Do Backlinks Still Matter in 2026?
An honest look at what backlinks still do, what changed, why small sites over-index on them, and what to spend that effort on instead.
Updated Aug 26, 2026 · 8 min read
Every new site owner asks this within about a week of starting. Here is the honest answer: what backlinks actually do, what changed, and why chasing them first is usually the wrong move.
The plain-language definition
A backlink is a link on somebody else's website pointing at yours. That is the entire concept, and it has not changed since 1998.
What has changed is what search engines do with it.
The original idea was elegant. A link was a cheap proxy for a human judgment: somebody read this page, found it useful enough to reference, and spent their own credibility saying so. Count those votes, weight them by who cast them, and you get a rough map of what the web considers worth reading.
That map is called PageRank, and a version of it still runs. But a link now does two separate jobs, and people conflate them constantly:
- It passes a ranking signal - some amount of credibility flows from the linking page to yours, scaled by how credible and how relevant that page is.
- It sends actual humans - people click links, and a link from a page your buyers already read can be worth more in customers than it ever was in rankings.
The second job is the one nobody optimizes for. It is also the one that survived intact.
A running example
Say you and one other person sell check-in and waiver software to climbing gyms. Roughly 400 gyms in your country could plausibly buy it. Your site is nine months old and has eleven pages.
Your main competitor has been around since 2016 and has links from a few thousand domains: trade press, gym-industry podcasts, a decade of conference sponsorships, and several hundred customer sites with a "powered by" footer.
You are never catching that number. The good news is that the number is not what you are actually competing on.
Why it matters
Because the belief that links are the bottleneck sends small sites down a six-month detour.
Here is how it usually goes. You read that you need backlinks, so you buy a guest-post package: fourteen placements on marketing blogs with names like growthhackerdaily, at $120 each, over three months. Nothing moves.
Nothing moves because nothing was ever going to move.
Those sites have no relationship to climbing gyms, no readers who run climbing gyms, and a link profile that makes their outbound links close to worthless. Meanwhile the query you actually needed to win, "climbing gym waiver software", returns six results, two of which are generic listicles that mention climbing once. That SERP was winnable with a page. Not with links.
The trap in one sentence: links are the most visible thing your competitors have, so they look like the reason your competitors win, even when the real reason is that they answered 200 questions and you answered eleven.
What actually changed
Four shifts, each of which quietly moved links down the priority list.
Bad links mostly get ignored now, not punished. Google's spam systems moved from penalizing manipulated links to simply neutralizing them, which means a bought link usually has no effect in either direction. That sounds like good news until you notice what it implies about the money you spent.
Google itself stopped ranking links near the top. Google's search advocates have said publicly for years now that links are no longer among the most important signals, and the guidance the company publishes for site owners barely mentions them. Take the specifics with salt, but the direction has been consistent and one-way.
Behavior signals turned out to matter a lot. Testimony and documents from the US antitrust proceedings against Google made clear how heavily the ranking systems lean on aggregated click and interaction data. A page people choose and stay on is doing something no link can do for it.
AI answers changed what gets cited. When an AI engine assembles an answer, it retrieves passages that match the question and quotes the ones that state something clearly. Site-level authority still tilts the odds, but a precise, well-structured paragraph on a small site gets quoted constantly, which is a genuinely new opening.
One clarification that saves a lot of confusion: Domain Rating, Domain Authority, and Authority Score are not Google metrics. They are three vendors' models of link strength, built from three different crawls. Google does not see them, does not use them, and has never published anything equivalent.
Where links still decide the outcome
All of the above is not "links are dead". It is "links are conditional". They matter most exactly where the money is, and least where most small sites should be playing.
| Query type | How much links decide it |
|---|---|
| Broad commercial head terms ("crm software", "personal injury lawyer") | A lot. Everyone has good pages, so off-page signals break the tie. |
| YMYL topics (health, finance, legal) | A lot. Trust signals are weighted heavily and links are one of them. |
| Specific long-tail questions | Little. Relevance and completeness usually win outright. |
| Narrow B2B and niche vertical terms | Little. The competing pages are often mediocre and thin. |
| AI answer citations | Indirect. Passage clarity and specificity do more of the work. |
Read that table as a map of where to spend the next quarter. If your niche lives in the bottom three rows, links are not your constraint and buying them will not make them one.
What to do instead
Concrete, in the order I would actually do them.
- Cover the question set completely first. Before any outreach, get to the point where every real question in your niche has a page. For the climbing-gym example that is waiver law by state, insurance requirements, check-in hardware, day-pass pricing, and the twenty comparison and migration questions buyers ask. Depth is the thing you control entirely.
- Get listed where buyers already look. Vertical directories, industry association member pages, integration and partner directories, app marketplaces. These are links, but more importantly they are the pages your buyers open on purpose.
- Publish one thing per year that has data in it. Survey 60 gym owners about what they pay for software and publish the numbers. Original data is close to the only asset that reliably earns unpaid links, because journalists and bloggers need something to cite.
- Fix your internal links before your external ones. You control every link on your own site, and on a small site the internal structure is usually leaking more than the outreach could ever recover. It is free and it works this week.
- Reclaim unlinked mentions. Search your brand name and find the forum threads, roundups, and posts that mention you without linking. A polite email converts a decent share of them, and the relevance is already perfect.
Notice what these have in common. Every one of them produces something a customer might read, with a link as a side effect. That is the direction the causality is supposed to run.
The healthy version of link building looks like a business doing normal business things in public: shipping something worth mentioning, showing up in the places your industry gathers, and publishing numbers nobody else has. If a tactic only makes sense as an SEO tactic, it is probably one of the ones that stopped working.
Going deeper
- Relevance beats raw strength. One link from a climbing-industry newsletter with 3,000 readers is worth more than ten from high-scoring general blogs, both as a signal and as customers. Vendor metrics cannot see this, which is exactly why link-buying markets price on them.
- You probably do not need the disavow tool. Google has been clear that most sites should never touch it, because the systems already discount junk links. Disavowing aggressively is a way to accidentally remove links that were helping.
- Nofollow is a hint, not a wall. Since 2019 Google treats nofollow, sponsored, and ugc as hints it may or may not follow. A nofollow link from a page your buyers read is still a good link.
- Deep links beat homepage links. Most outreach ends with a link to a homepage, which helps the page you least need help with. If you are going to spend effort, point it at the article you want to rank.
- Link velocity panic is mostly folklore. There is no public evidence of a natural-looking growth rate you must not exceed. Real coverage arrives in bursts, and search engines know it.
- Where the standard advice does hold: if you are competing for broad commercial terms against funded competitors, the honest answer is that you need a real PR function, not a content strategy. Pick a narrower fight or budget for the wider one.
One aside about tools, including this one
RankSpiral does not build backlinks. It does not have an outreach module on the roadmap, it will not find you guest post opportunities, and it cannot make a domain older than it is.
What it does is the other half: mapping every question in your niche and getting genuinely good pages written against them, which is the half a small site can actually win on. If you want the concepts behind that, the clustering and topical authority pages are the ones to read next.
Any tool that promises to build links at scale is selling you the thing search engines got good at ignoring.
Topical Authority: Why Depth Beats Breadth→ Keyword Clusters and Pillars, Explained→ How AI Search Reads Your Content→Was this page useful?
Tell us what was missing and we'll fix the page.