What Should You Verify Before Hiring an SEO Agency?

Begin with readiness. SEO work can identify and address search-related constraints, but it cannot compensate for every business, product, conversion, analytics, or implementation problem. A buyer should understand which foundation exists before purchasing ongoing production.

Review three areas first: technical access, useful information, and conversion capability. Technical access means important pages can be crawled, rendered, canonicalized, and indexed as intended. Useful information means the organization can provide subject knowledge, evidence, and pages that answer real customer needs. Conversion capability means visitors can understand the offer, complete the next action, and be measured reliably.

The source described a founder who paid for six months of work and discovered at month four that key pages were blocked, then repeated the six-month duration. Treat that scenario as an internal illustration without a supporting URL.

The operational lesson is to validate indexation, templates, analytics, and implementation ownership before scaling content or authority work.

Run a pre-agency review. Confirm site ownership and access, Search Console configuration, analytics events, conversion definitions, sitemap and robots behavior, canonical patterns, priority templates, mobile usability, content inventory, internal-link structure, current vendors, and the backlog of unresolved development work.

The source used a mobile-load threshold of under three seconds. Because no supporting URL appears, treat it as a historical operating example rather than a universal pass condition. Use current field and laboratory data, page purpose, user conditions, and documented performance guidance to identify specific bottlenecks.

Assess internal capacity. Identify the executive sponsor, day-to-day contact, analytics owner, engineering contact, content approver, legal or compliance reviewer, and subject-matter experts. Record realistic response times and competing priorities. An agency cannot approve its own recommendations on the client's behalf.

Test the conversion path. Review the landing page, offer, form or checkout, follow-up process, lead routing, qualification, and attribution. If visitors cannot complete or be credited for the desired action, increased search traffic will not provide a reliable commercial evaluation.

If two or more readiness areas are materially unresolved, consider a scoped diagnostic rather than an open-ended retainer. Preserve the source's one-time audit token as a service-format example, not a guarantee that one project resolves every foundation issue.

A readiness review passes when access, measurement, ownership, implementation, and page purpose are sufficiently clear to support the proposed work. It fails when the buyer expects the agency to discover the business model, repair every department, or deliver growth without access and implementation.

When readiness is uncertain, ask candidates to price a diagnostic that produces a baseline, risk register, prioritized backlog, required resources, and a decision about the next delivery phase.

The Model Match Test can be used as a structured conversation about fit. It should not be treated as a proprietary score or a substitute for references, contracts, and delivery evidence.

Evaluate four axes. The first is business-model alignment. Determine whether the organization earns revenue through transactions, subscriptions, leads, advertising, marketplaces, recurring relationships, or another model.

The agency should connect search intent, page type, content, measurement, and conversion to that model, including the needs of a B2B enterprise where relevant.

The source contrasted transactional businesses with relationship businesses and used a 12-month SaaS sales cycle as an example. Preserve that duration as an illustration, not as a universal sales-cycle fact.

Ask how the agency handles long journeys, assisted conversions, branded demand, account-based sales, and offline outcomes when relevant.

The second axis is timeline alignment. The source cited a 4-12 month horizon, 60-90 days, and a domain under two years old. None of those ranges is supported by a URL in the source. Treat them as historical planning examples.

Require the agency to distinguish setup, discovery, implementation, recrawling, early indicators, and business outcomes rather than collapsing them into one promise.

The third axis is delivery alignment. Determine whether the need is technical remediation, information architecture, content strategy, writing, digital PR, link-risk review, local search, analytics, conversion support, or program management.

Ask which work is performed in-house, which work is subcontracted, how reviewers are qualified, and who signs off on quality.

The fourth axis is reporting alignment. Agree on the business question and the supporting indicators. The source referenced success at 90 days, 6 months, and 12 months, then a conflict in month five.

Preserve those tokens as examples without treating them as mandatory review points. Define the cadence that matches the engagement and the data available.

Ask the agency to describe its last three relevant engagements by client model, problem, baseline, scope, implementation conditions, team, measurement, and lessons. Case details may need anonymization, but the explanation should be specific enough to show causal caution and operating knowledge.

Compare staffing. Confirm who sells, strategizes, analyzes, writes, builds links, manages the account, and reviews deliverables. Request named roles in the agreement or a clear substitution process. A senior pitch does not prove senior delivery.

Document dependencies and exclusions. A proposal should state what the client must provide, which systems are outside scope, how additional work is approved, and what happens when implementation is delayed.

The fit review passes when the agency explains why its model, team, and sequence match the client's constraint and can identify conditions that would change the plan. It fails when the answer relies on generic best practices or an unrelated case study.

If two candidates appear equally strong, compare the clarity of assumptions, staffing, evidence, implementation support, and exit terms rather than choosing by presentation quality.

The Red Flag Audit Framework is a rapid proposal review. It identifies questions that require follow-up; it does not prove that an agency is unethical or incapable.

First, flag guarantees. The source referenced Page 1 claims. Rankings and traffic depend on search systems, competitors, implementation, demand, content, links, and user behavior. An agency can guarantee its own deliverables and service standards, but it should not guarantee a specific organic position it cannot control.

Second, flag vague scope. Terms such as ongoing optimization, authority building, or monthly SEO work need definitions. Require outputs, responsibilities, review standards, assumptions, dependencies, change-control rules, and acceptance criteria. Do not replace vagueness with arbitrary production quotas.

Third, flag proposals written without a baseline. A candidate may provide an initial hypothesis before access, but a detailed strategy should acknowledge missing data. Ask what the agency inspected, which data it still needs, and which recommendations may change after discovery.

Fourth, flag reporting that isolates rankings and sessions from conversions, page purpose, implementation, and business context. Revenue attribution may not be available or reliable in every organization, so the correct standard is transparent measurement, not a mandatory revenue claim.

Fifth, flag one-sided contract risk. The source cited a 90-day pilot and a 12-month lock-in. Treat those lengths as historical examples rather than universal standards. Review term, termination, notice, renewal, payment, data access, intellectual property, confidentiality, subcontracting, non-solicitation, and transition support with qualified counsel where appropriate.

Review the proposed link work. Require a description of sourcing, relevance checks, editorial control, disclosure, payment practices, risk management, and reporting. Avoid providers that hide methods behind vague networks or guaranteed volume.

Review content production. Ask for briefs, source requirements, subject review, authorship, AI-use policy, plagiarism controls, corrections, and responsibility for regulated claims. A sample should be evaluated for accuracy and usefulness, not just optimization scores.

Review technical delivery. Confirm how issues are reproduced, prioritized, documented, tested, and handed to developers. Ask whether the agency implements changes, supports implementation, or only supplies recommendations.

The proposal passes when scope, team, evidence, assumptions, delivery, reporting, and commercial terms are inspectable. It fails when accountability depends on promises that cannot be measured or methods the buyer is not allowed to understand.

If several warning signs appear, request a revised proposal or remove the candidate. Do not negotiate a guarantee into a smaller guarantee; resolve the underlying misunderstanding of control and evidence.

A discovery call should be a two-way evaluation. Prepare the same core questions for every candidate and record the answers so confidence, personality, and presentation do not replace evidence.

Ask about a lost, paused, or unsuccessful engagement. The useful part is not the failure itself; it is whether the agency can explain its contribution, the client's contribution, the evidence available, the corrective action, and the process change that followed. Be cautious when every failure is blamed entirely on the client or an algorithm.

Ask how the agency responds when results do not match the plan. The source referred to month three. Preserve the phrase as part of the original question, but require a staged diagnostic rather than a predetermined rescue routine.

A credible process checks implementation, measurement, indexation, search intent, content, competition, links, seasonality, and business changes before choosing a response.

Ask who will work on the account. Request role, seniority, allocation, location where relevant, subcontractor involvement, reviewer, and escalation path. Confirm how staffing changes are communicated.

Ask about client retention carefully. The source requested the percentage of clients retained after 12 months. Retention can be influenced by project type, contract structure, seasonality, budgets, and successful completion, so it is not a standalone satisfaction metric. Request context and a range if the agency is willing to share.

Ask when the agency advised against SEO or against a retainer. A credible answer should identify the client's stage, the constraint, the alternative action, and the evidence required for reconsideration.

Ask for the onboarding sequence for the first 30 days. The process may include access, analytics verification, stakeholder interviews, technical discovery, content inventory, baseline reporting, risk review, and prioritization. Confirm which parts are standard and which depend on scope.

Ask how the agency handles corrections and disagreements. Review who approves factual claims, how errors are corrected, how disputed recommendations are documented, and how a client can challenge an assumption.

Ask for a mid-engagement reference when permission exists. The source suggested a client in the middle stage of an engagement. Preserve that range as an example, not a mandatory reference profile. Confirm that the reference understands what information may be shared and do not pressure clients for endorsements.

The question stage passes when answers are specific, internally consistent, and compatible with the proposal. It fails when delivery staff, methods, data ownership, or escalation remain unclear.

If an answer is plausible but unverifiable, mark it as an assumption and request evidence in the next stage rather than treating confidence as proof.

SEO pricing should be evaluated through the work, responsibilities, risk, and decision supported. A low or high fee is not meaningful until the buyer understands what is included, excluded, and dependent on client implementation.

Four common models appear in the source. A monthly retainer can support continuing strategy, production, analysis, and implementation coordination. It needs clear responsibilities, service levels, meeting cadence, change control, and a way to adjust priorities without hiding work.

A project price can suit a technical review, migration plan, content architecture, analytics setup, or another bounded decision. Define the inputs, assumptions, deliverables, revisions, acceptance criteria, and what happens after the project.

Performance-linked pricing can create alignment in some settings, but only when attribution, baseline, control, data access, implementation, and risk allocation are credible. It can also encourage easy targets or disputes about credit. Review the formula and exclusions with commercial and legal expertise.

Hourly pricing can be suitable for advisory sessions, investigation, training, incident response, or uncertain work. It is less suitable when the buyer expects a defined outcome but the scope remains unmanaged. The model should match the uncertainty, not a blanket rule about strategy.

Request a budget allocation explanation. Ask how fees support senior strategy, account management, technical analysis, writing, editing, digital PR, tools, reporting, and subcontractors. Do not treat hours as the only value measure, but confirm that the staffing model is plausible.

The source used a heavily content-weighted retainer as a warning example. The correct allocation depends on the diagnosed constraints and may change across phases.

Evaluate opportunity cost carefully. Delayed implementation, weak measurement, unnecessary content, risky links, or duplicated pages can consume time and resources. Do not describe hypothetical penalties or lost revenue as certain without evidence.

Compare proposals on a normalized sheet: objective, term, deliverables, staffing, client inputs, exclusions, measurement, intellectual property, data access, termination, transition, and total expected cost.

Pricing passes review when the buyer can connect the fee to a defined operating model and can identify which changes require additional approval. It fails when recurring revenue is protected by vague scope or when performance incentives depend on unverifiable attribution.

If prices differ substantially, ask which assumptions, staffing, production volume, risk, and responsibilities explain the difference. Do not assume the cheapest is efficient or the most expensive is strategic.

An agency's website, articles, profiles, and public search presence can reveal how it communicates, documents evidence, and maintains its own assets. They are useful samples, but they do not prove that the same team or process will be assigned to your account.

Review the service pages. Can you identify the client, problem, process, deliverables, limitations, and next step? Generic language may indicate weak positioning, but it may also reflect a broad client base. Ask for the reasoning behind the public message.

Review published articles. Check originality, sources, subject review, update practices, accuracy, and whether the content helps a reader make a decision. Do not equate length, frequency, or keyword use with quality.

Review public research and methodologies. A documented process can show intellectual investment, but a named framework is not evidence by itself. Verify whether the method contains reproducible steps, evidence boundaries, quality controls, and examples.

Review authorship and accountability. Identify who writes, reviews, and corrects material. Look for accurate bios and permitted profiles. Do not infer expertise from schema markup or social-profile activity alone.

Review public links only as one evidence source. Link tools can be incomplete, and a referring domain's value depends on relevance, context, editorial quality, disclosure, and risk. Ask the agency to explain its acquisition policy and show a permitted sample.

Search the agency's brand and service combinations if useful. The source gave an example query using the agency name and SEO strategy. Preserve the concept without adding an example domain or URL. A weak result is a data point, not an automatic disqualification, because branding, site age, priorities, and query interpretation vary.

Ask the agency how it markets itself and why. The answer should connect audience, channel, content, referrals, partnerships, and measurement. A provider may choose not to rely on organic acquisition for valid reasons.

Compare public quality with the proposed client deliverables. Request a sample brief, report, technical ticket, content draft, or implementation log when confidentiality allows. Confirm whether the sample was produced by the team assigned to you.

The evidence review passes when public work and permitted samples support the agency's stated process and quality standards. It fails when the pitch depends on unverifiable traffic figures, borrowed thought leadership, or assets the delivery team did not create.

If public evidence is sparse, request a paid discovery, references, team interviews, or a smaller initial scope rather than assuming either excellence or incompetence.

A capable agency can still fail inside a weak operating model. Before launch, convert the proposal into a working agreement that defines discovery, delivery, implementation, measurement, communication, and exit.

The source recommended a 90-day pilot. Preserve that duration as a historical structure, not a universal rule. A pilot or initial project should be long enough to complete meaningful discovery and agreed work, while keeping the decision risk proportionate to the evidence available.

Define the initial objective and deliverables. Examples include a baseline, prioritized roadmap, technical remediation plan, content architecture, measurement repair, or an implementation phase. Separate actions the agency controls from outcomes influenced by search systems, competitors, and the client.

Assign internal ownership. The client contact needs authority or a reliable escalation path, access to stakeholders, and time to coordinate approvals. Also identify engineering, analytics, content, product, legal, and subject experts as required.

Agree on reporting. Define the business question, leading indicators, implementation status, data sources, attribution limits, blockers, decisions, and next actions. The source suggested a standing 30-minute review call. Preserve that duration as an operating example, not a requirement.

Create change-control rules. Specify how new discoveries alter priorities, how additional work is estimated, who approves scope changes, and how urgent issues are handled.

Create an implementation tracker. Record recommendation, affected URLs or templates, evidence, priority, owner, dependency, status, deployment date, acceptance test, observation, and conclusion. This separates recommendations delivered from recommendations implemented.

Protect data and continuity. Confirm account ownership, permissions, exports, documentation, subcontractors, confidentiality, intellectual property, and transition support. The client should retain access to its own systems and records.

Set review and exit criteria. Define when to continue, narrow, pause, or end the engagement. Include unmet dependencies, quality issues, staffing changes, strategic shifts, and completion of the agreed scope.

The engagement passes setup when both sides understand the objective, responsibilities, data, implementation process, review cadence, and termination path. It fails when the agency is expected to act autonomously without the information and approvals only the client can provide.

If early delivery reveals a different primary constraint, document the evidence and revise the sequence through the agreed change process rather than protecting the original pitch.

Do not hire an SEO agency merely because organic search appears attractive. The channel and provider must match the business stage, evidence, timeline, and implementation capacity.

If the offer and target customer are still changing rapidly, first clarify positioning, customer research, and the conversion journey. Search content built on unstable assumptions may require repeated rework.

If the organization needs revenue inside 60-90 days, do not rely on SEO as the only acquisition response. The source framed paid search or direct outreach as faster alternatives, but channel speed varies. Compare current demand, economics, sales capacity, and evidence before selecting the mix.

If the site has unresolved manual actions, security incidents, migration failures, extensive indexation problems, or mass low-value content, begin with diagnosis and remediation. Do not scale production on a foundation that has not been assessed.

If the primary need is continuous cross-functional leadership inside a complex product or platform, an in-house hire may provide better context and coordination. An agency may still support specialist work, but the ownership model should be explicit.

If the organization cannot provide access, approvals, subject experts, engineering support, or reliable measurement, repair those dependencies before purchasing a broad retainer.

If the budget supports only a fraction of the required work, narrow the problem. A scoped audit, advisory engagement, implementation project, analytics review, or specialist assessment may create a clearer next decision.

Evaluate conflicts. Confirm whether the agency serves direct competitors, how information is separated, and whether exclusivity is required or commercially reasonable.

The no-hire decision passes when it identifies the blocking condition and the work required to become ready. It fails when the organization postpones indefinitely without assigning an owner or validation criterion.

If you remain uncertain, commission a bounded independent review that separates technical, content, authority, analytics, conversion, and organizational constraints. Use that diagnosis to decide between an agency, in-house role, fractional advisor, project specialist, or internal work.

Generic buying advice often says to review case studies, check testimonials, inspect reviews, and request references. Those steps can help, but each asset is curated. A favorable reference demonstrates that one relationship worked under specific conditions; it does not prove that the same team, scope, market, or implementation environment will exist in your engagement.

Another problem is evaluating tactics before diagnosing the business. Asking whether an agency offers links, technical audits, or content production is less useful than asking which constraint should be addressed first and why.

A technically strong provider can still be a poor fit when the primary problem is positioning, product-market fit, conversion, measurement, internal implementation, or subject-matter review.

Business-model fit also matters. E-commerce, local services, publishers, B2B SaaS, marketplaces, and complex professional services can differ in page architecture, search intent, content depth, attribution, sales cycles, compliance, and implementation.

Experience in another model may still transfer, but the agency should explain which principles transfer and which assumptions require testing.

The most useful buyer question is not a broad request for capabilities. Ask the agency to explain why its proposed sequence is appropriate for your site now, which evidence supports that sequence, what dependencies could block it, which outcomes remain outside its control, and how it would revise the plan if early findings contradict the pitch.

Buyers also underestimate their own role. Delayed access, unclear approvals, unavailable experts, weak analytics, and slow engineering can make a sound recommendation impossible to implement or evaluate. Agency selection must therefore assess the joint operating model, not only the supplier.

When candidate answers remain vague, do not fill the gaps with optimism. Request a narrower discovery project, clarify the brief, or postpone the decision until the underlying business and measurement questions are resolved.

Early agency evaluations can be dominated by presentation quality. Confident forecasts, polished case studies, and broad capability lists create comfort even when they do not address the buyer's actual constraint.

A stronger process begins with the site's condition, business model, measurement, and implementation capacity. The central question becomes: why is this sequence appropriate now, which evidence supports it, and what would cause the agency to change course?

That question improves the conversation because it requires the agency to engage with assumptions rather than repeat a standard pitch. A credible candidate may recommend a diagnostic, a narrower scope, another provider, internal hiring, or readiness work before a retainer.

The buyer must show the same honesty. Share known constraints, approval delays, resource limits, attribution gaps, prior failures, and commercial priorities. Hiding those conditions produces a proposal that cannot be evaluated fairly.

The most useful relationship is not automatically the most strategic-sounding one. It is the relationship with clear responsibilities, inspectable work, truthful reporting, effective implementation, and a shared process for revising assumptions. That standard can be applied to a project, retainer, advisory engagement, or hybrid team.

Days 1-3

Complete the readiness review: verify crawlability, Search Console, analytics, conversions, content inventory, technical backlog, internal owners, and implementation capacity.

Outcome: A documented decision about whether to request a diagnostic, specialist project, or ongoing agency proposal.

Days 4-7

Write the fit brief: define the business model, customer journey, priority problem, site baseline, realistic dependencies, primary business question, and required delivery capabilities.

Outcome: A one-page brief that gives each candidate the same context and reduces generic pitching.

Days 8-14

Review three to five candidates using the proposal warning-sign checklist, public work, staffing evidence, methods, sample reporting, references, and commercial terms.

Outcome: A shortlist with documented strengths, unresolved assumptions, evidence requests, and disqualifying risks.

Days 15-21

Run structured discovery calls covering failed engagements, diagnosis under uncertainty, assigned staff, a 30-day onboarding example, corrections, subcontractors, implementation support, and the client's stage.

Outcome: A ranked comparison based on fit, evidence, accountability, and operating maturity rather than presentation alone.

Days 22-27

Negotiate a 90-day pilot or another proportionate initial scope, then define deliverables, responsibilities, measurement, reporting, change control, data ownership, review criteria, and exit terms.

Outcome: A written initial agreement that creates evidence while limiting ambiguity and one-sided risk.

Days 28-30

Prepare implementation: assign the client owner, configure access, create the tracker, confirm reviewers, schedule decision-focused meetings, and brief internal teams on dependencies and approvals.

Outcome: A launch-ready operating model in which recommendations can be implemented, tested, and maintained.