usdigitalmarketingindex.com SEO
Which SEO elements should be smoke-tested when launching a new page template?
Updated November 15, 2026
The ones that, if broken, affect every page the template produces: whether the main content and links are in the HTML a crawler receives, whether the title and headings are generated correctly, whether the canonical tag and robots directives say what is intended, whether internal links are real and working, whether structured data is valid, and whether the page returns the right status code. A smoke test is a short pass over these on a handful of representative pages, done before launch, with a recorded result for each. It does not prove the template is perfect, and it does not review the content of any individual page. It establishes that the template has no defect that would be multiplied across the site on day one.
Rendered content and metadata
Test two views of each sample page: the HTML as served, and the HTML after rendering.
| Element | Check | Pass condition |
|---|---|---|
| HTTP status | Request the URL | 200 for a live page; real 404 for a missing one |
| Main content | Fetch without JavaScript; search for a sentence from the body | Present in served HTML, or reliably present after rendering |
| Title element | Read it | Present, unique to the page, generated from the right field, not a placeholder or the template name |
| H1 | Read it | One main heading, matching the page's subject |
| Heading order | Read the outline | Headings reflect the content structure |
| Meta description field | Read it | Populated from the intended field, or intentionally empty; not a default string repeated on every page |
| Language attribute | Read the html element | Correct for the page |
| Images | Inspect img elements | <!–INLINECODE0–> present in HTML; alt attributes populated from content |
| Structured data | Run the Rich Results Test | Valid; values match visible content |
| Social tags | Read Open Graph tags | Populated; correct URL |
| Rendering | URL Inspection on a staging URL where accessible, or a rendering crawler | Content and links present in rendered HTML |
| Mobile | Load at a phone viewport | Same content as desktop |
| Performance | Lab test of the template | No gross regression against the template it replaces |
Placeholders are a template fault to look for: a title of “Page title here”, an empty H1 where a field was not mapped, a description identical on every page. Test with pages that have unusual data, such as a very long title, a missing image, or an empty optional field.
Internal link functionality
| Element | Check | Pass condition |
|---|---|---|
| Navigation | Fetch without JavaScript; list anchors | Menu links present as <!–INLINECODE1–> elements with <!–INLINECODE2–> |
| Breadcrumbs | Inspect | Present; links correct; markup matches |
| In-content links | Click through a sample | Resolve with 200, no redirects |
| Related or listing modules | Inspect | Real links to canonical URLs |
| Pagination | Inspect on listing templates | Anchor links to following pages |
| Footer | Inspect | Present and unchanged |
| Link format | Read <!–INLINECODE3–> values | Absolute or root-relative; correct host and protocol; no staging host |
| Inbound links | Check that existing pages link to pages using the new template | The template's pages are reachable |
| Orphan check | Crawl staging from the home page | Sample pages found |
Google’s guidance on links says its crawlers follow links that are a elements with an href attribute (Google’s guidance on crawlable links and anchor text). A template that builds navigation from click handlers passes a visual review and fails this test.
Index directives and canonical
These are the elements whose failure is silent and site-wide.
| Element | Check | Pass condition |
|---|---|---|
| Meta robots | Read the tag in served and rendered HTML | No <!–INLINECODE6–> on pages meant to be indexed; no leftover from staging |
| X-Robots-Tag header | Read response headers | No <!–INLINECODE7–> |
| Canonical tag | Read it | One tag; absolute URL; the page's own production URL; not the staging host |
| Canonical consistency | Compare served and rendered | Same value |
| robots.txt | Test the template's URL pattern against the production file | Allowed |
| hreflang | Read, where used | Correct codes; self-reference; production URLs |
| Pagination canonicals | Read on page 2 and beyond | Self-referencing |
| Parameters | Load with a tracking parameter | Canonical names the clean URL |
| Sitemap | Confirm the template's URLs will be included | Present after launch |
| Status of variants | Request with and without trailing slash, uppercase | One canonical form; others redirect |
Check directives on production immediately after launch as well as on staging beforehand. Staging environments are deliberately blocked from indexing, and the launch is the moment that block has to be absent.
Representative page sampling
Choose samples that exercise the template’s branches.
| Sample | Why |
|---|---|
| A normal page | The common case |
| A page with the minimum content | Empty optional fields |
| A page with the maximum content | Long titles, many images, long body |
| A page with special characters in the title | Encoding and escaping |
| A page in each language, where applicable | Localization paths |
| A page in each state the template supports | For example in stock and out of stock |
| The first and a later page of a paginated listing | Pagination logic |
| A page that should not be indexed, if the template has such a state | Directive logic in the other direction |
| A missing page | Error handling |
An executable smoke-test checklist with evidence fields:
| # | Test | URL tested | Method | Expected | Observed | Pass or fail | Evidence |
|---|---|---|---|---|---|---|---|
| 1 | Status 200 | [staging URL] | HTTP request | 200 | Response headers saved | ||
| 2 | Body text in served HTML | Fetch, search for sentence | Found | Source excerpt | |||
| 3 | Title correct | Read title | Page-specific | Value recorded | |||
| 4 | Single H1 | Count H1 | 1 | Value recorded | |||
| 5 | No <!–INLINECODE8–> | Read meta and header | Absent | Values recorded | |||
| 6 | Canonical self, production host | Read tag | Own URL | Value recorded | |||
| 7 | Navigation links in HTML | Fetch without JS | Present | Anchor count | |||
| 8 | Internal links resolve | Request each | 200 | Status list | |||
| 9 | Structured data valid | Rich Results Test | No errors | Test result saved | |||
| 10 | Rendered content | Rendering tool | Matches served | Screenshot and HTML | |||
| 11 | Mobile parity | Phone viewport | Same content | Screenshot | |||
| 12 | Missing page returns 404 | [invalid URL] | HTTP request | 404 | Response headers | ||
| 13 | robots.txt allows pattern | Robots test | Allowed | Result | |||
| 14 | Images have <!–INLINECODE9–> and alt | Inspect | Present | Values recorded |
Run it per sample page. Record who ran it, when, on which build, and attach the evidence. A row marked pass with no observed value and no evidence is not a result.
Decision rule:
| Outcome | Action |
|---|---|
| Any failure in status, directives, canonical, or content presence | Launch blocked |
| Failures in structured data or social tags | Fix before launch where possible; otherwise logged with an owner and date |
| All pass on staging | Proceed, then repeat the directive and status tests on production |
What the smoke test does not cover: the accuracy and quality of each page’s content. A template can pass every row and still publish an article with a wrong fact or a thin page. Page-level editorial review is a separate process, and a passing template checklist is no substitute for it.
After launch, repeat rows 1, 5, 6, and 13 on production URLs within the hour, and schedule a crawl of the template’s pages for the following day.
A template test covers a whole class of pages. Verifying that one specific defect has been fixed calls for a different kind of evidence, described in the guide to what proof a developer should provide before an individual technical SEO bug fix is accepted.
The rule: before a template goes live, test status, content in the HTML, title and heading, directives, canonical, links, and structured data on pages that exercise every branch, and record what was observed.
Add your comment to this answer