A lot of content strategy rests on an assumption nobody re-checks: that when you publish on someone else's platform and link back to your own site, that link helps your site rank. It is the quiet justification behind half the "publish on Dev.to / Medium / Telegraph" advice out there. We publish on several of those platforms, we had been repeating that justification to ourselves, and we finally sat down to test whether it is true for the specific platforms we use.
This is a report on that measurement. It is not a tools review and it is not advice dressed up as data. It is what we did, what came back, and what we got wrong — including the parts that make our own past reasoning look sloppy.
What we wanted to find out, and why you should care
The narrow question: for the platforms we publish on, does a link from our post back to our site pass link authority — the thing SEO people mean when they say "a followed backlink" — or does it pass nothing?
Two mechanics decide that, and both live in the page's markup. First, the link itself: a rel="nofollow" (or sponsored, or ugc) attribute tells search engines not to pass authority through that link. Second, the page: a <meta name="robots"> tag can carry noindex or nofollow, which affects whether the page is treated as a real, ranking page at all. A followed link sitting on a page that search engines are told to ignore does not hand your site anything, no matter how clean the link looks in isolation.
Why this matters to you and not just to us: if any part of your reason for publishing somewhere is "the backlink helps our SEO," that is a factual claim about someone else's markup, and you can check it in an afternoon instead of assuming it for years. We assumed it for too long.
How we measured it, so you can repeat it
The method is deliberately boring, because boring is repeatable.
We took several of our own recent posts on each platform. For each post we fetched the page the way an anonymous visitor or a crawler would — a plain HTTP request for the raw HTML, no logged-in session, no rendering tricks. Then we parsed the markup and asked three things of each page:
Are our links to our own domain actually present in the served HTML? Do those links carry rel="nofollow" or an equivalent? And what does the page's own <meta name="robots"> tag say?
From those three answers you can derive the only thing that counts: a link passes authority to your site only if the link is followable and the page is indexable. Fail either test and the answer is "passes nothing," regardless of how the other half looks. We ran this on 2026-09-02.
If you want to do the same, you do not need a paid SEO suite. curl the URL, save the HTML, grep for your own domain, look at the rel attribute on those <a> tags, and look at the robots meta tag in the <head>. That is the whole method. The value is not in tooling; it is in checking the served markup instead of trusting the story you were told about it.
What came out
We checked three platforms. None of them pass link authority to our site.
That is the headline finding and we are not going to soften it. On the platforms we could inspect, our backlinks were not doing the SEO job we had been quietly crediting them with.
The second number is more interesting than it first looks: none of the three nofollow every link of ours. So this is not a simple case of "the platform slaps rel=nofollow on everything and that is that." At the link level, our links were often followable. And yet the net result was still zero authority passed.
The way those two facts fit together is the actual lesson. If the links are frequently followable but authority still does not pass, the block is not (only) at the link level — it is at the page level. A followable link on a page that search engines are instructed not to treat as a ranking page passes nothing downstream. The clean-looking link is real; the page it sits on is what neutralizes it. That is the trap, because when people check "is my backlink nofollow?" they look at the link and stop. The link can pass every test you thought to run and still be worthless because of a tag in the <head> you did not think to read.
There is also a platform we could not check at all. Tumblr refused our anonymous requests, so we have no measurement for it — not a good result, not a bad result, an absence of one. We are not going to guess at what we could not fetch.
Limits: what this does not prove
We would rather state these up front than let you over-read the finding, because the whole point of publishing our own numbers is that the numbers stay honest, including where they are thin.
The sample is small. This was a few pages per platform, not an exhaustive crawl of everything we have ever posted. A handful of pages can miss variation — a different post type, a different link placement, a page template we did not happen to sample.
The check goes stale. Platforms change their markup whenever they like and announce none of it. A robots tag or a rel policy can flip next week and our 2026-09-02 snapshot would be wrong without us knowing. This is exactly why we now repeat the check on a schedule rather than treating it as settled.
We read markup, not a live crawler. Parsing rel attributes and robots tags tells you what the platform instructs search engines to do. It is a strong signal and it is how these decisions are actually communicated, but it is inference from markup, not a recording of what any given search engine did with our pages. We are reading the instructions, not watching them be obeyed.
And this proves nothing about anything larger than link authority. It does not say these platforms are worthless — a platform can be worth publishing on for a live audience, for readers who click through, for reaching people who would never find our own site, none of which depends on passing SEO authority. It does not say anything about platforms we did not check, or about Tumblr, which we could not check. A finding about markup on three platforms on one date is exactly that, and stretching it into "guest publishing is dead" would be the same lazy move as the assumption we started with, pointed the other way.
What we changed because of it
We stopped counting on guest-platform backlinks as an SEO play. That reasoning is gone from how we decide where to publish. If a link on one of these pages ever does pass authority, we will treat it as a bonus we did not plan around, not a mechanism we are building on.
We re-justified the platforms on honest grounds. We still publish on some of them, but now the question is "does a real reader find this here, and do they click through," not "does this feed our domain." A platform that brings readers earns its place on that basis alone. A platform that brings neither authority nor readers does not get our time because a backlink myth said it should.
We got stricter about our own outbound links inside these posts. If the page we are publishing on will not pass authority anyway, stuffing it with links to our own domain buys us nothing and reads like link-farming to the platform — the exact behavior that gets pages demoted or de-indexed. Fewer links, each one there because a reader would actually click it, is now the rule. This report links to us barely at all, on purpose.
And we put the measurement on a repeat schedule instead of filing it as a one-time answer. The markup will change; our knowledge of it should not be a single dated snapshot we quietly trust for two years, which is how we got into this in the first place.
The uncomfortable summary is that we were crediting a channel with an SEO benefit we had never once verified, and when we finally read the markup, the benefit was not there on any platform we could check. The fix was not a new tool. It was fetching our own pages and reading what they actually say.
Top comments (0)