Short answer

Divide a rebuild into diagnosis, scope agreement, content and journey design, a test implementation, acceptance and a controlled launch. Every stage should leave a verifiable deliverable. Protecting SEO means reducing risk and monitoring, not guaranteeing that traffic will be maintained.

What are you buying: a new look or a solution to a problem?

Before requesting a quote, record what currently hinders customers and the team. Is a service hard to find, are form messages not arriving, is the offer outdated, or does adding content require a developer? Separate observations from hypotheses. “We did not receive the test message” is a fact; “the button colour puts customers off” still needs testing.

The process below is my recommendation for organising the project, not a universal schedule. I do not give a price or number of weeks without knowing the scope. A small service website and a store with integrations require different tests. In both cases, commission specific deliverables rather than just a promise of a modern website.

1. Audit: a shared baseline

The first stage should answer three questions: what works, what does not and what must be preserved. The provider needs an inventory of public URLs, information about forms, access granted through appropriate roles, and traffic and enquiry data where available. Record missing analytics as a diagnostic limitation rather than replacing it with an invented conversion rate.

The deliverable should be more than a presentation of screenshots. Ask for an issue register: URL, observation, evidence, business significance and proposed action. The business owner sets priorities; the provider explains technical dependencies. Separate defects needing repair from ideas that can be implemented after launch.

2. Scope: decisions before visual design

Agree the page inventory, language versions, features and integrations. Specify who prepares copy, photos and legal information, and who approves them. Changing CMS does not automatically require changing every URL. Keeping correct URLs can reduce migration work, but content and technical settings still need checking.

Record exclusions in the scope document: what the quote does not cover. A new booking system, moving mailboxes or translating dozens of articles should not appear unexpectedly on acceptance day. Price and approve scope changes separately before they affect the deadline.

3. Design: the customer journey first

The wireframe should show how a visitor moves from a question to an action. Check the homepage, service page, contact and form messages as one sequence. Does someone unfamiliar with the business understand the offer? Do they know what happens after sending an enquiry? Does important information disappear on a phone?

Approve content at this stage too, not just colours. Later removing a substantial service page and replacing it with two sentences can change the URL’s topic and usefulness. For a technical preservation checklist, see the SEO checklist before a rebuild. This guide instead covers organising the whole project and deciding whether to accept it.

4. Staging: a working version, not a set of pictures

The test environment should run the actual templates and features intended for publication. A mock-up alone cannot test form error handling, navigation or redirects. Test data must be safe; do not unnecessarily copy real customer data to a location accessible to outsiders.

Staging should not compete with production in search. Google recommends preparing a list of noindex directives that must be removed from the appropriate pages at launch. Access control and exclusion from indexing are different things: noindex does not protect confidential materials. Agree protection of the environment with the provider and check production settings again after deployment.

An acceptance matrix: what to require from the provider

The table below proposes a division of responsibilities. Assign a named person to each role and add criteria relevant to your business. Acceptance should rely on evidence rather than a general “done” message.

Stage and deliverableWho approvesAcceptance evidence
Audit and prioritiesBusiness ownerIssue list with URLs and observations
Scope and content mapOwner and providerPage, feature, exclusion and responsibility inventory
Test versionThe person accepting the projectKey journeys exercised on phone and desktop
URL migrationTechnical specialistOld → new URL map and response-test results
LaunchDesignated decision-makerTest record, backup and rollback plan
Post-launch careOwner and website maintainerAgreed monitoring scope and issue-reporting channel

5. The launch gate: when to say “not yet”

Before launch, plan data and file backups and a way to restore the previous version. Knowing that the hosting provider makes backups does not establish what they contain or how long restoration takes. Agree who decides on rollback and how to preserve enquiries or orders created after the switch.

Test both valid and invalid form completion, and receipt of the message by an authorised person. The screen message and actual arrival of the enquiry are two separate checkpoints. Agree tests so that an artificial enquiry does not trigger unnecessary customer-service work.

6. After launch: separate technical acceptance from observation of results

After publication, repeat the tests on the public domain. Operation on staging does not prove production has identical settings. Check final URLs, essential content, contact, consent for optional analytics tools and the way data is collected. Record the deployment date so that later you can distinguish website changes from other business events.

For URL changes, Google’s documentation recommends updating links and sitemaps and monitoring the migration. It also warns about possible ranking fluctuations. An HTTP 200 response does not mean the URL is indexed or traffic has been preserved. Compare data over comparable periods, accounting for seasonality and campaign changes, instead of judging the entire rebuild after one day.

How to compare rebuild proposals

Compare proposals by the deliverables in the table, not the number of impressive screens. Ask whether the price includes an inventory, content migration, mobile tests, redirects and post-launch care. Agree who retains access and how documentation is handed over. An offer without these answers may simply cover different work, rather than being a cheaper version of the same work.

My recommendation: begin with a short description of the current website and its problems. Choose a full rebuild or smaller upgrade only after diagnosis. Not every difficulty requires a new website, but every substantial change needs a decision owner and verifiable acceptance criteria.

Source and limits of the conclusions

The recommendations on URL migration, noindex, mapping, links and monitoring are based on Google Search Central’s site-move documentation. The stages, responsibility matrix and acceptance gate are the author’s recommendations. This guide promises no SEO outcome, presents no fictional case study and does not replace an assessment of a particular website.

Agree the rebuild scope before requesting a quote

In the brief, describe your current website, goals and problems. This is a starting point for discussing scope, not an automatic quote or a guarantee of results.

Open the website brief (in Polish) ↗