spreadsheet, crawler, or backlog view · privacy-safe cropEvery page had a reason to exist and an owner after launch.
The content system combined search data with what customers and field agents actually knew. Legal review stayed part of the process — without becoming the place where clarity and usefulness went to die.
- OpportunityDemand gaps, customer questions, agent feedback, competitor patterns, onsite search, and business priorities became one ranked opportunity set.SEO + product
- BriefEach brief named the audience, intent, page job, evidence, expert input, internal relationships, next action, and measurement plan.SEO + content
- DraftWriters worked from subject-matter interviews and source material, using plain language while preserving the precision required for insurance decisions.Content + SME
- ReviewBrand, product, legal, and compliance reviewed the claims relevant to them through a defined path instead of parallel email threads.Compliance
- PublishMetadata, links, schema, analytics, accessibility, and indexability shipped with the page, not as a follow-up checklist weeks later.Platform + SEO
- MaintainPerformance, age, product changes, broken journeys, local feedback, and search shifts triggered refresh, consolidation, or retirement decisions.Named page owner
I didn’t measure content quality by how much got published. I measured it by whether it helped someone make their next decision with less uncertainty.
Content portfolio and refresh decisions
Portfolio view separating create, improve, consolidate, redirect, and retire decisions by intent, performance, age, risk, and commercial contribution.
1600×900px · dashboard or planning board · anonymise page URLs if neededSearch-led content brief and review workflow
A real brief showing the customer question, evidence, subject-matter input, internal links, conversion role, schema notes, and approval state.
1200×900px · document or CMS workflow · remove names and unpublished claimsA distributed brand needed one reliable local truth
The program supported roughly 400 agents across 160 offices without erasing local ownership.
Local search connected digital records to real offices and people.
This covered profiles, on-site location pages, citations, reviews, and operational changes. Accuracy mattered a lot here — a wrong phone number or office status has an immediate, real consequence for a customer.
- Profiles
Google Business Profile governance
Business groups, access roles, categories, services, hours, photos, links, attributes, duplicate handling, and bulk quality checks were managed through a common standard.
- Entities
Office and agent relationships
Office pages and agent profiles expressed who worked where, which services were available, and how entity changes should flow across search surfaces.
- Citations
Consistent external records
Core directories, data aggregators, insurer references, maps, and community listings were audited against the approved name, address, phone, and URL source.
- Reputation
Review requests and response care
Review opportunities, monitoring, human response guidance, privacy boundaries, complaint escalation, and response-time reporting formed one reputation practice.
- Change control
Openings, moves, and departures
Office changes, temporary hours, agent moves, closures, and duplicates followed a checklist so customer access and accumulated trust were not casually discarded.
- Offline loop
Field signals improved search
Agent questions, call reasons, office feedback, community activity, and review themes fed content priorities and exposed gaps that search tools alone could not see.
Reputation management stayed human at scale.
The goal was consistent care, not copy-pasted replies. Clear boundaries let local teams sound like actual people while still protecting customer privacy and staying within regulatory bounds.
- Positive reviews received specific acknowledgement rather than a repeated template. The response reflected the experience described without introducing policy or personal details.Human response
- Service complaints moved into a private recovery path with a named owner. Public replies acknowledged the concern and explained the next safe step without debating the customer.Escalation path
- Suspicious or policy-violating reviews were documented and flagged through the platform process; they were not answered with accusations or improvised legal language.Policy route
- Themes were reported across offices without turning star rating into a vanity contest. Recurring friction became input for service, content, and operational improvements.Learning loop
Office and agent profile health
Estate-level scorecard for ownership, verification, NAP accuracy, duplicate risk, hours, categories, review coverage, and unresolved exceptions.
1600×900px · local platform or BI dashboard · blur personal contact detailsLocal actions and review themes
A paired visual combining office-level calls, directions, and website actions with review volume, response time, and recurring service themes.
1600×900px · map plus trend lines · aggregate low-volume officesAnswer readiness began with better publishing
No magic markup—just accessible pages, clear facts, attributable expertise, and maintained relationships.
The same foundation served classic results and AI-assisted answers.
I treated AEO as a quality layer on top of technical SEO, content, and entity management — not a second publishing system built just for answer engines.
- Answer design
Lead with the useful answer
Important questions received direct, plain-language answers before supporting detail. Headings and page sections made the information easy to locate without flattening the nuance.
- Evidence
Name expertise and sources
Subject-matter review, author or reviewer context, update dates, definitions, and sourceable facts made important insurance guidance easier to verify.
- Entities
Keep people, places, and products aligned
Organisation, office, agent, coverage, article, and service relationships stayed consistent across visible pages, internal links, local profiles, and structured data.
- Markup
Schema followed the visible page
Structured data described facts customers could see. It was validated, monitored, and removed when the supporting content no longer existed.
- Retrieval
Build strong internal pathways
Related questions, product hubs, local entities, expert pages, and supporting guidance were deliberately linked so both people and retrieval systems could follow the topic.
- Maintenance
Freshness meant factual care
Refresh work followed product, legal, service, and demand changes. Dates changed only when the underlying guidance was genuinely reviewed or improved.
Entity and structured-data quality view
Representative page, schema output, validation result, and entity relationship map shown together to demonstrate visible-content parity.
1600×1000px · browser + validator composite · use a public or anonymised URLReporting moved from rankings to decisions
Visibility, behavior, local action, leads, and sales were read as one chain—with its limitations made explicit.
Each layer answered a different business question.
No single tool told the whole story. Search Console showed search exposure, GA4 showed onsite behavior, Business Profile showed local actions, and CRM data closed the loop on what actually sold.
- AvailabilityCan the intended pages be crawled, rendered, indexed, and served with a healthy experience?
- VisibilityWhich query groups, pages, devices, locations, and search appearances are gaining or losing impressions, clicks, and click-through rate?
- EngagementDo organic visitors find the promised answer, continue to a useful next page, return, or abandon the journey?
- DemandHow many people view, start, and complete a quote or contact path—and where does meaningful drop-off occur?
- Commercial valueWhich leads, quotes, appointments, and policy sales can be connected to organic and local journeys, including assisted paths?
- Local actionWhich offices and agents earn profile views, calls, directions, website visits, reviews, and qualified onsite actions?
- ReputationAre review volume, response time, unresolved themes, and customer sentiment improving in ways operations can act on?
- Content healthWhich pages should be expanded, refreshed, consolidated, redirected, or retired based on demand, quality, age, and contribution?
The same data was edited for different decisions.
A practitioner, a product owner, an agency leader, and an executive don’t need the same dashboard. They do need the same definitions, and a clear path from signal to action.
Weekly working view
Anomalies, release impact, indexation, broken journeys, local exceptions, and urgent content or reputation actions.
Monthly program view
Demand movement, funnel performance, office patterns, completed work, experiments, risks, and the next set of decisions.
Quarterly leadership view
Commercial contribution, portfolio movement, capacity trade-offs, enterprise dependencies, and investment choices.
Field feedback view
Local visibility, profile actions, review themes, customer questions, and office-specific corrections returned to agency teams.
Organic journey from visibility to policy action
Full-funnel view joining impressions, visits, useful engagement, quote starts, leads, calls, appointments, and downstream sales with attribution caveats.
1600×1000px · GA4 + Search Console + CRM composite · indexed values acceptedQuarterly search decision brief
Leadership view of growth, portfolio movement, local health, shipped work, risks, capacity constraints, and the decisions required next.
1600×900px · slide or BI export · remove confidential revenue valuesSearch work moved through the organisation, not around it
Shared language and visible decisions made cross-team delivery faster and less fragile.
Every team received a useful contract with the program.
I translated search into whatever language each team actually cared about: clarity for customers, quality for the platform, confidence for compliance, usefulness for the field, and measurable demand for leadership.
- Marketing and content
“A prioritised portfolio, useful briefs, clear search intent, and evidence for refresh decisions.”
Shared editorial roadmap - Engineering and platform
“Reproducible defects, template-level requirements, acceptance criteria, and release priority.”
Product backlog + release gates - Legal and compliance
“Visible claims, named evidence, controlled reviews, and a traceable approval path.”
Defined review contract - Analytics and CRM
“Stable event definitions, identity boundaries, source preservation, and reconciled outcome data.”
Shared measurement dictionary - Agency operations
“Accurate local facts, simple exception reporting, useful review guidance, and visibility into field impact.”
Local ownership model - Leadership
“A concise account of value, risk, capacity, trade-offs, and the decisions only leadership could make.”
Quarterly decision narrative
The calendar protected the work from becoming reactive.
A regular cadence gave urgent defects, planned improvements, local maintenance, content decisions, and leadership trade-offs each their own place — instead of forcing all of it into one meeting.
- ContinuousMonitoring, critical technical issues, inaccurate office information, severe reputation concerns, and release regressions.On-call owner
- WeeklyWorking backlog, shipped changes, blockers, data anomalies, content reviews, and local exceptions.Program team
- MonthlyFunnel movement, content portfolio decisions, office and review health, experiments, and cross-team dependencies.Workstream leads
- QuarterlyRoadmap, commercial contribution, platform priorities, resource choices, and policy decisions.Leadership
- AnnuallyMeasurement definitions, governance roles, tool value, access controls, local estate, and portfolio-wide content health.Program sponsor
Cross-team workflow and decision rights
RACI-style view showing intake, prioritisation, content review, engineering acceptance, local exceptions, release validation, and reporting ownership.
1600×1000px · process diagram · role names rather than individual namesThe value reached beyond a search report
Customers found clearer paths; teams gained reusable systems; leadership gained better evidence.
Growth and operating value reinforced each other.
The growth numbers mattered. So did the less visible part — the infrastructure that made those gains repeatable across content, technology, local operations, and reporting, instead of a one-time spike.
- Organic demandTop-funnel organic traffic grew 400% year over year as technical coverage and the content portfolio expanded together.
- Commercial actionBottom-funnel sales grew 60–70% year over year, keeping the program accountable to qualified action rather than traffic alone.
- Answer visibilityMore than 60 featured-answer placements extended reach for useful insurance guidance and exposed new content opportunities.
- CustomersPeople gained clearer educational, local, quote, and service paths across the moments when insurance decisions are often least familiar.
- Agents and officesRoughly 400 agents across 160 offices benefited from a more reliable local presence, clearer routing, reputation support, and shared maintenance standards.
- MarketingThe content team gained a demand-led portfolio, reusable briefs, editorial governance, and stronger evidence for where to invest or consolidate.
- EngineeringSearch requirements became platform acceptance criteria, shifting recurring defects from page-level cleanup to durable template and component fixes.
- EnterpriseLeaders gained one decision framework connecting visibility, customer behavior, local action, leads, sales, delivery capacity, and risk.
The program left behind habits, not just outputs.
These principles remained useful as search surfaces, business priorities, and tools changed.
- What worked
Fix systems before pages
Template, component, workflow, and data-source improvements created broader and more durable gains than repeating the same correction across individual URLs.
- What worked
Measure the customer path
Search metrics became more useful when read beside onsite behavior, local actions, lead quality, sales outcomes, and the time between those events.
- Looking forward
Keep local ownership close to the field
Central standards protected quality, while agents and office teams supplied the operational truth that profiles, pages, and review practices depended on.
- Looking forward
Earn answer visibility through clarity
Accessible technical delivery, useful writing, attributable expertise, visible evidence, and consistent entities remained a stronger strategy than chasing speculative AI shortcuts.
- What I'd change
Retire weak work sooner
A mature content portfolio needs consolidation and removal as much as creation. Clearer retirement thresholds would have reduced maintenance load earlier.
- What I'd change
Join downstream data earlier
Bringing CRM, call, and sales definitions into planning sooner shortens attribution debates and improves the event design before implementation begins.
This worked because search stopped being a specialist’s report nobody read, and became a shared way for the whole business to improve how it’s found, understood, and chosen.