For public product, service, case-study, article and FAQ pages, start with an approach that returns the main text and links in the initial HTTP response. Server-side rendering (SSR) suits frequent editing when someone maintains the application server. Static site generation (SSG) suits a predictable publishing schedule when someone owns generation and deployment. Signed-in workspaces and complex operating interfaces can use client-side rendering, with their interaction requirements assessed separately from public content pages.
Turn the procurement question into observable tasks: what can a customer read by opening a product URL directly, when will an update appear, and how can they submit an enquiry? This checklist is for marketing owners, website buyers and delivery teams. Use the self-hosted website and maintenance guide to review the wider operating scope.
1. Choose by publishing needs and maintenance ownership
SSR generates page HTML on the server at request time. SSG creates page files ahead of time and publishes them for visitors. Client-side rendering (CSR) relies mainly on browser JavaScript to retrieve data and construct the interface. SPA describes navigation within an application; an SPA can also combine server or static output. Here, the comparison refers specifically to a client-rendered SPA whose initial response mainly contains an application shell.
| Approach | Consider it for | Assign an owner for |
|---|---|---|
| SSR | Products and news edited frequently, with timely public updates | Application, database, capacity and cache updates |
| SSG | Relatively stable public content and clear release windows | Generation, deployment, link integrity and update delay |
| Client-rendered SPA | Signed-in workspaces and complex ongoing interactions | Scripts, APIs, direct URL refreshes and public content output |
A website can combine approaches by page task. List the required freshness, public access and maintenance owner for product details, news, enquiries and signed-in tools before choosing an implementation. An SSG website still needs a service to receive form submissions. An SSR website may use JavaScript for interactive controls. Specify page generation, interaction and business processing as separate delivery requirements.
2. How CatchCMS currently serves public pages
CatchCMS uses Blade server-side rendering for its public website. Themes arrange CMS content into pages; the server outputs body content, titles, navigation, breadcrumbs and applicable structured data. Product, feature, case-study, article and FAQ pages can use this approach. Check each implemented theme and page against the actual acceptance requirements.
The administration interface uses a Vue SPA for editing and management. Customer-facing pages use a separate Blade runtime. Confirm these two responsibilities during selection. Static-site export, a separate SPA frontend, a prerendering service or automatic CDN invalidation needs an explicit project scope for development, deployment and verification.
Connect an existing AI-built design to content models through the AI website and theme delivery workflow. Use the product catalog guide to agree how themes retrieve and display model numbers, specifications and files.
3. Check the initial response, then the browser presentation
- Select real samples. Prepare a product detail page, a news item, a category listing and their English counterparts. Record the model number, important specifications, title and destination links that should appear.
- Inspect the original HTML. Use View Page Source or the main document request response in browser developer tools. Find the selected text and links. The Elements panel shows the DOM after scripts have run; record it separately from the initial response.
- Check links. Product-to-document and article-to-enquiry links should have accessible destinations. Inspect actual a elements and href attributes. Record the loading conditions for important content that appears after scrolling, clicking or an API response.
- Use a diagnostic comparison. Temporarily disable JavaScript and reopen the sample to check core text, image descriptions and basic links. Restore scripts and check menus, filters and other controls. These observations assess content output and interaction implementation separately.
Google's JavaScript SEO documentation describes crawling, rendering and indexing. Checking the initial response helps identify when content becomes available. Verify a search platform's actual rendered result with its inspection tools.
4. Review status codes, SEO and both languages together
Paste a deep URL into a new window, open it directly and refresh it. Check that a normal public page returns HTTP 200 with the expected content. A confirmed nonexistent URL should return an actual 404. For an approved URL migration, check the specified permanent redirect, such as 301, and its final destination. Record the visible message and the network response status separately.
- Check each title, description and main heading against the product or question being answered. Match important body facts to the current CMS record.
- Check that canonical identifies the accepted canonical address. Verify accurate reciprocal language links between equivalent Chinese and English pages, with each language retaining its own body and canonical address.
- Check that structured data parses and that page identity, language and facts match visible content. Review robots directives, response headers, the sitemap and internal entry links.
- Repeat the original-response check for English pages, including model numbers, units, document languages and versions. Use the multilingual update checklist after changes.
The website SEO feature guide explains the relevant fields. After page-output acceptance, use the search discovery and indexing guide to record platform inspection results.
5. Verify mobile tasks and enquiry outcomes
Read long headings, specification tables and English paragraphs at a phone viewport width. Check horizontal scrolling, obscured text, menus, document links and form labels. Ask an actual content maintainer to find a product from a listing, read its specifications, open its document and reach the enquiry form. Observe whether each destination is clear.
The CatchCMS form endpoint returns JSON, and theme scripts present submission feedback. Run the complete enquiry test with scripts enabled. Have the responsible person approve an enquiry test, using clearly labelled test information and agreed test recipients. Record submission time, product and language, page feedback, the new backend record and, where email notification is enabled, notification status and actual receipt. Page loading, submission feedback, record preservation and notification delivery each have their own evidence. Follow the enquiry form and email troubleshooting checklist for mail diagnosis. Handle test data according to the agreed project process after acceptance.
6. Hand over content updates, caching and performance checks
For SSR, specify which pages must change after an editor saves content, and who maintains application, full-page and CDN caches. For SSG, specify the generation trigger, deployment owner, failure detection and withdrawal of old files. In either approach, check details, listings, related links and English pages. Cache keys must distinguish the actual site, language and address.
Use cache policies appropriate to the access scope of sessions, personal information and form security fields. Before enabling shared caching, have technical staff verify isolation and freshness. Write executable steps for clearing the relevant cache, confirming the new public content and restoring the previous content.
Compare performance using the same page samples and similar devices and networks. Record date, location, tool, cache conditions and sample size. Lab tests help diagnose problems; field data describes actual user experience. Keep LCP, INP and CLS results identified by source. Images, scripts, server response and caching all contribute. Consult web.dev's rendering overview and Web Vitals guidance, then prioritize changes using measured results.
7. Example: accept a bilingual PX-100 product update
This is a fictional acceptance exercise. A company maintains 30 products and updates Chinese and English specifications weekly. A marketing editor owns publishing and technical staff own the runtime environment. The team chooses SSR for public detail pages and listings, then includes public verification in its editing workflow.
The exercise changes the PX-100 document version from R2 to R3. First preserve the previous text, file addresses and version records in both languages. Complete the reviewed body and document-link changes. Open the Chinese and English product URLs directly and confirm the model number and R3 in the initial response. Check the listing, document download destination, canonical and reciprocal language links, then complete mobile reading and an approved test enquiry.
The delivery owner records observations and outstanding fixes in the acceptance record. For the same task under SSG, add generation and deployment receipts before checking R3 at the public URL. For a client-rendered SPA, add script loading, API response and deep-URL refresh checks. This example illustrates a procedure; project measurements establish actual speed, indexing and enquiry outcomes.
8. A reusable delivery record and consultation brief
- Page: site, language, public URL, page task, expected core content and change time.
- Output: initial text and links, status and redirects, title, description, canonical, reciprocal language links and structured data.
- Tasks: mobile reading, menus and downloads, test enquiry marker, backend record, notification and inbox evidence.
- Operations: publishing method, cache-update steps, owner and restoration steps; performance samples labelled as lab or field data.
- Decision: accepted items, outstanding fixes, evidence location, recheck date and owner. Record search indexing, organic traffic, enquiries and actual AI citations as separate measured outcomes.
Stable text, internal links and structured information help search and AI systems understand pages. Google's guidance for AI search features also starts with crawlability, indexing eligibility and content quality. Verify actual appearances and citations from observed results. See AI data operations and GEO for ongoing content work.
Prepare site and language counts, sample URLs, update frequency, interaction requirements, the deployment environment and the maintenance owner. Review source-code plans and updates, then use project enquiries to agree the theme, development and delivery scope.