SEO Discovery Phase: Turn Diagnosis Into Signed Scope

Updated 2026-09-07 · guide · SEO,discovery,services,sales,roadmap

An SEO discovery phase is a bounded diagnostic process that converts business goals, search evidence, technical constraints, buyer paths, and stakeholder decisions into a feasible roadmap and priced scope. It matters because many engagements fail before delivery starts: the agency assumes growth targets, the client assumes guaranteed rankings, engineering assumptions stay hidden, and the statement of work describes tasks instead of outcomes. By the end of this guide, you should be able to run a discovery that produces a baseline, prioritized findings, measurement rules, risks, dependencies, and scope options a client can sign.

Why discovery is the commercial foundation

Discovery is not a free sales call and not a generic audit dump. It is where you learn whether the opportunity is real, whether the organization can execute, and whether your team is the right fit. A good discovery protects both sides: the client avoids buying activity that cannot produce revenue, and the service provider avoids owning outcomes that depend on undocumented constraints.

The deliverable should be decision-ready. It should not say, “Create more content and improve technical SEO.” It should say which markets, pages, systems, people, and metrics matter; what evidence supports the diagnosis; what can be done in the next 90 days; what the client must provide; and what package makes sense.

This phase connects directly to AI service proposals and SEO statement of work basics. Discovery findings should become proposal assumptions, SOW deliverables, dependencies, and payment boundaries.

When discovery is necessary

A discovery phase is appropriate when:

• The engagement is expected to exceed a low monthly retainer.
• The client has multiple services, markets, or buyer types.
• Past SEO work created unclear ownership or technical debt.
• A migration, replatform, JavaScript rendering change, or redesign is involved.
• Lead quality, analytics, or conversion tracking is uncertain.
• Stakeholders disagree about the primary revenue goal.
• The site has complex service pages, localization, product-led flows, or regulated content.
• The client needs a roadmap before committing to a retainer.

For a small, clearly defined repair task, a paid diagnostic may be enough. But if the engagement involves strategy, production, engineering coordination, or revenue accountability, discovery should be formal.

What discovery is not

Do not confuse discovery with:

• A keyword list. Keywords matter, but they do not prove feasibility or business fit.
• A 200-page audit. Long documents without prioritization delay decisions.
• Free strategy. Free work attracts clients who collect advice rather than execute.
• A guarantee. You can diagnose likelihood and constraints; you cannot guarantee algorithmic outcomes.
• A replacement for onboarding. Discovery defines the opportunity and scope; onboarding operationalizes delivery.

If the client expects a fixed ranking position or a specific revenue number from discovery alone, address that in writing before proceeding.

The core model: five questions

Keep the discovery around five questions.

1. Commercial question: What business result matters, and who judges it?
2. Demand question: Which qualified search and AI visibility opportunities exist?
3. Experience question: Which pages and buyer paths should capture that demand?
4. Execution question: What technical, content, organizational, and tracking constraints exist?
5. Decision question: What scope, sequence, resources, and measurement plan will make progress credible?

Every finding should map to at least one of these questions.

Step 1: Run a scope interview

Start with a structured interview involving the economic buyer and operational owner when possible.

Questions about the business

Questions about success

Questions about execution

• What product or service produces the best margin?
• Which segments do you want more of in the next 12 months?
• Which customers are unprofitable or strategically undesirable?
• What is the current primary acquisition channel?
• What is the average deal size, sales cycle, and close rate?
• What capacity or inventory constraints exist?
• What happened before this project, and why did it stop?
• What would make this engagement successful beyond traffic?
• What leading indicator can you observe monthly?
• What result would make the executive sponsor defend the budget?
• What result would make you consider the project unsuccessful?
• Who receives reports, and who can approve changes?
• Who owns website releases, content, legal review, analytics, and CRM?
• What is the normal approval cycle?
• What technical debt or release freezes exist?
• Which tools are available: Search Console, analytics, CRM, CMS, staging, logs?
• What has been tried before, and what evidence exists?

Write answers with attribution. “Sales wants more enterprise leads” is different from “Marketing wants more MQLs.”

Step 2: Define eligible opportunities

Before auditing every page, define what qualifies as an opportunity. Use three filters.

Business fit

The demand should relate to services or products the client wants to sell. A high-volume topic that attracts students, job seekers, or free-tool users may be irrelevant.

Search and AI fit

Look for evidence of:

• Clear commercial or solution-intent queries.
• Question patterns buyers use before sales contact.
• Local or regional demand where relevant.
• Buyer comparisons, alternatives, pricing, implementation, or proof topics.
• Prompts in AI engines that resemble real purchasing questions.

Use the clustering logic in keyword research for AI products and search intent mapping. Avoid counting every keyword as an opportunity; cluster by buyer task and page purpose.

Execution fit

The opportunity should be feasible given the CMS, engineering resources, legal constraints, content operations, and brand rules. A strong opportunity may still be invalid if the client cannot update service pages, approve content, or fix rendering issues.

Step 3: Baseline the account

Discovery needs a starting line. Record data with source, date range, and known limitations.

Commercial baseline

Search baseline

Technical baseline

• Qualified inquiries, meetings, opportunities, and revenue where available.
• Sales cycle and lead acceptance rate.
• Current cost per lead from other channels.
• Revenue or margin by service and segment.
• Capacity to serve new demand.
• Non-brand impressions, clicks, CTR, and average position for qualified clusters.
• Current visibility by service, location, segment, or product category.
• Top landing pages and their conversion behavior.
• Indexed pages versus pages that should drive demand.
• Existing rankings for commercial queries.
• AI visibility evidence for representative prompts.
• Crawl coverage and internal-link depth.
• Indexation rules, canonicals, redirects, and duplicate patterns.
• Rendering status for important content and links.
• Core Web Vitals and mobile behavior.
• Structured data validity.
• Sitemap, robots, staging, and environment hygiene.

Use web analytics for AI products to validate events before you promise conversion reporting. If intake_submit, calls, trials, demos, or checkout events are unreliable, your roadmap must include tracking repair.

Content baseline

• Which pages support buyer paths?
• Which pages cannibalize each other?
• Which pages are thin, outdated, or duplicated?
• Which proof assets support service claims?
• Which topics are covered but not linked to conversion pages?

A baseline can be short. One page of numbers with dates is more useful than a chart with no interpretation.

Step 4: Audit buyer paths

Map how qualified buyers move from question to confidence to inquiry.

Path components

For each priority segment, document:

1. Entry query or AI prompt.
2. Matching page.
3. Proof needed.
4. Action available.
5. Qualification fields.
6. Handoff to sales or self-serve flow.
7. Follow-up measurement.

Example for a consultancy:

• Entry: “How do we choose an AI implementation partner?”
• Page: Service hub with capability, process, and proof.
• Proof: Relevant case study, certifications, delivery method.
• Action: Request a scoped discovery or technical assessment.
• Qualification: Company size, systems, timeline, budget range, decision maker.
• Handoff: CRM routing and sales SLA.
• Measurement: CTA click, form submit, accepted meeting, opportunity.

Use the model in service business site architecture to decide whether pages belong in service hubs, proof sections, local pages, comparison pages, or education clusters.

Common path failures

Discovery should name path failures, not only page defects:

• High-quality blog traffic with no service CTA.
• Service pages that describe features but omit proof.
• Contact forms that ask too much too soon.
• Calls routed to an unmonitored inbox.
• Multiple pages competing for the same buyer question.
• Strong case studies hidden below unrelated articles.
• Mobile CTA requires desktop-style navigation.

Each failure should have a severity and business consequence.

Step 5: Audit technical feasibility

Technical discovery answers whether the site can deliver the intended pages and signals.

Review:

• Crawlability of priority pages.
• Rendering of primary content, navigation, links, and forms.
• Indexation and canonical behavior.
• Redirect chains and stale URLs.
• Duplicate templates or parameterized pages.
• Page speed and interaction quality on real devices.
• Structured data for Article, Service, FAQ, Organization, Product, LocalBusiness, or Breadcrumb types.
• Accessibility issues that suppress usability.
• Staging, deployment, rollback, and QA workflow.
• Log evidence of crawler behavior, if available.

Use the priorities in technical SEO checklist for AI products. For client-side applications, include JavaScript rendering for SEO in SPAs and verify the rendered HTML, not only source code. For migration or redesign programs, study canonicals, redirects, and migrations before recommending URL changes.

Step 6: Analyze competition honestly

Competitive analysis should support strategy, not produce envy.

For each priority cluster, identify:

• Who currently earns qualified visibility.
• What proof they show.
• How their service page or product page is structured.
• Which technical or UX advantages they have.
• What they miss: pricing clarity, implementation detail, local proof, risk handling, segment specificity, or post-sale support.
• Whether AI engines describe them accurately.

Summarize three levels:

1. Hard to displace now: Strong brand, product, links, and buyer fit.
2. Contestable: Demand exists, but incumbent pages are generic or outdated.
3. Open: Real buyer need with weak or misleading coverage.

Use competitor content analysis to turn observations into content decisions. Do not copy the competitor; answer the buyer better.

Step 7: Score findings

Create a simple scoring model. The goal is not mathematical precision; it is shared judgment.

Example:

Use a multiplier or weighting only if the client agrees. Then create three bands:

• P0: Do first; high fit, feasible, and close to revenue.
• P1: Do next; strong opportunity but needs dependencies.
• P2: Later; useful but not enough evidence or capacity.

For each P0 and P1 item, record evidence, action, owner, effort, dependency, and expected signal.

Step 8: Estimate feasibility and effort

Estimate three dimensions separately.

Agency effort

Client effort

Time to signal

• Discovery report writing.
• Technical fixes.
• Content strategy.
• Content production.
• Internal linking.
• CRO changes.
• Tracking setup.
• Migration planning.
• Training and governance.
• Access and credentials.
• Stakeholder interviews.
• Content approval.
• Engineering time.
• Legal review.
• CRM integration.
• Sales feedback.
• Data sharing.

Distinguish:

• Implementation complete: Work shipped.
• Indexing signal: Page crawled, indexed, or rendering corrected.
• Search signal: Impressions, clicks, or ranking movement.
• Behavior signal: CTA clicks, scroll, form starts.
• Commercial signal: Qualified inquiry, meeting, opportunity, sale.

Do not promise a ranking date. Promise what will be built, measured, and reviewed.

Step 9: Define measurement plan

The measurement plan should be written in discovery, not invented after traffic changes.

Key event definitions

Define each event with exact names and examples:

• Qualified lead.
• Unqualified lead.
• Demo request.
• Discovery call request.
• Trial start.
• Pricing CTA click.
• Service CTA click.
• Form submission.
• Phone call.
• Sales-accepted lead.
• Opportunity created.

Use AI SEO lead qualification to align marketing and sales definitions. If form spam is common, define filtering before reporting begins.

Reporting cadence

Baseline rules

Step 10: Identify risks and dependencies

• Weekly internal: blockers and delivery risk.
• Monthly client: progress, search movement, inquiries, next actions.
• Quarterly: business review, scope fit, roadmap update.
• Record source system and date range.
• Separate brand and non-brand demand.
• Segment by service, market, or product where useful.
• State when data is unavailable.
• Avoid claiming causality without a clear change and control logic.

Every scope option needs a risk section.

Common discovery risks:

• Analytics cannot prove qualified leads.
• Engineering capacity is limited.
• Legal approval slows publishing.
• Stakeholders disagree on target segment.
• Brand guidelines prevent useful copy.
• CMS cannot support service hub requirements.
• The client expects rankings without production changes.
• Sales response time is slow.
• Previous agency owns documentation or access.
• A redesign or migration is planned but not scheduled.

For each risk, state:

1. Likelihood: low, medium, high.
2. Impact: low, medium, high.
3. Mitigation.
4. Owner.
5. Decision required.
6. Effect on timeline or price.

Do not bury a dependency that can make the engagement fail.

Step 11: Create scope options

End discovery with options rather than a single take-it-or-leave-it plan.

Option A: Focused pilot

Best when evidence is promising but the organization is not ready for a large retainer.

Includes:

• One service cluster or segment.
• Technical repair on priority pages.
• Two to four buyer-path assets.
• Tracking validation.
• One conversion path improvement.
• 60-90 day review.

Benefit: lowers commitment while proving collaboration quality.

Option B: Core growth program

Best when discovery shows feasible demand and delivery capacity.

Includes:

• Service hub and supporting pages.
• Technical roadmap.
• Content production.
• Internal linking.
• CRO experiments.
• Monthly reporting.
• Qualification and sales feedback loop.

Benefit: aligns search, conversion, and delivery operations.

Option C: Technical or migration recovery

Best when the main problem is visibility loss, rendering, indexation, or architecture.

Includes:

• Deep technical diagnosis.
• URL and redirect plan.
• Rendering QA.
• Template changes.
• Crawl and indexation monitoring.
• Phased rollout and rollback criteria.

Benefit: prevents content spend on a broken foundation.

Option D: Local or multi-location program

Best for service businesses competing by geography.

Includes:

• Service-area architecture.
• Local proof pages.
• Review and reputation workflow.
• Call and booking tracking.
• Seasonal demand planning.

Use local service area SEO for AI consultancies when geography drives buyer selection.

For each option, state deliverables, assumptions, exclusions, client responsibilities, price, timeline, and success signals.

Discovery report structure

A useful report is decision-oriented. Use this structure:

1. Executive summary: Goal, current state, main opportunity, recommended option.
2. Commercial context: Products, segments, constraints, sales process.
3. Evidence base: Sources, dates, limitations, access reviewed.
4. Demand findings: Qualified clusters and AI visibility patterns.
5. Buyer-path findings: Entry pages, proof, CTAs, qualification gaps.
6. Technical findings: Indexation, rendering, speed, structured data, infrastructure.
7. Content findings: Coverage, cannibalization, proof, decay.
8. Competitive findings: Contestable and open clusters.
9. Prioritized opportunities: Scores, actions, owners, effort, dependencies.
10. Measurement plan: Events, definitions, dashboards, cadence.
11. Risks and dependencies.
12. Scope options and pricing.
13. 90-day roadmap.
14. Decisions requested.

Appendix material can include crawls, logs, keyword tables, screenshots, and analytics validation. Do not let the appendix replace the summary.

A one-page discovery template

Use this in proposals.

Account snapshot

Qualified demand

Buyer path

Technical constraints

Top five findings

Scope option

Sample 15-day discovery schedule

Days 1-2: Access and alignment

Days 3-5: Baseline and diagnostics

Days 6-8: Demand and buyer-path review

Days 9-11: Competitive and content analysis

Days 12-13: Roadmap and options

Days 14-15: Presentation and decision

• Company:
• Primary product/service:
• Priority segments:
• Target markets:
• Commercial goal:
• Baseline date range:
• Economic buyer:
• Operational owner:
• Cluster 1:
• Cluster 2:
• Cluster 3:
• Evidence:
• Current page:
• Gap:
• Entry query:
• Matching page:
• Proof needed:
• CTA:
• Qualification fields:
• Routing:
• Tracking event:
• Rendering:
• Indexation:
• Speed:
• CMS:
• Release process:
• Owner:
• Deliverables:
• Client responsibilities:
• Exclusions:
• Timeline:
• Price:
• Success signals:
• Risks:
• Confirm scope, price, and confidentiality.
• Collect Search Console, analytics, CMS, CRM, and staging access.
• Interview economic buyer and delivery owner.
• Define priority segments and ineligible opportunities.
• Create an evidence log.
• Validate tracking events and form routing.
• Export baseline search, conversion, and sales data.
• Run crawl and rendering checks.
• Review indexation, canonicals, redirects, sitemap, and robots.
• Identify technical blockers on priority pages.
• Build candidate clusters from search, sales notes, support tickets, and AI prompts.
• Map queries to existing pages.
• Review service pages, proof, CTAs, forms, and routing.
• Interview sales or customer-facing staff.
• Record lead-quality examples.
• Compare current and desired buyer paths.
• Analyze incumbents for proof, clarity, and coverage.
• Identify content decay, cannibalization, and orphaned pages.
• Use content pruning and consolidation where existing coverage is messy.
• Draft prioritized opportunity list.
• Convert findings into actions, owners, effort, and dependencies.
• Create 30/60/90-day roadmap.
• Define measurement plan.
• Prepare scope options with assumptions and exclusions.
• Review internally for feasibility and margin.
• Present summary, findings, options, and risks.
• Answer feasibility questions.
• Document client objections.
• Request a decision.
• Move accepted scope into proposal, SOW, or contract.

For longer enterprise programs, extend interviews and technical review, but do not let the report become a book.

Pricing discovery

Discovery should be paid when it produces custom recommendations, access review, workshops, and roadmap ownership.

Pricing models

• Fixed fee: Common for a defined two- to four-week discovery.
• Audit fee credited: A fixed discovery fee credited toward a signed retainer.
• Diagnostic sprint: Small cross-functional sprint with a defined report.
• Hourly advisory: Useful for uncertain internal politics or complex procurement.
• Phase one of program: Discovery plus first implementation sprint.

Avoid free discovery unless the work is short, public information, and clearly designed to qualify fit.

What to include in price

What to exclude unless agreed

• Stakeholder interviews.
• Data validation.
• Crawl/render diagnostics.
• Demand cluster review.
• Buyer-path analysis.
• Competitive summary.
• Scoring workshop.
• Report preparation.
• Presentation meeting.
• One revision round.
• Scope options.
• Roadmap handoff.
• Full content production.
• Development work.
• Link acquisition.
• Ongoing reporting.
• Migration execution.
• Legal, PR, or paid media strategy.
• Custom CRM implementation.
• Unlimited revisions.

State exclusions explicitly.

Discovery presentation tips

Executives decide using the first five pages. Engineers decide using constraints. Operators decide using workload.

Present in three layers:

1. Decision layer: Goal, opportunity, recommended option, price, risks.
2. Evidence layer: Demand, buyer paths, technical findings, competition.
3. Operational layer: Roadmap, owners, dependencies, reporting.

Use one sentence per finding. Example:

“Enterprise service pages render primary pricing proof only after user interaction, which likely suppresses crawling and mobile conversion; we recommend a server-rendered proof block in the next release.”

That is more useful than “improve technical SEO.”

Handle objections with evidence:

Internal-linking and architecture recommendations

• If the client says “our competitors don’t do this,” explain why their buyer path differs.
• If the client says “traffic used to be higher,” show what changed and what cannot be compared.
• If the client says “we just need more content,” show which pages already fail to convert or render.
• If the client says “can you guarantee first place,” explain the constraints and propose leading indicators.

Discovery often reveals that pages exist but cannot be found. Document:

• Number of clicks from home to priority pages.
• Missing parent hubs.
• Orphaned service or proof pages.
• Repeated anchor-text problems.
• Navigation structure that hides commercial services.
• Faceted URLs consuming crawl budget.
• Blog clusters not connected to service pages.

Use internal linking strategy to define rules for hubs, sibling pages, proof, conversion pages, and anchors. Linking is not a post-publish task; it is architecture.

Conversion discovery

Even with correct demand, inquiries can fail because of form design, routing, proof, speed, or qualification friction.

Review:

• CTA visibility on mobile.
• Number of required fields.
• Whether form explains what happens next.
• Spam filtering and routing.
• Response time SLA.
• Sales qualification questions.
• Proof near CTA.
• Error handling and confirmation page.
• Event tracking across starts, submits, and failures.

Use SEO CRO audit to separate traffic problems from conversion problems. A page can need neither more content nor more links; it may need a clearer next action.

Documentation and handoff

Discovery output should be usable by delivery.

Hand off:

• Goals and segment definitions.
• Baseline exports and date ranges.
• Access list and owners.
• Technical findings.
• Opportunity scores.
• Roadmap and dependencies.
• Event definitions.
• Scope assumptions and exclusions.
• Risk register.
• Client responsibilities.

Then transfer into SEO service onboarding assets. Discovery should not disappear into a slide deck; it should become the first version of the client operating system.

When not to proceed

Discovery can reveal that the engagement should stop or shrink. Consider declining or redesigning the offer when:

• The client’s target buyer does not exist in qualified search demand.
• Revenue goal depends on one keyword position.
• Engineering cannot change templates or rendering.
• Legal cannot approve any buyer-facing content.
• The client refuses to share lead feedback.
• Sales cannot respond to inquiries.
• The primary problem is product-market fit, not SEO.
• The previous agency owns all context and the client refuses to grant access.
• Budget is below the minimum needed for viable delivery.

A respectful decline is better than a renewal doomed by impossible assumptions.

Enterprise discovery considerations

Large organizations need extra governance:

• Multiple buyer personas and regional owners.
• procurement, security, privacy, and brand reviews.
• Separate staging and production systems.
• Central SEO versus local business ownership.
• Legacy CMS, templates, and release freezes.
• Internal politics between marketing, product, engineering, and sales.
• International URL and language rules.
• Data-sharing restrictions.

Use a RACI for findings approval, technical implementation, content approval, analytics access, and roadmap changes. Without this, discovery becomes the first casualty of internal politics.

Common mistakes

Selling a report instead of a decision

A report is not successful because it is long. It is successful when the client accepts a feasible scope.

Starting with tools

Tools produce evidence, not strategy. Begin with business fit and buyer paths.

Ignoring lead quality

Traffic growth can coexist with zero useful inquiries. Define qualified demand early.

Promising rankings

Search systems are not under your control. Commit to work, diagnostics, and measurement.

Missing client dependencies

The most common delay is not SEO skill; it is access, approvals, engineering time, and sales feedback.

Auditing every page

Audit enough to find patterns and prioritize. A complete inventory can follow after the scope is signed.

Not validating tracking

If forms, calls, and events are broken, every later report is uncertain.

Hiding assumptions

If the roadmap assumes engineering time, publishing capacity, or CRM access, state it.

Presenting too many options

Three to four scope options are usually enough. More choices create delay.

Losing context at handoff

The roadmap must include evidence and assumptions, otherwise the delivery team will reinterpret it.

Client questions to anticipate

Prepare answers for:

• “Why can’t you just tell us what to do for free?”
• “Can you guarantee more traffic?”
• “Why do you need sales data?”
• “Can we skip technical fixes?”
• “Can we start with blog content?”
• “Why do you need engineering time?”
• “How fast will rankings improve?”
• “What if we already have an agency?”
• “What happens to the discovery report if we don’t continue?”
• “Can we pay only after traffic grows?”

Answer plainly. If a request is unreasonable, explain the risk and offer an alternative.

Discovery quality checklist

Before presenting, confirm:

Renewal and phase-two thinking

• Commercial goal documented by the client.
• Priority segment chosen.
• Baseline data dated and sourced.
• Tracking events validated.
• Priority buyer paths mapped.
• Technical rendering checked on mobile and desktop.
• Indexation and canonical issues reviewed.
• Competitive gaps summarized.
• Findings scored.
• Top opportunities have owner, effort, and dependency.
• Risks and client responsibilities written.
• Measurement plan defined.
• Scope options priced.
• Exclusions stated.
• Roadmap has 30/60/90-day actions.
• Decision requested.
• Report handoff owner assigned.

Discovery should prepare future phases, not exhaust the team. Record ideas intentionally deferred:

• New market expansion.
• Product-led or self-serve flow.
• Additional service line.
• Local proof program.
• Technical debt cleanup.
• Content refresh system.
• AI visibility testing.
• Conversion experiment queue.
• Documentation or support content.

These become roadmap candidates later, not unpaid additions to current scope.

Use discovery to decide whether a lead magnet belongs in the buyer path and which fields reveal true fit.

Discovery findings should recommend one of a few prebuilt AI service packages rather than another vague scope.

The discovery phase should confirm the promises made by service page conversion copy about scope, access, cadence, and responsibilities.

Discovery findings should strengthen the follow-up: the proposal follow-up system turns those findings into proof, sequencing, and decision prompts.

The first call is not full discovery; the SEO discovery call script decides whether the broader discovery phase should begin.

A diagnostic is the deliverable of a scoped review; pair this SEO diagnostic deliverable standard with the SEO discovery phase that defines the engagement.

Bottom line

Discovery turns ambiguity into a signable plan. Diagnose business fit, demand, buyer paths, technical feasibility, and organizational constraints before promising growth. Your next action is to replace your generic audit template with a fifteen-day discovery framework that ends in scored opportunities, risks, and priced scope options.

FAQ