Before ending an SEO relationship, establish business-controlled ownership or administrator access for every critical account, export the working data and configuration records, test recovery paths, and only then remove the vendor.
- The business should control the registrar, DNS, analytics, Search Console, repository, and recovery identities.
- Removing a vendor before testing internal access can create an avoidable outage.
- Offboarding should preserve change history, redirects, tracking configuration, source files, and license records.
01
Preconditions
Do not begin an SEO vendor offboarding by deleting the vendor’s accounts. Begin by establishing what the business owns, what the vendor controls, and which systems can affect revenue or public availability. Record the contract end date, notice requirements, transition contacts, authorized internal owner, and a maintenance window for credential changes. Preserve the current state before anyone starts rotating keys or transferring domains.
The business needs organization-controlled identities that do not depend on a departing agency employee’s personal email. At minimum, two internal administrators should be able to authenticate, recover access, and receive security notices. Shared passwords are a poor substitute when a platform supports individual roles. Google Search Console, Analytics, and Business Profile all distinguish ownership or administrative authority from lower-permission access, although their role names and transfer mechanics differ. [1] [2] [3]
Create an asset inventory covering the registrar, DNS, hosting, CDN, content management system, repository, deployment platform, Search Console, Analytics, tag manager, Business Profile, product feeds, call tracking, rank tracking, shared drives, email accounts, API credentials, automation jobs, content calendars, redirect maps, media libraries, and paid tools. For each item, record the current owner, administrators, recovery address, billing identity, export method, and operational purpose.
The rollback owner must be named before access changes begin. If the organization cannot identify who can restore DNS, deployments, analytics, and critical integrations, pause the termination sequence and repair that governance gap first.
02
Ordered process
- Capture the contractual and technical baseline. Save the executed agreement, statements of work, invoices, change logs, current access list, DNS zone, deployment status, analytics configuration, tag-manager version, robots directives, XML sitemap state, redirect inventory, and current production crawl. Record which deliverables are complete, open, disputed, licensed, or dependent on third-party subscriptions. The purpose is not theatrical paperwork. It is to preserve evidence of the working state before authority changes.
- Establish business-controlled identities. Create or verify organization-owned administrator accounts using business email addresses and approved multifactor authentication. Confirm that at least two internal people can receive recovery messages. In Search Console, verify that the organization holds an owner role rather than relying solely on delegated user access. Google documents that owners have broader control over users and settings than full or restricted users. [1] In Analytics, confirm internal administrators can manage account or property access and are not dependent on the vendor’s organization. [2]
- Secure the domain and delivery chain. Log in to the registrar through a business-controlled identity, confirm the registrant and recovery details, export the DNS zone, and verify registrar lock settings. ICANN’s transfer process relies on an authorization code obtained through the registrar relationship, so the business should be able to retrieve that code without asking the departing vendor. [4] Confirm access to hosting, CDN, source repository, environment configuration, build provider, deployment logs, and rollback artifacts. Do not transfer the domain merely for symbolism if the current registrar is working; control and recoverability are the objective.
- Export the working record. Preserve Search Console exports, Analytics reports and annotations, tag-manager containers, crawl reports, keyword and landing-page analyses, content briefs, editorial calendars, redirect maps, canonical rules, robots and sitemap configuration, structured-data templates, outreach records, reporting definitions, source graphics, media licenses, repository history, deployment instructions, open tickets, and vendor-created dashboards. An exported PDF is not equivalent to the underlying data or configuration. Ask for editable source files and machine-readable exports where they exist.
- Transfer platform-specific ownership. Complete ownership or administrative transfers while the vendor still has access to assist. For Business Profile, Google separates ownership transfer from ordinary management access and documents a waiting period for some actions after a new owner or manager is added. [3] Confirm the business holds the appropriate role before removing the former provider. Review product feeds, merchant accounts, call-tracking numbers, and advertising-linked properties even when the engagement was nominally “SEO,” because agencies often connect systems across service boundaries.
- Test recovery and real operations. Use internal accounts to log in without the vendor, trigger approved recovery checks, read Search Console data, open Analytics administration, inspect DNS, clone the repository, view deployment logs, download owned media, and verify billing access. In a staging environment, perform a harmless deployment or content change and roll it back. Confirm that internal staff can restore the site from documented instructions. “You have access” is a claim. A successful independent login and recovery test is evidence.
- Rotate secrets and remove access in dependency order. After internal control is verified, disable vendor users, revoke OAuth grants, rotate shared passwords, replace API keys, remove SSH keys, review service accounts, update recovery channels, and inspect recent administrative logs. Change secrets according to a dependency map. Rotating every credential simultaneously can break legitimate integrations and make the source of failure harder to identify. Preserve emergency contact information until the transition window closes, but do not leave unnecessary production access active indefinitely.
- Monitor the transition. Crawl the site after access removal, test forms and analytics events, inspect deployment and error logs, verify Search Console ownership, review DNS and registrar notices, and compare production behavior with the baseline. Watch for missing scheduled jobs, expired tool subscriptions, disabled call tracking, or dashboards that depended on the vendor’s account. Record each variance, its owner, and the corrective action.
03
Failure cases
Stop if the vendor remains the only verified Search Console owner, the only registrar recovery identity, or the only person who can deploy or roll back the site. Removing that account first can turn a commercial disagreement into an outage. Also stop when the domain’s registrant information, transfer authorization, or recovery process is unclear. The domain and DNS chain take priority because losing control there can disable the website, email, and verification records simultaneously.
Do not rotate credentials blindly. A service account may operate deployments, feeds, reporting, or cache purges long after its creator left. Revoke access only after identifying the dependency and a replacement. If the business discovers disputed code, content ownership, licensed media, or unpaid third-party subscriptions, preserve the evidence and obtain appropriate legal or contractual review rather than deleting files or making unsupported ownership claims.
A failed recovery test invalidates the transition plan. So does a platform transfer that leaves the business with a lower role than expected. If production monitoring shows broken forms, missing analytics, changed DNS, failed builds, or inaccessible repositories, use the documented rollback and restore the last verified state before continuing. The goal is controlled separation, not an especially expensive demonstration of resolve.
04
Completion criteria
Offboarding is complete when the business controls every critical account through organization-owned identities, at least two internal administrators can recover access, and no essential system depends on a former vendor’s personal email, phone, billing method, or password vault. The registrar, DNS, hosting, CDN, repository, deployment platform, Search Console, Analytics, tag manager, Business Profile, and critical feeds must each have a verified internal owner.
The evidence package should contain the pre-transition inventory, final access matrix, raw data exports, configuration exports, source files, media-rights records, change history, open-issue list, credential-rotation log, recovery-test results, and post-transition crawl. Production forms, analytics events, scheduled jobs, redirects, canonicals, robots directives, sitemaps, and deployments should remain functional after vendor removal.
The final review should also identify what was not transferred and why. Some vendor tool subscriptions or proprietary systems may end with the contract. That is acceptable when the dependency was disclosed, the business retained its own data, and replacement arrangements are documented. The standard is not possession of every tool the vendor ever used. It is business continuity, recoverable ownership, and a clear record of the site’s operational state after the relationship ends.
References
Sources behind this record
- Managing owners, users, and permissions in Search Console — Google Search Console Help (accessed August 2, 2026)
- Manage access and data-restriction roles in Google Analytics — Google Analytics Help (accessed August 2, 2026)
- Transfer primary ownership of a Business Profile — Google Business Profile Help (accessed August 2, 2026)
- Authorization codes for domain transfers — ICANN (accessed August 2, 2026)
Corrections
Correction history
No corrections recorded.
To report an error, use the public corrections path.
Platform roles and transfer procedures can change.
Contractual ownership and intellectual-property disputes may require legal review.
This operational playbook does not establish who legally owns disputed assets or work product.