How to earn dofollow backlinks as a new product

There is no rel="dofollow". It has never existed. “Dofollow” is shorthand for the absence of an attribute that suppresses the link — the default state of an ordinary <a href>. Three values suppress it:

• nofollow — the original: do not pass ranking signal through this link.
• sponsored — the correct marking for a paid or compensated placement.
• ugc — user-generated content: comments, forum posts, profile fields.

A link can carry more than one. The practical consequence is simple: the only way to know what a site gives you is to look at the markup of a real listing page. Not the sales copy, not the pricing table, not a third-party list — the markup.

We wanted a directory list we could defend, so instead of copying one, we wrote a checker: fetch an actual live listing page on each directory, find the outbound link to the listed product, read its rel. No inference, no assumption from domain authority, no taking the directory’s word for it.

Across the 51 directories we have run that check against so far:

The reasons for the unverifiable group are mundane rather than sinister. Twenty-one return HTTP 403 to automated checkers — a bot-protection default, not a hostile act. One rate-limited the check. Two render their outbound links client-side, so there is no rel in the HTML at all to read.

Sit with that number for a second. A deliberate, automated, purpose-built check could confirm a followed link on fewer than half the directories it looked at. Now consider a listicle claiming “dofollow” beside every one of two hundred entries. Either it solved a problem we could not, or nobody checked and the column was filled in from other listicles that also did not check. It is not the first one.

Our results, per directory and with the date each was last looked at, are on the directories page. Anything we could not verify is labelled unverified rather than quietly assumed.

Unverifiable is not a verdict on the directory. A site that returns 403 to a script is behaving exactly like a great many well-run sites behind a CDN; several of those 27 probably do give followed links. What we cannot do is tell you so without evidence.

If one of them matters to you, the check takes a minute by hand, because a browser is not a script:

1. Open a live listing page on that directory — any product, not necessarily yours.
2. Right-click the outbound link to the product and inspect it, or view source and search for the destination domain.
3. Read the rel on that <a>. Absent or harmless values means followed; nofollow, sponsored or ugc means it passes nothing.

The two client-side cases are worth a note of their own. If the link is injected by JavaScript, then whatever a crawler that does not render sees, it is not that link — which is a decent reason to treat the placement as traffic rather than as a backlink.

Three ways, in ascending order of effort.

In the browser

View source, find your domain, read the surrounding tag. That is the whole technique. Devtools works too, but view-source shows you the HTML as delivered, which is what a crawler gets.

With curl

curl -sS https://directory.example/products/your-product \
  | grep -o '<a[^>]*yourdomain\.com[^>]*>'

Crude, and perfectly sufficient. You are looking for whether the tag it prints contains a rel at all.

With the tool on this site

Point it at a listing page before you spend an evening submitting to forty directories. Point it at ours, too — that is the intended use.

Nobody links to a product because it exists. Links come from having done something linkable, and for a new product there are a handful of realistic sources. Roughly in order of effort.

Directories and launch platforms

The floor, not the strategy. They are fast, mostly free, and each individual listing is worth little — a page with a hundred outbound links is not a strong endorsement of any one of them. Their honest value is that they exist, they are indexable, and a few of them send real traffic. Do the verified ones, skip the ones that want money for a link, and stop when it stops being fun. An evening, not a project.

Some, including this one, ask for a reciprocal badge on your site in exchange for the listing. That is a trade, and you should decide whether you want it; we are not going to pretend it is a gift.

The tools you build on

Integration pages, app marketplaces, plugin registries, “built with” showcases and community pages belong to companies with an active interest in listing you, and they are systematically under-used. If you integrate with something, look for its directory. If you built on a framework, look for its showcase. These are usually followed, usually free, and usually topically relevant, which is more than most directories manage.

Open source

A genuinely useful library, CLI, template or dataset attracts links that no amount of outreach will. Awesome-lists, READMEs of projects that depend on you, docs pages, and issue threads. This is slow and it is real; the artifact has to be worth using.

Being the source of a number

The most reliable link-earning move available to a small team is to measure something nobody has bothered to measure and publish the method. This guide is an example: the reason anyone would cite the 51-directory figure is that we said how we got it and what we could not confirm. Numbers get quoted; opinions do not. You do not need scale to do this — you need one honest measurement in a corner of your field.

People

Podcasts, interviews, guest posts, conference and meetup pages, newsletters that cover your niche. Slower, unpredictable, and the only category that compounds into something other than links. If any of it is paid or compensated, it should carry rel="sponsored" — a publisher who offers you a followed link in a paid placement is telling you something about how they treat their own site.

Communities

Forums, Reddit, Discord, Stack Overflow, Hacker News. Nearly always nofollow or ugc, so treat them as distribution rather than as backlinks. They are frequently where the first real users come from, which is worth more than the link would have been.

Buying followed links. Selling or buying links that pass ranking signal is against Google’s link spam policy. A paid placement is supposed to carry sponsored. The reason we are firm about this is not squeamishness: it is that our own directory links are followed precisely because they are earned by a public rule rather than bought, and a site that sells followed links has nothing left to say about anyone else’s.

Bulk submission services. “We’ll submit you to 500 directories.” Most of those 500 are pages nobody reads and search engines discount, and you have no way to check what you got. If the seller cannot tell you which links are followed, they did not check either.

Private blog networks. Networks of sites built to link to each other. This is the manipulation that the policies exist to describe, the risk lands on your domain, and the people selling it are insulated from that risk by design.

Chasing a metric instead of a link. Domain authority scores are third-party estimates, not something any search engine publishes. Optimising the estimate is not the same as optimising the thing.

Here is the part most guides skip, because it makes the rest sound less exciting.

Nobody outside a search engine can measure the ranking contribution of an individual link. Not us, not the tool vendors, not the people selling packages of them. What you can measure is a short and useful list: whether the link exists, whether its rel suppresses it, whether the page carrying it is indexable, and whether it sends you referral traffic. Everything past that is estimation dressed as fact.

So the reasonable posture is: get the mechanical part right, since it is cheap and checkable, and do not build a plan on top of a number nobody can see.

And no, listings do not buy you AI citations

You will read that directory listings make assistants recommend you. We sell directory listings, and we are telling you there is no evidence for it. The clearest counterexample is sitting in plain sight: G2 and Capterra carry enormous authority in exactly this category and receive close to no citations. If listing density and domain strength were the mechanism, they would be the answer to every software question asked of an assistant. They are not.

What is worth doing for that purpose is upstream and unglamorous: being reachable and readable in the first place, and describing yourself the same way everywhere you appear. That is covered in the companion guide, including which of it is measured and which is folklore.

1. Week one: fix your own house. Your site should be server-rendered and reachable, with one clear sentence saying what it is. A link pointing at a page that cannot be read is a wasted link.
2. Week one: the verified directories. Work from a list where the link type was checked, not asserted. One evening, no money, and do not go back for the unverified long tail.
3. Week two: the tools you build on. Every integration, framework, marketplace and 'built with' page connected to your stack. This is the highest ratio of relevance to effort you will find.
4. Weeks three and four: make one linkable thing. A measurement, a free tool, a small open-source piece. Something with a method someone else would want to cite.
5. Ongoing: verify what you got. Re-check the links you earned. Sites redesign, and a followed link quietly becomes a nofollow one without anyone telling you.