Webflow SEO is the set-up work the platform leaves to whoever runs the site. The sitemap, SSL certificate, 301 redirects and fields for titles and meta descriptions are ready out of the box, while canonical tags, the language declared in the code, structured data and rules for crawlers have to be switched on by hand in the settings. In our test of 75 pages on 15 Italian sites built with Webflow, run on 9 October 2026, the meta description was there 83% of the time, but 4 sites out of 15 had no canonical tag at all and 8 had no structured data on any of the pages we checked.
In short, on Webflow the copy is fine. It's the settings that get left empty.
And the people using it are rarely hobbyists: in our sample, Webflow is the CMS of startups, software companies and design agencies.
It runs 0.8% of all websites and 1.2% of those with an identifiable CMS, according to W3Techs (October 2026). In the Chrome data collected by HTTP Archive for August 2026, there are 4,200 Webflow sites visited from the UK on mobile, against 125,625 on WordPress: roughly one for every 30.
So where do Webflow users go for instructions?
Mostly to Webflow itself. When we searched "webflow seo" on Google UK from a mobile on 10 October 2026, the first page showed an AI Overview, four pages from webflow.com, a Reddit thread, then Seobility, Flux Academy, Webstacks and three videos. Most of what ranks is written by Webflow or by people who sell Webflow sites, and none of it measures what real sites actually get wrong.
So the same mistakes keep turning up:
- the global canonical never set, because "Webflow does it all for you";
- the lang attribute missing or set to the wrong language, such as English on sites written in Italian;
- the sitemap switched off, sometimes declared in robots.txt but nowhere to be found;
- no structured data, because Webflow has no plugin to add it.
We've been there ourselves. In 2024 we started rebuilding the Visilay site on Webflow, with languages and CMS already organised, and stopped before launch: given our skills and how we manage things in-house, we chose to stay on WordPress. Before we could downgrade the site plan we had to switch off the multilingual feature first, a detail no guide mentions.
The yardstick we use to judge a site, whatever CMS it runs on, comes from the projects we work on. On Macropix, a Milan-based maker of LED screens, every product line has its own page, and for "monitor pubblicitario" (advertising display) the site went from position 88 to 2 on Google Italy (the full case is in our guide to SEO for manufacturing).
On Sunscape, an Italian solar farm developer, we built an architecture of pages by region: in six months organic impressions rose from 329 to 5,809 a month.
What does that have to do with Webflow?
Both results sit on a clean technical base: one address for every page, the right language, a sitemap Google can read. Those are exactly the settings we found switched off in the test.
In this article:
- what Webflow does for you and what it doesn't, with links to the official documentation;
- the results of our test on 75 pages, check by check;
- how to fix canonicals, language, sitemap and redirects from the dashboard;
- how to add structured data without a plugin;
- when Webflow isn't enough and another CMS makes more sense.
Let's start.
Is Webflow good for SEO? (Yes, if you open the right settings)
Yes: Webflow produces clean code, generates the sitemap, includes SSL and lets you write titles, meta descriptions, slugs and alt text without plugins. The catch is that several SEO features exist but don't switch themselves on.
The clearest case is the canonical. Webflow's official guide to canonical tags, updated in May 2026, says the global tag is set in Site settings, under SEO, by entering the site's base address including the protocol. If it's missing or set up wrongly, the same page warns, search engines may pick the preferred URL themselves.
Here's how the work is split:
| Element | What Webflow does | What you have to do |
|---|---|---|
| XML sitemap | Generates it automatically if the option is on | Check it's switched on and submit it to Google Search Console |
| SSL certificate | Included in Webflow hosting | Nothing, apart from checking the connected domain |
| Canonical tag | A global field and a field per static page, empty until you fill them in | Set the global URL including the protocol |
| Title and meta description | Fields on every page and in CMS templates | Write them, and in the CMS link them to dedicated fields |
| Language and hreflang | Automatic with the Localization add-on, in the code and the sitemap | Check the lang attribute on single-language sites |
| 301 redirects | A panel with wildcards and CSV import, on a paid plan | Map every address that changes |
| Structured data | None by default | Add it with custom code, CMS fields or Webflow AI (how to do it without duplicates) |
| Robots.txt and AI crawlers | Editable; search engines and AI bots are allowed by default | Decide which crawlers to block |
| .webflow.io subdomain | "Staging indexing" option to keep it out | Switch it off, on a paid plan |
The test: 15 Italian Webflow sites, 75 pages (and one error in five is in the settings)
The test can be repeated in an hour. On 9 October 2026 we checked more than 140 domains belonging to Italian companies and startups, kept only those serving pages generated by Webflow (recognisable by the data-wf-site attribute in the code) and, for each one, downloaded the home page and five pages taken from the sitemap. Three sites had no reachable sitemap, so we analysed the home page only. We ran it on the Italian market, where most of our clients are; the settings involved are the same for a UK site.
Total: 15 sites and 75 pages. They include B2B software companies, fintechs, mental health platforms, a legaltech, a publishing site, a design agency and a personal portfolio. The sitemaps ranged from 18 to 16,716 URLs.
| Check | Result | Share |
|---|---|---|
| XML sitemap reachable | 12 sites out of 15 | 80% |
| Robots.txt present | 13 sites out of 15 | 86.7% |
| No canonical tag | 16 pages out of 75, on 4 sites | 21.3% |
| Canonical pointing to another domain | 6 pages out of 75, on 1 site | 8% |
| Meta description missing or empty | 13 pages out of 75 | 17.3% |
| Empty title | 2 pages out of 75, on 1 site | 2.7% |
| Exactly one H1 | 48 pages out of 75 | 64% |
| No H1 | 12 pages out of 75 | 16% |
| Two or more H1s | 15 pages out of 75 | 20% |
| Lang attribute missing | 3 sites out of 15 | 20% |
| Lang set to English on Italian text | 2 sites out of 15 | 13.3% |
| Hreflang tags present | 9 sites out of 15 | 60% |
| Pages with JSON-LD structured data | 31 pages out of 75 | 41.3% |
| Sites with no structured data on any page | 8 sites out of 15 | 53.3% |
| Noindex pages inside the sitemap | 3 pages, on 2 sites | n/a |
| Images with empty alt | 2,691 out of 3,637 | 74% |
| Images with lazy loading | 3,287 out of 3,637 | 90.4% |
| AI crawlers named in robots.txt | 1 site out of 15 | 6.7% |
The sample is small and we'll say so straight away: 15 sites are not a national statistic, and almost all of them are looked after by teams with a dedicated designer.
That's exactly why the comparison with the test we ran the same day on Squarespace is telling. There, the meta description was missing on 81% of pages and only 30% had a single H1. On Webflow the description is missing on 17.3% and a single H1 reaches 64%.
The errors don't disappear: they move from the copy to the configuration, where only someone who opens the source code will see them.
Canonicals on Webflow: the field left empty (and the site pointing to another domain)
The canonical tag is the line of code that tells Google which address is the main one when a page exists in more than one version. On Webflow it isn't there until you set it: in the test 16 pages out of 75, on 4 sites, had none, including two sites with no canonical on any page.
And it isn't a theoretical problem.
On one site the same page returned a 200 status code both with and without the trailing slash, meaning two addresses with the same content. The canonical pointed to the version with the slash, the sitemap listed the one without: two opposite signals sent to Google about the same page.
The worst case was another one. A software company had the canonical on all six pages we checked set to a different subdomain, which sent visitors back to the main site with a 301 redirect (we checked this on the home page). Google is told "the right version is elsewhere", goes to look and gets sent back.
Does Google work it out anyway?
Sometimes. Google Search Central explains that the canonical is a strong signal but not an instruction, and that Google picks the canonical URL by weighing it against redirects, sitemaps and internal links. When the signals contradict each other, Google makes the choice, not you.
The fix takes five minutes:
- in Site settings, SEO tab, fill in "Global canonical tag URL" with the exact address of the domain, for example https://www.yoursite.co.uk;
- if you use Google Tag Manager or another tool to insert canonicals, leave the global field empty: the documentation warns that otherwise you end up with two tags per page;
- for static pages that need to point elsewhere, use the "Page canonical URL" field in the page's SEO settings;
- publish and check the source code of an inner page.
How the tag works in general is explained in our guide to the canonical URL.
Language and hreflang: when an Italian site says it's English
The lang attribute is the line at the top of the code that declares the language of the page. In the test it was missing on 3 sites out of 15, and another 2 declared lang="en" on pages written in Italian: a publishing site on every page, a software company on its legal pages and its Italian case studies. On a UK site the equivalent slip is a template left on a different language or region code.
Does it matter for rankings?
Not much: Google says it doesn't use the lang attribute to detect language, which it works out from the text. Screen readers and browser translation tools do read it, though: it's an error your visitors pay for, not your rankings.
Multilingual sites are trickier. 9 sites out of 15 had hreflang tags, the annotations that link the different language versions of the same page. With the Localization add-on, according to the guide to localised SEO updated in July 2026, Webflow sets the language of each version itself and adds hreflang both in the code and in the automatically generated sitemap.
So far so good. The problems start with content that's only half translated:
- on a site in five languages two articles had an empty title, and one of them, with an Italian slug, sat in the /nl/ folder with the code declaring it Dutch;
- the same guide warns that 301 redirects don't apply to localised slugs: every language version needs its own rule;
- if you switch off the automatic sitemap to use a custom one, you have to write the hreflang in the sitemap yourself, page by page.
The add-on is paid for separately. On the pricing page, checked on 10 October 2026, Localize Essential costs $9 a month for up to 3 locales and 10,000 translated words a month, and Localize Advanced $29 for up to 10 locales. Webflow lists prices in US dollars only, with tax added at checkout. How to set up hreflang properly is covered in our guide to hreflang tags.
Is your Webflow site sending Google mixed signals?
In the test 4 sites out of 15 had no canonical, one pointed to another domain and 5 declared the wrong language or none at all. You can't see these errors from the Designer. We check your site setting by setting and tell you what to fix first, starting from the searches that bring you customers.
Book a strategy consultationSitemap, robots.txt and redirects: 3 sites out of 15 without a sitemap
The XML sitemap is the list of addresses a site asks search engines to crawl. Webflow generates it itself, but only if the "Auto-generate sitemap" option is on: the sitemap guide, updated in August 2026, explains how to turn it on without saying whether it starts on.
In the test 3 sites out of 15 returned a 404 on /sitemap.xml. One of them even declared it in robots.txt: Google is sent to an address that doesn't exist.
Two other sites had no robots.txt at all. And on two sites the sitemap contained pages carrying a noindex tag, including the thank-you pages shown after a form: the site asks Google to crawl pages it is also telling Google not to index.
Then there's the staging subdomain, the one ending in .webflow.io. The indexing guide explains that you keep it out from Site settings, SEO, Indexing, by switching off "Staging indexing", and that the option needs a paid plan. Webflow then publishes a robots.txt just for the subdomain, which blocks crawling but doesn't remove pages that are already in the index: for those you need noindex and the removals tool in Search Console.
Redirects, finally, live in Site settings, Publishing, 301 redirects. According to the redirects guide updated in May 2026:
- there's no hard limit on the number of rules, but Webflow advises staying under 1,000 redirects, because each one ends up in a file the browser downloads;
- wildcards are written as
(.*)in the old address and%1in the new one, so/blog/(.*)to/articles/%1moves a whole section in one line; - rules can be imported and exported as CSV.
When and how to use them is in our guide to 301 redirects, and for sitemaps in general see our sitemap guide.
Structured data on Webflow: 8 sites out of 15 without a single line
Structured data is the code, usually in JSON-LD format, that describes to Google what's on the page: an article, a company, a product, a set of FAQs. Webflow doesn't add any on its own, and in the test only 31 pages out of 75 (41.3%) had any.
8 sites out of 15 had none on any of the pages we checked.
So can't it be done on Webflow?
It can, and the test shows it. The two mental health platforms had structured data on every page we checked, with types such as MedicalOrganization, FAQPage and BreadcrumbList, and the practitioner profiles on one of them reached 18 different types. The difference isn't the CMS, it's whoever configures it.
There are three ways to add it:
- custom code in the page settings, in the field for code inside the head tag;
- CMS fields inside the JSON-LD of the page template, so every article or product page fills the markup with its own data;
- the "Generate schema markup" button in Webflow AI, which according to the official guide updated in September 2026 needs a paid plan and uses AI credits. The same page asks you to review all generated markup before publishing it.
Which types you actually need and what they change in the results is explained in our guide to schema markup. Google reminds you that markup must describe content that's visible on the page.
H1s, alt text and speed: 74% of images have an empty alt
The H1 is the main heading on the page. In the test 48 pages out of 75 (64%) had exactly one, 12 had none and 15 had two or more. The extreme case was a home page with 21 H1s.
The likeliest reason is visual, as on Squarespace: the heading level is chosen element by element, and anyone who wants big text picks H1. The right logic is in our guide to heading tags.
On images the figure is starker: out of 3,637 images checked, 2,691 had an empty alt attribute, 74%. Only one had no attribute at all.
Do they all need fixing?
No. An empty alt tells screen readers the image is decorative, which is correct for icons, backgrounds and repeated logos. But product photos, charts and article images need describing, because Google Images uses the alt to understand what they show. Watch out for one limit: according to the official guide, Webflow's Audit panel doesn't check the alt of images inside the CMS, which are exactly the ones in your articles. How to write it well is in our guide to alt text.
Webflow handles deferred loading well: 90.4% of the images in the test had lazy loading. And on Core Web Vitals, the speed and stability metrics Google measures on real Chrome users, the HTTP Archive figures we pulled for August 2026 show that on mobile 78.9% of Webflow sites visited from the UK pass (3,216 out of 4,078), against 71.2% of Webflow sites worldwide and 65.8% of UK WordPress sites.
AI crawlers on Webflow: 1 site out of 15 blocks them
On Webflow you write robots.txt by hand in Site settings, SEO. Webflow's official blog, in October 2025, pointed to it as the way to block individual AI crawlers by name, from OpenAI's GPTBot to Google-Extended.
In the test 14 sites out of 15 don't name any AI crawler.
The only one that does is a legaltech, and it's an example worth copying: it blocks four crawlers used to train models (GPTBot, CCBot, ClaudeBot, Google-Extended) and explicitly allows OAI-SearchBot, ChatGPT-User and PerplexityBot, the ones that bring the site into answers that cite sources.
Should you do the same?
It depends what you sell. If you live off proprietary content, you have a reason to opt out of training. If you want to appear in ChatGPT or Perplexity answers, bear in mind that blocking every bot at once also removes your visibility in AI search. The difference between the various crawlers is explained in our guide to AI crawlers.
When Webflow isn't enough (and when it's the right choice)
Webflow is the right choice for company sites, startups and brands that change their pages often and want full control over design without touching a server. In the test the tidiest site was also the biggest, with 16,716 URLs in its sitemap: when the CMS template is set up properly, it repeats itself identically on every page.
It gets tight in three cases:
- very large sites: on the pricing page the Basic plan ($15 a month billed yearly, listed in US dollars only) doesn't include the CMS, and the plan with the CMS goes up to 20,000 items;
- advanced SEO needs that a plugin solves on WordPress, such as automatically generated structured data or canonical rules for filters;
- teams without a technical lead, because every setting we found switched off in the test needs someone who knows where to look for it.
The comparison between platforms is in our guide to the best CMS for SEO. For the closest alternatives, see SEO for WordPress, SEO for Wix and SEO on HubSpot.
If you decide to switch, onto Webflow or away from it, the real risk is the migration: every old URL has to be mapped with a 301, or the rankings you've earned are lost. The right sequence is in our guide to SEO for a website redesign.
Webflow SEO FAQs
Yes. It generates the sitemap, includes SSL, handles 301 redirects and lets you write titles, meta descriptions and alt text without plugins. Canonicals, structured data and crawler rules have to be set by hand, though: in our test of 15 Italian sites on 9 October 2026, 4 had no canonical at all and 8 had no structured data.
It depends. Webflow makes fewer technical mistakes out of the box: in August 2026, 78.9% of Webflow sites visited from the UK passed Core Web Vitals on mobile, against 65.8% of WordPress sites (HTTP Archive data). WordPress offers more automation through plugins, for example for structured data, and is easier for very large sites or sites run without a technical lead.
No. The global canonical is set in Site settings, on the SEO tab, under Global canonical tag URL, including the protocol. For individual static pages there is the Page canonical URL field in the page SEO settings.
No. The SEO fields are already in the settings of pages and CMS templates, and on paid plans the Audit panel with Webflow AI flags missing titles, meta descriptions, alt text and structured data. The panel does not check the alt of images inside the CMS, though.
Yes, with the Localization add-on, which is paid for separately: it sets the language of each version and adds hreflang in the code and in the automatic sitemap. 301 redirects do not apply to localised slugs, however, so every language version needs its own rule.
Yes. They are set in Site settings, Publishing, 301 redirects, on a paid plan. You can use wildcards with (.*) and %1 and import rules from CSV; Webflow advises staying under 1,000 redirects.
Webflow SEO in brief
Webflow generates the sitemap, SSL and redirects, and lets you fill in every SEO field without plugins.
There's no canonical until you set one: in the test it was missing on 4 sites out of 15.
The language in the code, the sitemap and robots.txt need checking by hand, one by one.
Structured data is down to you: 8 sites out of 15 had none.
Every AI crawler gets in, unless you decide who to let through.
What we took away from the test is something else. Webflow is almost always chosen by people who care about design, and people who care about design look at the page, not the source code. That's why the sites in the test were well looked after, with lazy loading on and meta descriptions written, but with canonicals, language and sitemap left as they were on handover day. On Webflow, SEO is decided on the day the designer hands the site over: if nobody opens Site settings that day, nobody ever does. It's the first thing we check in our SEO consultancy.
Until next time!
