How to Compare SEO Proposals Without Comparing Vague Deliverables

01

Preconditions

Proposal comparison begins before the proposals arrive. The buyer must define the decision, the site boundaries, the internal resources available for implementation, and the evidence needed to approve work. Otherwise each vendor can choose a different problem, measurement window, and responsibility model while presenting the result as a comparable monthly package. Preserve the same brief for every bidder and require written assumptions. The purpose of normalization is not to force identical strategies. It is to expose where differences come from scope, risk, labor, evidence, or ownership rather than from vague labels and decorative deliverable counts.

Before comparing vendors, write a one-page buyer brief containing:

• The business outcome that matters
• Revenue-critical services, products, or locations
• Current site and platform
• Known migration, redesign, or tracking constraints
• Search Console and analytics availability
• Internal engineering, content, and approval capacity
• Regulatory or brand restrictions
• Budget range and minimum engagement period
• Who owns final production decisions

Without a shared brief, vendors are solving different imagined problems and their prices are not meaningfully comparable.

Google’s hiring guidance recommends learning the basics, asking for a technical and search audit, and being wary of anyone who guarantees rankings. Google also states that payment cannot buy inclusion or ranking in organic results. Do you need an SEO?^[1]

02

Ordered process

1. Separate diagnosis from recurring work.
2. Normalize the business objective.
3. Normalize each proposed deliverable.
4. Compare access and account ownership.
5. Compare evidence quality and measurement plans.
6. Compare production standards, risks, and exclusions.
7. Compare exit terms and hold a contradiction meeting.

1. Separate diagnosis from recurring work Mark each proposed item as one of:

• Baseline discovery
• One-time remediation
• Recurring monitoring
• Recurring production
• Experiment
• Reporting
• Account or platform administration

A proposal that charges indefinitely for a one-time setup item should explain the continuing work. A proposal that skips diagnosis should explain why the chosen actions are justified before the site is inspected.

2. Normalize the objective Rewrite the vendor’s stated goal in observable terms.

Weak:

Improve online visibility.

Stronger:

Increase qualified nonbrand organic entrances to the service pages for three priority markets while preserving conversion tracking and excluding irrelevant national queries.

The stronger statement is not a promise of growth. It is a testable direction that constrains the work.

3. Normalize each deliverable For every line item, ask:

“Technical SEO” is not a normalized line item. “Correct canonical and internal-link generation in the product template, validate against a controlled URL sample, and monitor recurrence after releases” is.

4. Compare access and ownership Record who owns and controls:

• Domain registrar
• DNS and CDN
• Hosting
• CMS and code repository
• Google Search Console
• Analytics and tag manager
• Business Profile
• Advertising accounts
• Email and form systems
• Content files and media licenses
• Rank-tracking or reporting accounts

The provider can receive delegated access. The client should not become dependent on an account that only the provider controls.

Google’s guidance specifically recommends granting limited access rather than handing over unrestricted control, and asks buyers to understand proposed changes before allowing implementation. Do you need an SEO?^[1]

5. Compare evidence quality Score evidence in descending order:

1. Direct observation on the client’s site
2. First-party search or platform documentation
3. Reproducible test or controlled experiment
4. Relevant third-party research with visible methodology
5. Expert inference with stated limits
6. Tool score without interpretation
7. Anecdote or testimonial

FTC truth-in-advertising guidance says objective advertising claims require a reasonable basis and that testimonials do not substitute for substantiation. FTC Advertising Substantiation Policy^[2]

This matters when a proposal uses claims such as “our process increases traffic 300%” or “clients reach page one in 90 days.” Ask for the population, denominator, baseline, exclusions, observation window, and typical result.

6. Compare measurement plans A defensible measurement section identifies:

• Baseline period
• Primary and secondary metrics
• Brand versus nonbrand segmentation where possible
• Page, query, country, device, and conversion scope
• Known attribution limits
• Seasonality and campaign effects
• Release and annotation process
• Decision cadence

A dashboard is not a measurement plan. The plan should explain what decision changes when a metric moves.

7. Compare production quality, not counts For content, define:

• Search intent and audience
• Evidence requirements
• Original contribution
• Review process
• Byline and disclosure policy
• Internal linking
• Media rights
• Update and correction process
• Acceptance standard

For links or outreach, define:

• Target criteria
• Prohibited tactics
• Disclosure and sponsorship handling
• Ownership of relationships and accounts
• Reporting of rejected, removed, or paid placements

For technical work, define:

• Environments
• Change control
• Testing
• rollback
• deployment responsibility
• post-release verification

The number of monthly outputs becomes meaningful only after the acceptance standard is visible.

8. Compare risk and exclusions Require explicit treatment of:

• Site migrations
• Template changes
• generated content
• paid links or sponsorships
• reputation management
• access to customer data
• legal or medical claims
• local listing changes
• production credentials
• subcontractors
• AI use and review

An exclusion is not necessarily a defect. Hidden exclusions are.

9. Compare exit terms Document what happens when the engagement ends:

• Account access returned or revoked
• Credentials rotated
• Content and working files delivered
• Licenses identified
• Redirects and technical changes documented
• Reporting history exported
• Open experiments closed or transferred
• Subscriptions canceled or reassigned
• Final production state and rollback recorded

A low monthly price can be expensive if the client cannot leave without rebuilding accounts, content, or infrastructure.

10. Hold a contradiction meeting Give each finalist the same three questions:

1. Which part of your proposal has the weakest evidence?
2. What observation would cause you to change the first 90-day plan?
3. Which requested activity would you advise us not to buy yet?

A provider who cannot identify uncertainty is selling certainty theater, humanity’s most renewable resource.

03

Failure cases

A comparison fails when the buyer allows each provider to define a different success metric, hides internal labor, or treats forecasts as guaranteed outcomes. It also fails when the apparent bargain depends on provider-owned accounts, unpriced engineering work, undisclosed subcontractors, or an exit process that leaves the client without files and credentials. Do not repair an incomplete proposal by inventing favorable assumptions on the vendor’s behalf. Mark the field unknown, request the missing evidence, and exclude the proposal from final comparison if the provider cannot state the responsibility or completion condition clearly.

• Choosing by deliverable count: A larger number can mean thinner work.
• Choosing by forecast alone: Forecasts can hide assumptions about implementation, competition, seasonality, and conversion.
• Letting every vendor define success differently: Prices remain incomparable.
• Ignoring internal labor: A cheaper proposal may require far more client writing, engineering, review, or project management.
• Treating tool scores as outcomes: A score can identify inventory; it does not establish business consequence.
• Accepting provider-owned accounts: Exit cost and operational risk increase.
• Signing before the audit scope is bounded: The provider may use the audit as either a sales document or an unlimited discovery project.

04

Completion criteria

The evaluation is complete when every finalist can be compared in one normalized record covering objective, baseline, scope, actions, evidence, ownership, dependencies, risk, measurement, acceptance, recurring work, exclusions, and exit. The buyer should also know the expected internal labor and which assumptions could materially change cost or timing. A selected proposal should have an explicit first decision window and a process for changing course when evidence contradicts the plan. Preserve the rejected alternatives and reasons so later performance reviews compare the engagement against the decision actually approved rather than against sales language remembered selectively.

The comparison is complete when every finalist has a normalized row for:

• Objective
• Baseline
• Scope
• Deliverables
• Evidence
• Access
• Ownership
• Dependencies
• Risks
• Measurement
• Acceptance
• Exclusions
• Exit
• Total client and provider effort

The selected proposal should win because its decisions and responsibilities are clearer, not because its PDF contains the most decorative rectangles.

References

Sources behind this record

1. Do You Need an SEO? — Google (accessed August 2, 2026)
2. Policy Statement Regarding Advertising Substantiation — Federal Trade Commission (accessed August 2, 2026)

Corrections

Correction history

No corrections recorded.

To report an error, use the public corrections path.