{{text}}
Share
Link obfuscation means replacing a standard HTML link with an element the browser makes clickable in JavaScript, but that a crawler does not recognise as a link. The visitor clicks and lands on the page. The crawler sees no destination to follow. The link leaves the graph without leaving the screen.
The technique has circulated in French SEO circles for years, almost always presented as a way to spread internal PageRank more sensibly. That is not wrong. It is mostly, for the vast majority of sites, an answer to a problem they do not have. Synqro is a Webflow Premium Partner Webflow agency in Paris, and we deploy obfuscation on a minority of projects, always for the same reason: a URL count that got away from its owner.
This article covers what the technique actually does, what separates it from nofollow, what Google writes in its own documentation, and the threshold above which it is worth the work it costs.
What obfuscation does, and what it does not
Google is explicit about what it can follow. Its documentation on crawlable links, updated on 10 December 2025, fits in one sentence: Google can only crawl a link if it is an HTML a element with an href attribute. The same page lists what its crawlers do not reliably follow: an a without an href, a span carrying an href, a link driven by an onclick handler, a javascript: URL.
Obfuscation applies that rule as written. It deceives nothing, it simply does not write a link. That is what separates it from every other crawl control method, which leave the link in place and add an instruction on top.
| Method | Link exists in the HTML | Google follows it | Target page indexable | What it solves |
|---|---|---|---|---|
| Standard link | Yes | Yes | Yes | Nothing to solve |
| nofollow attribute | Yes | Treated as a hint | Yes | Qualifying a paid or unvetted link |
| noindex tag | Yes | Yes, it must crawl to read the instruction | No | Removing a page from the index |
| robots.txt block | Yes | No, the target is not crawled | Possible without being read | Saving crawl on a directory |
| Obfuscation | No | No, it never sees it | Yes, if reachable another way | Removing a link from the internal graph |
One consequence is almost always left unsaid. An obfuscated link is not only invisible to Google, it is invisible to anything that reads HTML without running JavaScript. Keyboard navigation skips the element unless you rebuild it by hand, screen readers do not announce it as a link, and opening in a new tab stops working. That is a real accessibility cost, not a formality.
Why this is not a nofollow
The confusion is constant and it dates back to 2009. Matt Cutts announced then, as Search Engine Land reported on 16 June 2009, that Google had changed its algorithm more than a year earlier: nofollow links are no longer excluded from the distribution calculation. On a page carrying ten links, five of them nofollowed, value does not concentrate on the remaining five. It is still divided by ten, and the share allocated to the nofollowed links is passed to nobody.
So nofollow does not redirect value, it destroys it. Obfuscation removes the link before the count is taken: the page carries five, the division is by five. That is the only difference that matters, and it is also what makes the technique serious, since it genuinely changes the structure of the site as Google sees it.
The downside arrives with it. An obfuscation mistake removes a page from crawling without leaving any readable trace in the HTML, so without any alert. It is the kind of failure that takes three months to diagnose.
Does Google treat this as cloaking?
No, as long as the page served is the same for everyone. Google's spam policies, updated on 28 August 2026, define cloaking as presenting different content to users and search engines with the intent to manipulate search rankings and mislead users. The two examples given are a travel destinations page shown to search engines while users see a discount drugs page, and inserting keywords only when the user agent requesting the page is a search engine.
Obfuscation does neither. The content served is identical, only navigation changes. The nuance rests on a single condition: if you obfuscate by detecting the user agent, serving one HTML document to crawlers and another to humans, you have left obfuscation for cloaking. The test is one line long, the HTML must be the same whoever the visitor is.
One point deserves honesty. Google has never published a page that explicitly authorises the practice. Articles claiming that Google accepts it are extrapolating. What is verifiable is that the technique meets none of the criteria in the cloaking definition, and that it appears in no list of sanctioned practices, unlike the methods catalogued in our article on black hat SEO.
The three cases where obfuscation earns its keep
Catalogue facets and filters
An e-commerce catalogue with crossed filters generates one URL per combination. Three filters with ten values each already produce a thousand pages with no reason to exist. Canonical tags and noindex handle indexing, they do not reduce the number of links the crawler discovers on arrival. Obfuscation does.
Navigation blocks repeated at scale
An 80-link menu present on 50,000 pages adds up to four million internal links, nearly all of them pointing at a handful of destinations. Obfuscating the part of the menu that serves navigation rather than search changes the map of the site measurably. On a 200-page site, the same move changes nothing at all.
Pages useful to visitors, pointless in search
Internal comparison tools, login areas, booking funnels, multi-step forms. These pages must stay reachable and have no business weighing on the link graph. A noindex tag removes them from the index, but Google keeps crawling them to read the instruction. Obfuscating the links that lead there solves the problem further upstream.
The cases where it does nothing, which is almost all of them
The most common justification is crawl budget. Yet Google puts numbers on the scope itself. Its documentation on managing crawl budget, updated on 22 July 2026, targets sites with more than one million unique pages whose content changes moderately often, and sites with more than ten thousand unique pages whose content changes very rapidly. Google notes that these figures are a rough estimate and not exact thresholds, then adds a sentence that reads like an instruction: if your site does not have a large number of pages that change rapidly, or if your pages seem to be crawled the same day they are published, you do not need to read that guide.
A brochure site, a 200-article blog, an agency site, a SaaS with a few hundred URLs therefore gain nothing from obfuscating anything. The time spent writing the script, maintaining it and repairing the accessibility it breaks would be better spent fixing internal linking properly, deleting empty pages, cleaning up redirect chains.
Our position is blunt. Obfuscation is a bottom-of-the-list tool. It comes after architecture, after pruning useless pages, after fixing redirects. A site that deploys it before doing that work is moving a symptom instead of treating the cause.
How it is implemented
The principle rests on three pieces. A neutral element carries the visible anchor, a data attribute carries the destination in a form the crawler will not read as a URL, an event listener restores navigation on click.
In markup, the link <a href="/target-page">Anchor</a> becomes a <span class="obf" data-o="L3RhcmdldC1wYWdl" role="link" tabindex="0">Anchor</span>, where the value of data-o is the destination encoded in base64. A script listens for clicks and the Enter key on elements of class obf, decodes the value with atob and triggers navigation.
Base64 encoding is not security, it is a precaution so a plain URL is not recognised as one. Two things not to skip: without role="link", without tabindex="0" and without Enter key handling, you make keyboard navigation impossible. Handle middle-click too if your visitors open tabs.
On Webflow, in practice
Webflow offers no obfuscation option. Implementation goes through custom code, placed before the closing body tag at site or page level, and through elements you give a class and custom attributes in the Designer. A link coming from a CMS link field cannot be obfuscated from that field, it has to be rebuilt as a generic element.
In practice the question rarely comes up on Webflow. Sites reaching the URL volume where obfuscation becomes relevant are rarely Webflow sites, and when they are, the useful work sits upstream: cutting the number of URLs generated, ranking collections properly, checking what the crawler actually reaches. That is exactly what a technical SEO audit covers.
How to check that obfuscation works
Three checks are enough, in this order.
- The URL Inspection tool in Search Console, rendered HTML tab. The link must not appear as an
atag in either the source or the rendered code. - A crawler run twice, without and then with JavaScript execution. The target page must drop out of the internally discovered URL list on the first pass.
- A manual test with the keyboard, tab then Enter, plus a screen reader pass. This is the check people skip, and it is also the one that produces the most expensive regressions.
The reflex that matters more than the technique
For Les Fermes de Marie, a luxury hotel group in Megève, our work consisted of an SEO audit and more than fifty technical fixes. Obfuscation was not among them, and that is precisely the point: at that scale the gains were elsewhere, in structure and code hygiene. The detail is in the Les Fermes de Marie case study.
If you are hesitating between obfuscating and rebuilding your architecture, the answer is in your crawl logs, not in an article. Our Webflow SEO agency team knows how to read those files. Let's talk about your project.
Frequently asked questions about link obfuscation
Is link obfuscation penalised by Google?
No Google spam policy targets obfuscation. The cloaking definition, updated on 28 August 2026, requires serving different content to crawlers and users with intent to manipulate, which obfuscation does not do since the HTML served is identical for everyone. That said, Google has never published anything that explicitly authorises it.
What is the difference between obfuscation and nofollow?
Nofollow leaves the link in the HTML and asks Google not to follow it, but the share of value allocated to it is lost rather than redistributed. Obfuscation removes the link from the HTML, and therefore from the count: value spreads across the remaining links.
Does obfuscation improve rankings?
Not directly. It changes how internal links are distributed and what the crawler reaches. A gain only appears if the problem being treated really was a URL volume or link dilution problem. On a normal-sized site the effect is nil.
How many pages before it is worth considering?
Google places the crawl budget question above one million unique pages changing moderately often, or ten thousand unique pages changing very rapidly, while noting these are orders of magnitude rather than thresholds. Below that, the documentation says plainly there is no need to worry about it.
Does obfuscation create accessibility problems?
Yes, when done badly, which is the common case. An element that is not a link is not reachable by keyboard and is not announced as a link. You have to add a role, a tab order and Enter key handling, then verify with a screen reader.
Can you obfuscate a link on Webflow?
Yes, with custom code and attributes added in the Designer. A link generated from a CMS link field has to be rebuilt as a generic element, the option does not exist in the interface.
How can I tell whether a site uses obfuscation?
Compare the source code with the rendered HTML, then run a crawl without JavaScript execution. If elements that are clickable on screen appear nowhere as an a tag with an href, the site is obfuscating.
Should footer links be obfuscated?
Rarely. A footer of ten to twenty links on a normal-sized site causes no problem at all. The question only arises across tens of thousands of pages, and only for links with no search value, such as legal notices or account pages.
Synqro's latest achievement
