A PDF can rank, but only if Google can read its text, find it through links, and tell which copy is the real one. Per Google's Martin Splitt and John Mueller (13 October 2026), Google converts PDFs to HTML for indexing, so the checklist below is mostly about making that conversion clean: real text, a clear title, one canonical version, an HTML landing page, and a new filename when content changes. Use a PDF for print and download needs, and put the answer people search for on a web page.
Key Takeaways
- Run the copy-and-paste test first: if you cannot select text in the PDF, Google may not read it.
- Give every PDF an HTML landing page that carries the key facts, and link to the PDF from it.
- Use one canonical copy. For PDFs, Google documents a
rel="canonical"HTTP header (a Link header). - Do not block PDFs in robots.txt if you want them indexed, and do not use robots.txt to hide them either; use
X-Robots-Tag: noindex. - Publish updates under a new filename, because Google recrawls PDFs infrequently, per Mueller.
- Put fees, dates and eligibility on the HTML page, since PDFs go stale silently.
Why this checklist exists
My previous post on how Google indexes PDFs covers what Google said on Search Off the Record. Here I turn it into steps. The sources I used: Google's indexable file types, duplicate URL consolidation, robots meta tag and Googlebot documentation, the Search Engine Journal and PPC Land summaries of the podcast, and two older practitioner posts from Foxit (2022) and Ryte. The older posts are secondary and dated, so I mark anything that relies only on them.
Part 1: Decide whether it should be a PDF
1. Ask what the visitor needs
Ryte frames it as: can a PDF meet the expectations of a search visitor? If the answer is "read, compare, filter or buy", use HTML. If it is "print, sign or keep offline", a PDF is reasonable. Splitt and Mueller said the same: PDFs for fixed formatting, forms and print; HTML is more flexible.
2. Plan the HTML landing page first
Every PDF should have a web page that states what it is, who it is for, the key facts in text, and a download button. This is the page you optimize for the query; the PDF is the attachment.
3. Do not publish a PDF as your only source of truth for volatile facts
Fees, dates, batch start dates and eligibility belong on the HTML page. A PDF refetched rarely can sit in results with last year's numbers.
Part 2: Build the file correctly
4. Export with a real text layer
Export from the source document (Word, InDesign, Google Docs), not from a screenshot or print-to-image. Per Splitt and Mueller, rasterized pages with no text layer may not rank for their text.
5. Run the copy-and-paste test
Open the PDF, select all, paste into a plain-text editor. Check the order, special characters and tables. Foxit and Ryte both give the same test. If OCR is the only way to read it, Google's handling is less certain.
6. Write the visible title and headings properly
Use a descriptive title at the top of the first page and real headings. Google's title link documentation lists HTML sources for titles and does not mention PDF metadata, so I cannot confirm how Google picks a PDF title. Foxit's 2022 post says Google uses the title metadata in the file and anchor text of links; treat that as a secondary claim and make both consistent anyway.
7. Fill in the document properties
Set the Title, Author and Subject fields in the file. Mueller recommended populated metadata as one help against duplicate confusion, per the Search Engine Roundtable summary. Costs nothing, so do it.
8. Put a visible link to your site at the top
Mueller and Splitt suggested a visible homepage link at the top of the PDF, which survives conversion, so that copies that get re-hosted still point home. Use a full URL, not just a logo.
9. Add internal links inside the PDF
Links in PDFs are crawled because the converted HTML contains them. Link to the relevant landing page, the apply or enquiry page and a related guide. Foxit notes you cannot embed nofollow in a PDF, so only link to pages you are happy to endorse.
10. Keep the file small
Google's Googlebot documentation says it crawls the first 64MB of a PDF, but the hosts on the podcast could not name the limit and one trade summary quoted a different number. Aim for a few MB, since mobile users must download the whole file before it shows, per PPC Land.
11. Make it accessible
Tagged structure, reading order and alt text help screen-reader users, and rasterized pages are harder for them. I did not find a Google statement tying accessibility tags to ranking, so do this for users, not for rankings.
Part 3: Publish it correctly
12. Choose a stable, descriptive filename
Use words in the URL: data-analytics-course-brochure-2026-q4.pdf instead of Final_v7_new.pdf. I have no Google source that filenames rank directly; the benefit is clarity and a clean new URL for each version.
13. Serve the right content type
Google determines type from the Content-Type header, per its indexable file types page, so send application/pdf.
14. Link to it from crawlable HTML
A PDF no one links to is hard to discover. Link from the landing page and include it in your XML sitemap if you want it indexed. Foxit suggests listing the preferred URL in the sitemap when HTML and PDF versions both exist.
15. Do not block it in robots.txt unless you mean it
Ryte notes blocked PDFs can still be indexed and lose their link value. If you want one out of results, serve X-Robots-Tag: noindex; Google's robots meta tag page gives Apache and NGINX examples and warns a robots.txt block means the header is never seen.
Part 4: Handle duplicates and updates
16. One canonical copy
If the same content exists as HTML and PDF, Google's duplicate-URL documentation describes a rel="canonical" HTTP header for non-HTML files. Its example is a Link header on the file's response pointing to the canonical URL, with a PDF as the canonical on a Word version. In your case the pattern is the reverse: the PDF's response points to the HTML page. Test with a header checker after deployment.
Link: <https://www.example.com/courses/data-analytics>; rel="canonical"17. Decide: index the PDF, or fold it into the page
- If the PDF is a genuine standalone asset with its own demand (a map, a rate card, a whitepaper others cite), let it index with HTTP 200 and no canonical elsewhere.
- If it is a print version of a page, canonical it to the page. Ryte says to prefer canonical over noindex when the PDF has backlinks, since noindex wastes link power. That is a practitioner view, so treat it as a reasoned default, not a Google rule.
18. Publish updates under a new filename
Mueller said Google refetches PDFs rarely and that a new filename is the way to get the update crawled. Keep the old URL live with a link or redirect to the new one if you want users to land on the current version, and update the landing page link.
19. Watch CMS-generated PDFs
Splitt and Mueller warned that CMS-generated PDF copies of pages can sit out of sync with the HTML. Canonicalize or noindex them, and remember that Ryte says crawlers still spend time on them.
20. Measure them
Standard page analytics may not record PDF views, per Ryte. Use server logs, GA4 file-download events or UTM-tagged links on the landing page. Search site:yourdomain.com filetype:pdf to see what is indexed.
Templates by document type
Whitepaper
- Landing page: summary of findings in text, author bio, publish date, a short table of contents, download button.
- PDF: full report, canonical to the landing page unless it is a citable asset on its own. If others cite the PDF, let it index.
- Update rule: new filename and "last updated" date on the landing page.
Product brochure
- Landing page: features and specs in HTML, pricing statement, enquiry form.
- PDF: designed for print and sales calls; canonical to the product page.
- Watch: duplicated claims that no longer match the website.
Course prospectus (edtech, India)
- Landing page: programme overview, eligibility, duration, fee structure, admission steps, FAQs. This is the page that answers "fees and eligibility" questions, as I argued in my programme page structure post.
- PDF: official prospectus or brochure, linked from the page with a dated file label.
- Watch: fees, batch dates and scholarship terms. Put those in HTML and attribute them to the official source with a date, because PDF copies circulate on aggregator sites. See also my edtech site audit after the spam update.
- Compliance: if the prospectus makes outcome or placement claims, keep the same wording in HTML and in the PDF so the claim cannot differ between copies.
A one-page audit table
| Text layer OK? | HTML equivalent? | Canonical set? | Last updated | Action | |
|---|---|---|---|---|---|
| (name) | Y/N | URL | URL or none | date | keep / replace / noindex / fold |
FAQ
How do I make a PDF SEO-friendly?
Export it with a real text layer, give it a descriptive title and filename, fill in the document properties, link to your site from inside it, and link to it from a crawlable HTML page. Then decide whether it should be indexed or canonicalized to an HTML page.
Should I use noindex or canonical for duplicate PDFs?
Google documents a rel=canonical HTTP header for non-HTML files. A canonical keeps link value flowing to the HTML page, which a practitioner source from Ryte prefers over noindex for PDFs with backlinks. Use noindex via X-Robots-Tag when you simply want it out of results.
Can I put noindex inside a PDF?
Not as a meta tag. Google's documentation says to use the X-Robots-Tag HTTP header for PDFs, set on the server. Make sure robots.txt does not block the file, or Google will never see the header.
Does Google read scanned PDFs?
Not reliably. Splitt and Mueller said image-only PDFs may not rank for their text, and older practitioner posts say OCR is partial. Export with real text or provide an HTML version.
How do I update a PDF that is already indexed?
Publish the new version under a new filename and update the links on your landing page. Mueller said Google refetches PDFs rarely and a new linked URL is crawled as new content.
Do PDF file names matter for SEO?
I found no Google statement that filenames rank directly. They still help humans and give each version a clean URL, so use descriptive words and a date or version.
Are links inside a PDF followed?
Google's hosts said links work because they survive conversion to HTML. Foxit notes nofollow cannot be embedded in a PDF, so link only to pages you endorse.
How big can a PDF be?
Google's Googlebot documentation lists 64MB for PDFs and 2MB for HTML as of the page I read, but the podcast hosts could not name the PDF number. For mobile users, smaller is better.
Should a course prospectus be a PDF or a web page?
Both: a web page with the fee, eligibility and admission facts in text, and the PDF as a download. That keeps answers readable, updatable and consistent across copies.
Need a PDF and landing-page audit?
I'm Younus Fardeen, an organic growth, SEO and AEO strategist with 4+ years of marketing experience, working with edtech and startup teams. If your key information sits in brochures and prospectuses, I can help you audit them and build the HTML pages that answer the questions people actually search. See my work and reach me via the contact form at younusfardeen.in. This checklist is as of 14 October 2026 and reflects the sources linked above.