Most CMS migrations need a system ownership map before content moves. The safe approach is to define what the CMS owns, what it references, what remains external, and how every major URL type will be handled. In complex, product-heavy content ecosystems, that ownership work matters more than the content import itself. This article explains the blueprint decisions that matter before implementation: coverage mapping, product references, DAM ownership, editor permissions, design-to-CMS mapping, and migration treatment.
I’m Matija Žiberna, founder of BuildWithMatija. I work on CMS architecture, Payload implementation, content modeling, and migration planning for companies with messy digital ecosystems. This piece draws on patterns that show up repeatedly in large CMS migration blueprints and uses a composite scenario rather than a single client case.
A representative migration context
Consider a product-heavy company migrating from a legacy CMS to a more structured content platform.
Different sites may serve different audiences. One may be consumer-facing, another more professional or practitioner-oriented, and another may sit closer to ecommerce while inventory and order logic remain in operational systems.
The ecosystem often contains far more than standard pages. Typical areas include product pages, listing pages, blogs, recipes, events, giveaways, lead magnets, campaign landing pages, microsites, older educational content, quizzes, store locator areas, contact forms, gated resources, ecommerce-adjacent pages, community links, PDFs, downloads, and redirects.
The useful migration work starts once each area is classified by its future role. Some content belongs in the CMS as structured collections. Some content should stay external. Some content needs product-system references. Some pages deserve manual rebuild. Some URLs should redirect or retire.
That classification becomes the spine of the blueprint.
Start with a one-to-one coverage map
The first practical deliverable was a one-to-one website coverage map.
The goal was simple: every current website area needed a future home, a future owner, or an explicit retirement path.
A coverage map prevents the migration from becoming a vague promise to “move the site.” It gives leadership, developers, marketers, and stakeholders a shared view of what will happen to each part of the existing ecosystem.
This table format changed the conversation. Stakeholders could see that migration was a set of treatment decisions, rather than one bulk export/import task.
Define what Payload owns
Payload’s role needed to be defined before schema work.
In many organizations, there are already existing systems for product data, CRM capture, ecommerce operations, inventory, orders, community activity, account resources, and asset governance. The CMS should become the website content, publishing, workflow, and reference layer rather than a replacement for every neighboring system.
That ownership model keeps the CMS focused.
This table became a core architecture artifact because it stopped the project from turning into a general platform replacement.
Payload can technically be extended in many directions. The blueprint needed to decide which responsibilities belonged there.
Model product references instead of copying product truth
The most valuable discovery was product traceability.
Product information appeared across product pages, blogs, recipes, campaign pages, PDFs, educational content, lead magnets, banners, ingredient references, product images, and microsites.
That created a maintenance and compliance challenge. A product, ingredient, image, PDF, or claim-sensitive association could appear in many places outside the product detail page.
The first useful Payload model does not need to auto-update every instance across the website. In this project, the safer first target is visibility.
Payload should help the team see every URL where a product appears, every recipe connected to a product, every article mentioning an ingredient, every page using a specific product asset, and every campaign connected to a product line.
That turns content into a relationship graph.
A blog post becomes more than rich text. It can reference products, topics, ingredients, authors, claims, assets, downloads, and review status.
A recipe becomes more than a formatted article. It can connect ingredients, dietary preferences, health goals, related products, creator profiles, lead magnets, and images.
A product education page becomes more than a marketing page. It can connect approved assets, related products, claims, FAQs, downloadable files, and CRM-connected lead forms.
This is where Payload’s content model creates operational value. The CMS gives teams a way to find affected content later.
Define editor access by role, team, brand, and content type
Many organizations need marketing teams to manage repeatable content without sending every small request to developers.
That did not mean broad admin access.
One useful permission model starts with four base roles:
Role
Purpose
The important part was the layering.
Access needed defaults by team, configuration by group, exceptions by user, and restrictions by brand or content type.
A marketing editor might manage blog posts, recipes, campaign pages, and giveaways for a specific site or business line. A channel or campaign owner might manage retailer-related copy, banners, or resource pages. A product stakeholder might review product-related content. Legal or regulatory stakeholders might need approval rights without full editing rights. Developers and admins would retain access to technical configuration.
This structure gives teams autonomy inside safe boundaries.
Map the design system to CMS control layers
A redesign prototype and component library often introduce another critical decision layer.
The prototype already had reusable cards, product layouts, blog layouts, recipe layouts, event layouts, store cards, health-goal visuals, ingredient visuals, marketing blocks, banners, and featured product sections.
Those components still needed CMS interpretation.
A frontend component does not automatically become a CMS block. A visual section does not automatically become reorderable by editors. A content-looking area can still be locked if it affects brand consistency, performance, or design integrity.
The CMS mapping looked closer to this:
This mapping protected both sides of the project.
Editors get useful control over content, references, CTAs, banners, assets, and publishing workflow. Designers and developers keep control over core templates, interaction logic, and component behavior.
Treat DAM as an upstream system
Asset review usually shows that website media is only one part of the asset problem.
Many teams need a broader asset governance model covering approved product images, lifestyle photography, campaign assets, PDFs, social assets, sensitive files, versioning, statuses, metadata, brand visibility, and team permissions.
Payload can manage website media. It should consume approved assets for website content. The broader DAM requirement belongs upstream because it serves more than the website.
That distinction affects schema design.
Content records in Payload can reference approved assets, use asset metadata, and expose selected assets on pages. The DAM remains responsible for approval, version control, sensitive files, and company-wide asset organization.
This also helps with future product traceability. A product image used in blogs, recipes, microsites, banners, and PDFs should be findable through references, rather than lost as disconnected uploads.
Convert migration into treatment decisions
A large migration needs more than an export strategy.
Each URL type needs a treatment decision.
This classification gives the implementation team clear migration rules.
Automated migration makes sense for some repeatable content. Manual rebuild makes sense for important marketing pages, campaign pages, microsites, and content with weak structure. Retirement makes sense for outdated primers, old campaigns, or low-value URLs that should not move into the new system.
The migration plan becomes safer once every content type has a treatment pattern.
The blueprint prevents developers from making business decisions by accident
A vague implementation brief forces developers to make operating-model decisions inside the codebase.
They decide which fields are editable. They decide how products are referenced. They decide which templates are flexible. They decide how approvals work. They decide where form settings live. They decide how assets are selected. They decide how old URLs redirect.
Those decisions have business consequences.
The blueprint makes those decisions explicit before implementation.
For migrations like this, the blueprint needs to define site structure, multilingual setup, collections, reusable blocks, templates, product references, DAM direction, CRM boundaries, quiz options, store locator boundary, ecommerce boundary, community and portal exclusions, migration strategy, SEO preservation, redirects, roles, permissions, approvals, notifications, editor experience, infrastructure direction, risk register, open decisions, and roadmap.
That document gives the internal team a buildable foundation. It also gives leadership a clearer view of scope, risk, sequencing, and ownership.
Practical checklist for a serious CMS migration
Use this checklist before implementation starts:
FAQ
Should every CMS migration need this much discovery?
Small brochure sites usually need a lighter process. Multi-site, multilingual, product-heavy, ecommerce-adjacent, regulated, or workflow-heavy sites need deeper discovery because the architecture decisions affect permissions, source-of-truth data, migration safety, and long-term maintainability.
Why should product data stay outside Payload?
Product data often has its own approval workflow, regulatory rules, field mapping, label logic, SKU structure, and downstream dependencies. Payload should reference product records and manage website-specific enrichment around them.
When should content become a structured collection?
Content should become a structured collection when it repeats, needs filtering, has relationships, needs approval, requires migration rules, or should appear consistently across the frontend. Blogs, recipes, events, giveaways, resources, lead magnets, and microsites are common examples.
Should a DAM be separate from Payload?
A small website can use Payload media directly. A company with approved assets, versioning, sensitive files, trade materials, product imagery, metadata, brand-specific access, and external partner needs should treat DAM as an upstream source of truth.
How much control should editors have?
Editors should control structured content, approved blocks, references, CTAs, assets, scheduling, and review submissions. Core templates, product truth, global logic, complex integrations, and sensitive system settings should remain restricted.
Conclusion
A serious CMS migration starts with ownership.
In complex content migrations, the highest-value work is often defining what the CMS should own, what it should reference, what should stay external, and how each content type should move into the future system.
That clarity gives the implementation team a safer foundation. It gives marketing teams useful autonomy. It protects product truth, asset governance, SEO value, and approval workflows. It also prevents the new CMS from inheriting every structural problem from the old one.
Start with the ownership map before you start the build.
Thanks,
Matija
Current area
Future handling
Main brand pages
Payload pages or structured templates
Product detail pages
Fixed frontend templates fed by product-system references
Product listing pages
Payload-managed presentation with product-system data
Blog posts
Structured Payload collection
Recipes
Structured Payload collection
Giveaways and contests
Structured campaign or giveaway collection
Events
Structured event collection or external registration link model
Lead magnets
Payload landing pages with CRM-owned submission flow
Product microsites
Microsite model, template, bespoke rebuild, or retirement review
Educational primers
Preserve, redesign, consolidate, or redirect based on value
Quizzes
External quiz tool, embedded quiz, or future Payload quiz model
Store locator
External locator data with Payload-managed surrounding content
Ecommerce pages
Existing ecommerce and operational systems remain responsible
Gated resources
Boundary review across portal, account system, DAM, and Payload
Community areas
External community platform remains responsible
PDFs and downloads
Asset migration, DAM mapping, redirect review, or retirement
Old campaign URLs
Redirect, archive, rebuild, or retire based on traffic and relevance
Area
Payload owns
Payload references
External system owns
Standard pages
Page content, page status, SEO fields, preview, publishing
Approved assets, related content
No external owner unless the page embeds external functionality
Product pages
Website-specific enrichment, optional content blocks, related content
