How to Leave a Marketing Agency: Contractor Website and Ranking Checklist

Leaving a marketing agency need not mean moving your domain or rebuilding your contractor website. Decide what actually must change. Secure company-controlled access, prove a backup works, and test the replacement before anyone cancels the old service. Keep the incumbent available until calls, forms and the office handoff pass live checks. These checks help protect customer contact during the move. Monitor search performance after any URL or content change.[1]

Start with the smallest safe change. Keep a working site when the contract and access permit it. Repair a broken lead path before buying another build. Plan a rebuild only when the required system cannot transfer or be maintained on acceptable terms.

Plan the handoff before you give notice

KKS can discuss the current site, the assets you control and the lead paths that need testing. Agree the scope and timing of any migration work before the handoff.

Request a consultation

The main risk is an access gap

An agency change does not require a domain migration: the contractor can retain the domain while changing the people, hosting, content system or marketing tools. Keeping useful URLs at their existing addresses avoids one large source of search disruption, although the remaining technical and content changes still need testing.

Access gaps are operational. The old agency may control the registrar login while nobody at the contractor’s company has WordPress administrator access. A contact form may send to an agency mailbox, or Analytics may have reports without a company administrator who can add users. A call-tracking number on every page is another custody problem until its portability and routing are checked.

Find those gaps while the relationship still works, because a disputed departure makes account recovery and lead routing harder to coordinate.

Our contractor website ownership checklist covers the contract questions to settle before signing; this guide covers the practical handoff at the other end, when a contractor needs the site, accounts and customer contact routes to keep working after the agency leaves.

What the platform guidance actually covers

Google distinguishes a hosting change that keeps the same URLs from a move that changes them. Its checklists call for testing, monitoring and, where URLs change, mapping old pages to relevant new pages. A change of agency by itself does not require either type of move.[1][2]

Verified factKKS analysis
Google recommends preparing a URL map, testing the new site thoroughly, using redirects, and monitoring the move. Google also warns that significant site changes can cause temporary ranking fluctuations while pages are recrawled and reindexed.[1]Preserve good URLs unless there is a documented reason to change them. A new agency does not need a new address for every page.
When hosting changes without URL changes, Google recommends copying and testing the site first, then changing DNS, monitoring old and new hosting, and shutting down the old system only after users and Googlebot receive the correct content.[2]A hosting move needs an overlap period. Cancelling the old host before the new copy passes live tests creates avoidable downtime.
Google recommends permanent server-side redirects where possible when a page’s search URL must change.[3]Redirects belong in the migration plan before launch. They should not be improvised after 404 errors appear.
Google Analytics access can be granted at the account or property level. Adding or changing users requires the Administrator role at the relevant level.[5]Seeing reports is not control. The contractor needs an administrator account that the departing agency does not own.
A Google Business Profile can have multiple owners but only one primary owner. Managers cannot add or remove users.[6]The contractor should hold primary ownership. Agency personnel normally need managed access, not custody of the listing.
Google Tag Manager warns that a business can be locked out when its sole administrator leaves and recommends at least two active administrator accounts.[7]Use one company-controlled administrator and a second recovery administrator. Remove the old agency only after the new access works.

Keep useful URLs, test redirects and monitor the pages after the move.

Phase 1: build the asset register before notice

Create one record with the asset, public URL or account ID, legal owner, current administrator, recovery email, billing owner, renewal date, and transfer status. Do not store passwords in the spreadsheet. Use a password manager for credentials and the register for custody.

At minimum, inventory:

  • domain registrar, DNS host, nameservers, and domain renewal;
  • web host, CDN, SSL certificate, backups, and staging site;
  • WordPress or other CMS administrator accounts;
  • theme, child theme, plugins, custom code, and license keys;
  • all live pages, posts, media, forms, downloads, and redirects;
  • Google Business Profile, Bing Places, and major trade listings;
  • Google Analytics, Search Console, Tag Manager, Ads, and Merchant Center when used;
  • call-tracking numbers, recorded greetings, routing rules, and number-porting terms;
  • form destinations, CRM connections, online booking, email delivery, and spam controls;
  • creative files, logos, photos, video, ad copy, reports, and keyword research;
  • agency-created email addresses and any account that uses them for recovery.

Mark each item green, yellow, or red. Green means the contractor can sign in and add another administrator. Yellow means access works but ownership, billing, recovery, or export rights are unclear. Red means the contractor cannot control or retrieve the asset.

A red item affecting the domain or lead delivery is a hold, not a reason to rush the departure. A polished new homepage does not compensate for a domain the company cannot renew.

Phase 2: collect a working website package

Ask for more than “the website files.” A usable WordPress handoff normally needs the database, uploads, theme and child-theme files, active plugin list, custom code, redirects, form settings, and a record of external services. It also needs the information required to run those components legally after the agency leaves, including any license limitations.

WordPress has a built-in export that can create a WXR XML file containing posts, pages, custom post types, comments, custom fields, categories, tags, taxonomies, and users.[8] That export is useful, but KKS does not treat it as a full operational backup. It does not, by itself, prove that the theme, plugin files, media library, configuration, forms, redirects, or server behavior can be restored exactly.

Where paid licenses belong to the incumbent, the incoming provider should price replacements or secure transferable rights before treating a restore test as approval to cut over.

Request these deliverables:

  1. A recent full backup with a database and site files.
  2. A WordPress WXR export as a second, portable content copy.
  3. A list of active plugins, versions, license owners, and renewal dates.
  4. Theme and child-theme files, plus the source files for custom design work.
  5. Current redirect rules and any redirects managed outside WordPress.
  6. Form configuration, recipient addresses, integrations, and recent test results.
  7. A list of scheduled tasks, webhooks, API connections, and external scripts.
  8. A list of anything that will stop working when the agency account closes.

If the site is not portable, establish what does transfer. Our analysis of static files versus a working CMS handoff explains why a visual copy and an editable production system are different assets.

Phase 3: preserve the URL and search record

Export the current URL list before changing content. Pull URLs from the XML sitemap, CMS, analytics landing-page reports, Search Console, backlink data, paid landing pages, and a crawl of the live site. No single source catches everything.

For each URL, record:

  • current HTTP status;
  • page title and canonical URL;
  • indexability;
  • organic clicks or impressions when available;
  • meaningful external links;
  • internal links pointing to it;
  • conversion role;
  • keep, improve, merge, redirect, or retire decision;
  • exact destination when redirected.

The safe default is to keep a useful URL at the same address. If a page must move, map the old URL to the closest relevant new page. Do not send every retired service page to the homepage.

Google says a site move is processed per URL and recommends mapping old URLs to corresponding new URLs.[1] Use a page-by-page map rather than a single rule that sends the old site to the new site.

Use a permanent server-side redirect where possible for a permanent page move.[3] Update internal links so they point directly to the final canonical URL. A redirect should catch stale bookmarks and outside links, not become the site’s normal navigation system.

The Search Console Change of Address tool is for a domain or subdomain move, such as moving from one domain to another. It is not required merely because the agency changes while the domain remains the same.[4]

Who signs off on the move?

Name people, not only vendors. The owner approves the scope, contract notice and rollback authority. The incoming website provider restores the site, tests URLs and documents DNS, redirects and backup recovery. The incumbent supplies the permitted export, current routing and account handoff while its service is active. The office lead confirms that test calls and labeled form or booking requests reach the right person and can be followed up. Give each person a contact route and a cutover window in the handoff record.

Use the 17 ownership questions to settle rights and account custody. This page tests whether the handoff actually works. For a site connected to ServiceTitan, the ServiceTitan handoff guide adds product and account checks.

Phase 4: test the replacement before the cutover

The incoming provider should restore the site in a protected staging environment and test it without relying on production as a safety net. Google recommends testing pages, images, forms, and downloads before a hosting change.[2]

Run a practical acceptance test:

  • crawl every staged URL and compare it with the current inventory;
  • confirm one H1, the intended title, and the intended canonical on each important page;
  • with permission on the company-owned test or live environment, submit labeled test inquiries and confirm office receipt; do not submit a third party’s form;
  • call each company-controlled tracked number and have the office confirm routing and answer coverage;
  • test online booking from a phone and desktop browser;
  • verify analytics and conversion events in their live debug tools;
  • inspect robots.txt, noindex rules, XML sitemaps, and structured data;
  • check images, PDFs, project galleries, and before-and-after sliders;
  • verify mobile navigation, click-to-call buttons, and sticky CTAs;
  • confirm 404 handling and every planned redirect;
  • search the staged source for old agency email addresses, domains, phone numbers, and tracking IDs.

A green confirmation is not delivery evidence. The office must find the labeled inquiry in the intended inbox or CRM, check its fields and verify any promised notification and customer acknowledgment. Run tests only on systems the company controls or authorizes.

Build a marketing stack you can change one piece at a time

Your marketing stack is the set of tools connecting a homeowner’s first visit to an inquiry and a booked job. KKS recommends choosing those tools so that replacing one does not force a website rebuild or interrupt customer contact. This does not require buying every tool from a different company. It requires knowing which accounts, data and connections survive a change.

An all-in-one service can reduce administration. Separate tools can offer more choice but create more connections to maintain. Judge either arrangement by a practical question: if this component stopped meeting our needs, what would replacement require?

Write a replacement test for each important component

Ask the provider to demonstrate the export or handoff for each tool you plan to replace.

Component being replacedKeep stable where possibleEvidence needed before cancellation
Website hostDomain, useful page URLs, content and inquiry destinationsA restored site, preserved DNS records, working forms and a rollback copy
Website designer or managerCompany-controlled accounts, useful content, analytics history and customer contact routesAdministrator access, source files, license terms and a named maintenance owner
Form or booking toolThe public service pages and the person responsible for respondingField mapping, service-area and availability rules, and a labeled inquiry received in the correct destination
CRM or field-service softwareBusiness identity and the customer-facing websiteA trial import that checks contact and job records, notes, attachments and identifiers separately; a plan for anything the export omits
Call-tracking providerCustomer-facing numbers where a transfer is permitted and feasibleWritten transfer requirements, routing and text-message checks, and confirmed completion before the old account closes
Advertising managerCompany-controlled advertising and measurement accountsVerified access, billing responsibility, current campaign records and working conversion destinations

Account access, exportability and a working replacement are separate checks. WordPress documents a WXR content export, including posts, pages, custom fields and other content records.[8] KKS would still require a restore test before treating that file as evidence that the complete website can run elsewhere. Apply the same discipline to a CRM export: a downloaded spreadsheet alone does not establish that job history, files or automations transferred.

Map the connections, not just the subscriptions

For each connection, record the sending tool, receiving tool, fields passed, account that authorizes it, expected delivery time, failure alert and person responsible for fixing it. Keep credentials in a password manager, not in this register. Record whether an integration depends on a separate paid plan or on an agency account that will close.

For example, a website form might feed a CRM, which then alerts the office and sends a customer acknowledgment. Replacing the form is not complete when the new button works. The test inquiry must reach the correct record and recipient, with the expected service details. Check that the old connection does not also create a duplicate record or send a second acknowledgment.

Replace the weak component first, then verify the whole path

Choose the change from the observed problem. A missed form notification calls for a delivery investigation, not an automatic redesign. An unusable booking process may need different configuration or a replacement tool. Neither finding, by itself, proves that the domain, hosting or advertising accounts should change.

Google’s hosting guidance follows a staged sequence: prepare and test the replacement, change DNS, monitor both environments, and shut down the old hosting only after the new infrastructure serves visitors and Googlebot correctly.[2] KKS applies that continuity principle to the broader handoff: name the acceptance test and rollback owner before changing a live dependency.

A useful handoff note states what changed, what stayed unchanged, who tested it, what reached the receiving system, and when the old component can safely close. Follow the account-control checklist for access and ownership questions. Use the cutover sequence below for execution.

Preflight: pass, hold or recover

Before the cutover, the incoming provider records the restore result, URL map, DNS record copy, redirect checks, lead-delivery tests and a known-good backup. The owner signs off only when the office confirms receipt and a rollback operator has a tested route back. Hold the change if the domain or DNS cannot be controlled, a critical page is missing, the old environment is already scheduled to close, or a call or inquiry cannot be traced to its destination. The owner and incumbent then agree on more overlap or a revised notice date within the contract.

If the live change fails, the named incoming provider stops further changes, restores the known-good route where feasible and retests the exact failed path with the office. Keep a log of DNS changes, redirects, backup versions and test times. DNS changes can take time to propagate; allow for that delay in the recovery plan. The owner authorizes closure only after the site and lead paths pass again.

Phase 5: sequence the cutover

A controlled handoff gives the incoming operator enough time to verify access, restore the backup and test customer contact before the old operator is removed.

1. Establish company-controlled administrators

Add and verify the contractor’s administrator accounts in Analytics, Search Console, Tag Manager, Business Profile, the CMS, hosting, registrar, DNS, call tracking, and CRM. Do not remove the old agency yet.

2. Freeze unnecessary production changes

Pick a cutover window. Record any lead, content, form, or order data that changes after the final backup. A quiet weekday may be better than Friday afternoon for a contractor whose weekend emergency calls matter.

3. Take the final backup and export

Capture the database, files, content export, redirect map, DNS records, and account-user lists. Hash or timestamp the package so everyone knows which copy became the migration baseline.

4. Change one major dependency at a time

The incoming provider changes hosting or DNS, then checks the public site from an ordinary visitor connection. It tests the URL sample and redirect map with the owner. The office checks calls and labeled form or booking inquiries. Change another dependency only after the first test passes. Remove old users last, subject to contract rights; changing everything at once hides the cause of a missed lead.

5. Keep the old environment available

Google recommends monitoring traffic on both old and new hosting and shutting down the old infrastructure only after the new environment serves users and Googlebot correctly.[2] Keep rollback available until the office verifies calls, forms and booking, and the provider checks analytics and search crawling.

What to monitor for the first 30 days

Before changing anything, save a dated baseline for the same pages, calls, form deliveries and Search Console views you will check afterward. The incoming provider reviews errors and indexing daily during the first week, then weekly for the remainder of the month; the office reconciles inquiries with actual receipt each workday. The owner asks for the exception log and decides whether unresolved failures require recovery work.

Watch:

  • uptime and server errors;
  • 404s and redirect failures;
  • submitted forms, delivered forms, and booked appointments;
  • call volume and routing failures;
  • indexed pages and sitemap processing;
  • Search Console clicks, impressions, and important query groups;
  • organic landing-page traffic;
  • analytics and tag collection;
  • Google Business Profile calls, website clicks, and booking actions when available;
  • security, backup, and plugin-license failures.

Do not rewrite every service page after one quiet morning. Google notes that significant site changes can cause temporary ranking fluctuations while its systems recrawl and reindex pages.[1] First check whether the technical system and lead paths work. Then separate a real trend from normal daily movement.

When to give the old agency notice

Give notice after the contractor can answer yes to these questions:

  • Can the company renew and control the domain without the agency?
  • Can a company-controlled administrator add or remove users from each critical account?
  • Has a full working backup been restored and tested?
  • Is every current URL accounted for?
  • Are redirects prepared for every necessary URL change?
  • Have forms, calls, booking, analytics, and Search Console been tested?
  • Are contract notice periods, final invoices, license terms, and data-return duties documented?
  • Is there a rollback window if the cutover fails?

This is operational planning, not legal advice. If the contract is disputed or the agency claims ownership of a critical asset, have counsel review the signed agreement before taking an irreversible step.

KKS position: improve after continuity is secure

A new provider may propose a redesign, new navigation, renamed service pages and a different tracking stack in the same week. Defer changes that are not needed for continuity.

First preserve what works. Measure the baseline. Fix broken custody and conversion paths. Then decide whether the site needs repair, a managed growth program, or a rebuild.

KKS manages contractor websites through agreed work on the site and lead paths rather than a fixed article quota. The first action after an agency exit may be a service page. It may be a redirect repair, a form correction, a declining-page refresh, or a cleaner booking path. Read what monthly contractor website management should control before comparing retainers.

Bring the exit plan to a consultation

Tell KKS what must stay live, who controls the domain and where the office receives inquiries. We can discuss a scoped repair, ongoing management or a rebuild when the evidence calls for one. Bring the account and license terms so the next steps can be scoped.

Request a consultation

Frequently asked questions

Will changing marketing agencies hurt contractor website rankings?

Changing agencies alone does not require new URLs or a new domain. Risk rises when the move creates downtime, missing pages, changed URLs without relevant redirects, crawl blocks, lost content, broken internal links, or weaker pages. Significant site changes can also produce temporary ranking fluctuations while Google recrawls and reindexes the site.[1]

Should I tell the agency before securing administrator access?

Secure and test company-controlled access first, within the rights granted by your contract. Early notice can start account closure, license removal, or staff offboarding. The goal is not secrecy. It is avoiding a period when neither team can operate the site.

Is a WordPress export enough to move the website?

Not usually. The WordPress WXR export can carry posts, pages, custom post types, comments, custom fields, taxonomies, and users.[8] A working migration may also require the database, uploads, themes, plugins, custom code, configuration, forms, redirects, licenses, and server settings.

Do I need the Search Console Change of Address tool when changing agencies?

No, not when the site stays on the same domain. Google says the tool is for moving a site from one domain or subdomain to another.[4]

When should the old agency lose access?

Remove old users after the new administrator accounts work, the final backup is secured, the cutover has passed live tests, and the contract allows removal. Keep a dated user-access record so the contractor can show who retained control after the handoff.

Sources

  1. Google Search Central: Site moves and migrations. Rechecked October 24, 2026.
  2. Google Search Central: Changing hosting without changing URLs. Rechecked October 24, 2026.
  3. Google Search Central: Redirects and Google Search. Rechecked October 24, 2026.
  4. Google Search Console: Change of Address tool. Rechecked October 24, 2026.
  5. Google Analytics: Manage users and access. Rechecked October 24, 2026.
  6. Google Business Profile: Manage owners and managers. Rechecked October 24, 2026.
  7. Google Tag Manager: Managing users and permissions. Rechecked October 24, 2026.
  8. WordPress: Tools Export screen. Rechecked October 24, 2026.

Get website and marketing help for your contracting business

KKS builds and manages contractor websites, handles local SEO, and researches your competitors to guide the work. Explore our managed website plans or start with a free audit comparing your business with three competitors.

View website and marketing plans Request my free competitor audit