Many companies are migrating from WordPress or a proprietary website builder to a headless CMS in the belief that simply changing the technology will organise their content. After implementation, however, it turns out that the new system replicates the old problems. Service descriptions are duplicated across subpages, landing pages are built from scratch, CTAs come in different versions, articles aren’t linked to the services on offer, and the marketing team still needs a developer for every major campaign.
A headless CMS does not organise content automatically. It offers flexibility, but only a well-designed content model turns that flexibility into a business advantage.
Content modelling in a headless CMS is not simply a matter of creating an ‘article’ type, adding a title field and a large text editor. It involves designing the company’s entire knowledge system: services, industries, customer problems, personas, case studies, expert content, CTAs, forms, SEO data and the relationships between them.
As a result, the company does not manage a collection of separate subpages. Instead, it manages structured elements that can be reused across the website, campaigns, sales materials, language versions and responses generated by AI systems.
For B2B organisations, the difference is significant. A good content model enables you to launch campaigns more quickly, build topical authority, respond more effectively to customer intent, and measure which content actually supports lead generation.
What is content modelling, and why isn’t it about the structure of a blog?
Content modelling involves designing content types, their fields, relationships, taxonomies, validation and usage rules. It defines what information a company stores in its CMS, how that information is linked together, and where it can subsequently be published.
In the traditional approach, the team thinks in terms of screens: a service page, a blog post, a landing page, an industry-specific subpage. Each screen is treated as a separate document. The content is tied to one specific layout, which is why it is difficult to use it in other contexts without copying it.
In a structured content approach, we first design the information and only then consider how to present it. The service is therefore not merely a single subpage. It is a separate entity containing a description, the problems it solves, the sectors it is intended for, the technologies involved, FAQs, evidence of effectiveness and related materials.
The same service can later be used on a service page, an industry-specific landing page, in an article, a case study, an enquiry form or a section on recommended solutions.
A mature content model for a B2B company may include:
- services, products, sectors, personas and decision-makers,
- customer issues, use cases and stages of the purchasing process,
- case studies, authors, experts, articles and thematic clusters,
- FAQs, evidence modules, forms, CTAs and lead magnets,
- SEO data, relationships between content and language variants,
- information on campaigns, traffic sources and analytical events.
It’s much more than just a blog with categories and tags. It’s a content system designed to support marketing, sales, SEO, GEO and the further development of the digital product.
Why does the content model affect SEO, rather than just making life easier for the editorial team?
In a poorly designed CMS, SEO is an afterthought. After writing the content, the marketer adds the meta title, meta description and slug. They then manually search for places to link to, ask the developer to implement structured data, and ensure that the author, last updated date and canonical tag are included on the page.
In a well-designed model, SEO is an integral part of the content structure. The CMS doesn’t just store the text. It also makes sure the information needed to publish it correctly, link it to related content and help search engines understand it is in place.
An article model may include search intent, the main keyword, supporting keywords, a topic cluster, funnel stage, author, update date and related services. A service page model may highlight customer pain points, supported sectors, case studies, FAQs and CTAs. Thanks to these relationships, the system can automatically suggest internal links and display relevant content without having to copy it onto every subpage.
In the SEO layer, it is worth planning for things such as:
- meta title, meta description, slug and canonical URL,
- the option to set indexing or noindex,
- the structured data type appropriate to the content,
- the author, the expert and the date of the last review,
- service, sector, problem and thematic cluster,
- related FAQs, case studies and articles,
- recommended CTAs based on the user’s intent,
- relationships used for internal linking.
The data required to generate JSON-LD should be derived from the content model. Information about the organisation, author, service, article, event, product, breadcrumbs or FAQs should not have to be entered manually each time in the code or in an additional text field.
This approach reduces errors and lets you scale publishing. When a company has 20 subpages, many things can still be managed manually. With 200 or 2,000 pieces of content, manually managing links, metadata and relationships becomes impractical.
A good content model ensures that the basic SEO requirements are handled by the system itself. The specialist still plans the strategy and optimises the content, but does not have to correct the same technical issues with every publication.
GEO and AI search are increasing the importance of structured content
GEO, or Generative Engine Optimisation, refers to the optimisation of content for responses generated by search engines and AI systems. It is not a substitute for SEO. It builds on the same foundations: good content, correct indexing, domain authority, information structure and the technical accessibility of the website.
However, the way in which content is used is changing. A traditional search engine displays a list of results. A generative system can construct a direct answer from several sources, summarise a topic, compare solutions or quote a selected excerpt.
For content to stand a chance of being correctly understood, it must clearly present entities and relationships. The system should be able to recognise what a service is, who it is intended for, what problem it solves, who the author of the information is, and when the material was last updated.
That’s why a content model built for GEO should include short definitions, TL;DR sections, FAQs, sources, statistics, authors and their areas of expertise. Fields such as ‘when the solution makes sense’ and ‘when it doesn’t’ are also useful, as are comparisons that answer customers’ decision-making questions.
It is not simply a matter of creating an artificial format exclusively for language models. Content organised in this way is also easier for humans to read. It answers questions more quickly, guides users more effectively to the next subpage, and makes it easier to assess whether the offering is suitable for a particular problem.
From a business perspective, SEO and GEO should not operate as two separate projects. A company needs a single, well-organised knowledge base that can be shared with users, Google, AI tools, applications and other channels.
A content model for B2B should cover the entire demand generation process
One of the most common mistakes is designing a headless CMS primarily for a blog. The company creates articles, authors, categories and tags, but still manages its offering, campaigns and case studies as a collection of manually assembled subpages.
Meanwhile, in B2B, content plays a role at various stages of the purchasing process. A customer might start with an article describing a problem, then move on to a service page, look at a case study, read the FAQs, and only then arrange a meeting. The content model should support this entire journey.
Services as the core elements of the offering
Each service should be a separate type of content or an object with a precisely defined structure. A title and description alone are not enough.
A service model may include:
- a business description and the problems it solves,
- information on who the service is suitable for,
- situations in which implementation is not justified,
- expected outcomes and how results will be measured,
- technologies and integrations,
- related case studies, articles and FAQs,
- recommended CTAs and structured data.
This means that a single service can be presented in different places and contexts. On the homepage, a short description is enough. A dedicated subpage shows the full offer. On a landing page for a specific sector, the system can use only selected problems, evidence and arguments.
Changing the core service details then doesn’t mean editing several, or even a dozen or more, subpages.
Industries as context, rather than copies of service pages
B2B companies often market a similar service to several sectors, but the selling points vary depending on the market. For e-commerce, performance, catalogue integrations and conversion rates will be key. For medtech, reliability, data security and complex user processes may be of greater importance.
A separate industry model lets you define the characteristic challenges, objections, risks, typical integrations, communication style and evidence relevant to that particular sector. An industry-specific landing page does not, therefore, need to be a manually written copy of the service page.
It can be developed from the relationship between an industry, a service, a problem and a case study. Marketing retains control over the message, but does not repeat the same information across multiple documents.
Personas and decision-making roles
A B2B process rarely involves just one person. The business owner, CMO, CTO, sales director and operations manager may all assess the same investment using completely different criteria.
A persona model should describe the goals, concerns, KPIs, typical questions, preferred content formats and the arguments needed to make a decision. This enables the campaign to direct the business owner to the ROI calculator and information on TCO, and the CTO to details on architecture, security and integration.
This does not mean creating a separate page for each person. It is about being able to consciously tailor content to the audience, rather than preparing a single, one-size-fits-all message for everyone.
Problems and use cases
Companies often describe their services solely in terms of what they offer. Customers, however, usually start by identifying a problem: the website is slow, conversion rates are falling, WordPress is becoming difficult to maintain, multilingual content is causing chaos, and existing tools do not integrate with one another.
It is therefore worth modelling problems as separate entities. Each problem can be linked to a service, an industry, an article, a case study, an FAQ and the relevant CTA.
This approach supports SEO, as the content addresses specific search intent. It also helps with GEO, as the link between the problem and the solution is clear and explicit. For sales, it provides a ready-made pool of materials that address the situations potential customers come to you with.
Case studies as a source of evidence
A case study should not simply be a long post published in a single section of a website. It is a source of evidence that can be used in many different places.
The model can store details such as the client, industry, problem, project scope, solution, tech stack, outcome, metrics, quotes and related services. Individual data points can then be automatically displayed on service pages, landing pages and in sales materials.
For example, the 86 per cent increase in ClickLease’s conversion rate, as detailed in our case study, could be featured on a page about conversion optimisation, in the case studies section, or on a landing page aimed at companies facing a similar problem. OpenMed’s 3,782 per cent increase in organic traffic could support content relating to SEO, migration and website performance.
A single update in the CMS changes the data wherever it has been used. The marketing team does not have to copy the figures manually, and the company reduces the risk of presenting different versions of the same result.
Expert articles and the content hub
An expert article should form part of a larger system. In addition to the title, content and image, it requires information on the main topic, cluster, intent, funnel stage, author, sources, date of last update and related elements of the offering.
A blog designed in this way is no longer simply a chronological list of posts. It becomes a content hub, where each piece of content has a specific place and role.
An article about Astro could be linked to a website development service, the issue of poor Core Web Vitals, Cloudflare technology, a relevant case study and a call-to-action for a performance audit. The system can use this as a basis to generate recommended content and support internal linking.
As a result, the company doesn’t just publish more content; it builds topical authority in a controlled manner.
CTAs and conversion modules
CTAs are often entered manually into the content or embedded within website components. After a few months, there are many different versions of buttons, forms and messages in use on the site. It is unclear which ones are current or where they have been used.
A better approach is to create a separate CTA model that includes the funnel stage, user intent, service, industry, button text, supporting copy, test variant and place of use.
This allows the editor to choose from a set of pre-defined options: a discovery call, a quick quote, a Core Web Vitals audit, a redesign ROI calculator or a checklist for migrating to a headless architecture. The CTA is then not just a random element at the end of the page, but one that is tailored to the context.
FAQs as a shared database of answers
A single question may appear on the service page, in an article, on a landing page and in the knowledge base. Copying the answer to each of these places leads to inconsistencies.
It is worth modelling FAQs as separate entities with short and long answers, linked to a service, issue, intent, funnel stage and update date. You can also specify whether a particular answer is to be used in structured data.
This model supports SEO, visibility in AI search and sales, as the organisation builds a well-organised database of answers to real customer queries.
How does content modelling support B2B campaigns?
In many companies, the problem isn’t a lack of ability to set up a landing page. The problem is that each campaign is treated as a separate project.
The marketer rewrites the headlines from scratch, describes the service, looks for case studies, creates a call-to-action and sets up the form. More and more versions of the same offer pile up, and nobody updates them afterwards. The campaign runs for a few weeks, and the content that was prepared is then abandoned.
A well-designed CMS for B2B campaigns works differently. A landing page is made up of controlled elements: industry, problem, service, persona, proof, FAQs, a form and a CTA. Each of these has its own structure, owner and relationships.
This enables the company to create variants tailored to different segments more quickly. The same service will be presented differently for e-commerce, manufacturing or professional services, without having to copy the entire subpage and without losing consistency.
The content model may also include a campaign tag, a traffic source ID, a message variant and analytical events. This makes it possible to measure not only the final form but also the effectiveness of specific arguments, evidence modules and CTAs.
The company is therefore moving away from building individual landing pages. Instead, it is building a campaign system that can be developed, tested and optimised.
How can you avoid chaos in a headless CMS?
The biggest mistake is to replicate the old WordPress structure in the new tool. A headless CMS may run faster and offer greater front-end flexibility, but if the content remains inconsistent, the company will still incur the same operational costs.
Problems tend to arise most often when:
- a CMS models ready-made pages rather than information and relationships,
- a flexible page builder lets anyone create any layout without rules,
- there is no consistent taxonomy of services, sectors, issues and topics,
- the same descriptions, CTAs and FAQs are copied between documents,
- multilingual support was only added after implementation,
- the fields have no validation, owners or publication statuses,
- no one has planned internal linking or how to measure the content.
A CMS that is too rigid restricts marketing. One that is too flexible leads to chaos. A good content model strikes a balance: it gives the editor the freedom to assemble the required content, but within the confines of controlled components, fields and rules.
At WebProfessor, we design this layer in tandem with the information architecture, UX and marketing objectives. As a result, the CMS is not a separate technical task carried out at the end of the project. It forms part of the way in which the website is designed to generate traffic, leads and sales.
Content modelling, multilingualism and entering new markets
Multilingualism is one of the areas that needs to be planned for before implementing a content model. Adding a ‘second language’ as a copy of the document usually only works for small websites. With a larger volume of content, local campaigns and different versions of the offering, it quickly leads to problems.
The model should distinguish between language and market. The Polish version of a website for B2B clients will not always be translated on a one-to-one basis for the German or British markets. Sometimes the examples, CTAs, selling points, case studies, keywords and the way the offer is presented will change.
It is therefore worth taking into account the local URL, hreflang, translation status, the owner of the version in question, the date of the last update, and whether the content has been translated or localised.
A well-designed structure also makes it possible to share those elements that do not require a local version, and to separate those that need to be adapted to the market.
If a company is planning to expand, this model helps to minimise future rebuild costs. It helps to avoid duplicate content, incorrect hreflang tags, out-of-date translations and campaigns that direct users to the wrong version of the website.
Sanity, Payload or WordPress - how do you choose a CMS to suit your content model?
Technology should be driven by content architecture and business objectives, rather than by the team’s habits.
Sanity works well in projects where a company needs a flexible content hub, a comprehensive blog, landing pages, campaigns, relationships between content, and a user-friendly editorial environment. It lets you design schemas, validations and relationships precisely, while also integrating well with Astro and Next.js.
Payload CMS is a good choice when the content structure starts to overlap with application logic. This covers client portals, permissions, complex workflows, admin panels, product data and projects in which the CMS is part of a larger system.
WordPress may still be sufficient for simple implementations or where the client has an extensive, well-established ecosystem. However, when it comes to more advanced SEO, GEO, multilingualism, modular campaigns and a large number of relationships, its structure often requires a number of workarounds.
At WebProfessor, we use Sanity by default for content-based projects, Payload for more bespoke systems, and WordPress only when the client’s specific circumstances genuinely justify this choice.
We build the front end in Astro.js or Next.js, depending on the level of interactivity and application logic. Cloudflare ensures fast content delivery, a low TTFB, caching and security. The result is one system rather than a collection of randomly combined tools.
What does the content model design process involve?
The design of the content model should begin before the CMS is coded. First, you need to understand how the content is intended to support the company’s growth.
The process we follow comprises eight stages:
- Business discovery. We determine whether the content is intended to support SEO, GEO, lead generation, campaigns, sales, education, onboarding or customer service.
- Audit of existing content. We check what generates traffic and conversions, what is out of date, what has been duplicated and what is missing.
- Entity and topic map. We map out services, industries, personas, problems, technologies, case studies and the relationships between them.
- Content type model. We define fields, components, validations, statuses, editorial roles and reuse rules.
- SEO and GEO layer. We establish data for metadata, schema, authors, sources, FAQs, definitions, updates and internal linking.
- Campaign components. We design CTAs, hero sections, forms, proof points, comparisons, testimonials and lead magnets.
- Editorial workflow test. We check whether the marketing team can create a service page, landing page, article and campaign without involving a developer for every change.
- Measurement and iteration. We establish a baseline, implement the solution and measure visibility, conversion, lead quality, Core Web Vitals and content effectiveness.
This approach allows the model to be assessed not merely on the basis of whether ‘a document can be saved’. What is more important is whether the system works in day-to-day marketing operations and whether it will remain clear and understandable when the volume of materials increases many times over.
What does this mean for a B2B company?
Content modelling in a headless CMS has implications far beyond simply keeping the editorial dashboard organised.
For SEO, this means consistent topic clusters, improved internal linking, well-organised structured data and easier content updates. For GEO, this means clear entities, definitions, authors, sources, FAQs and relationships that help AI systems interpret the company’s knowledge correctly.
For B2B campaigns, it enables the rapid creation of landing pages using ready-made, tried-and-tested modules. For sales, it creates a library of materials tailored to the needs of different personas and addressing the objections that arise during the decision-making process.
It also has an impact on TCO (total cost of ownership). The company spends less time copying, editing and synchronising content. It is less likely to need to call on a developer for simple changes. Nor does it have to rebuild the CMS every time a new channel, language or type of campaign is introduced.
A good model should work just as well with 20, 200 or 2,000 pieces of content. If the architecture only works well with the first few subpages, it is not a structure that is ready to scale.
A headless CMS only offers an advantage if the content model has been designed with the company’s growth in mind. Without it, the organisation simply ends up with a more modern tool, but with the same chaos it previously experienced with WordPress, a website builder or a collection of separate landing pages.
