How Links Affect Your Google Rankings

Links are still the biggest ranking signal. Not the only one, but the biggest.

Key Takeaways

• The Original Idea Was Simple
• What the Studies Actually Say
• How Google Might Weight Links Differently Now
• The Relevance Question
• Anchor Text Still Matters, But Less Than You'd Think
• The Follow vs. Nofollow Saga

I know that's a bold claim in 2024, when Google keeps telling us they've moved beyond links, when every other conference talk focuses on E-E-A-T or user engagement metrics or whatever the latest emphasis happens to be. But the data keeps pointing the same direction. Study after study, year after year — the number and quality of backlinks pointing to a page remain the strongest predictor of where that page shows up in search results. You can argue about the margins, about whether it's links or content quality that matters "more" in some abstract sense. But strip away the noise, and links still sit at the top of the pile.

That doesn't mean they work the way they used to. Not even close.

The Original Idea Was Simple

Google's founding insight — the thing that made it better than AltaVista and Yahoo and every other search engine in the late '90s — was treating links as votes. Larry Page and Sergey Brin wrote about it in their original Stanford paper. If a page had more links pointing to it, it was probably more important. If those links came from pages that themselves had lots of links, even better. PageRank, they called it. Elegantly recursive. And it worked shockingly well for its time.

The web was smaller then. Fewer people understood how search engines operated. Links were mostly organic — someone created a page about astronomy and linked to NASA because, well, NASA had good information about astronomy. The signal was clean. Links really did represent a kind of democratic vote about what was useful on the internet.

That lasted maybe five years before the manipulation started in earnest. Link farms. Blog comment spam. Paid links on every sidebar that would accept a few bucks. The signal got polluted, and Google has spent the last two decades trying to separate the genuine votes from the manufactured ones. That fight is ongoing. It hasn't been won, and I'm not sure it ever will be.

What the Studies Actually Say

Let's talk numbers, because this is where things get interesting and also where people tend to get sloppy.

Backlinko ran a study of 11.8 million Google search results back in 2020 and found that the number of domains linking to a page had the strongest correlation with rankings of any factor they measured. The top result in Google had an average of 3.8 times more backlinks than results in positions two through ten. That's a massive gap. And the correlation between referring domains and rankings was stronger than anything related to content length, keyword usage, or page speed. This ties directly into Understanding Link Equity: A Complete Guide, which is worth reading next.

Ahrefs has published similar findings multiple times. Their analysis of over a billion pages found that 90.63% of all pages get zero organic traffic from Google, and one of the biggest reasons was a lack of backlinks. Pages with zero referring domains almost never rank. The correlation between referring domains and organic traffic was, by their measurement, the strongest single factor.

Now — and this is where I need to be careful — correlation isn't causation. Everyone says that, and everyone's right to say it. It's possible that pages ranking at the top earn more links because they rank at the top, creating a self-reinforcing cycle. It's possible that the kinds of content that attract links also happen to be the kinds of content that satisfies users, and it's the user satisfaction that Google actually cares about. These are real objections. I don't dismiss them.

But Google's own leaked documents and various patent filings suggest that links are, in fact, a direct ranking input — not just a side effect. The original PageRank algorithm has evolved into something much more sophisticated, but the core concept of using link data to evaluate page importance hasn't been abandoned. Former Google engineers have said as much in interviews. Gary Illyes said at one point that links are among the top three ranking factors, alongside content and RankBrain. That was a few years ago, and Google's messaging has shifted since then, but I don't think the underlying reality has changed as much as the messaging implies.

How Google Might Weight Links Differently Now

Here's where things get speculative, and I want to be upfront about that. Nobody outside Google knows exactly how their algorithm works. What follows is informed speculation based on patents, leaked information, observable behavior, and what Google engineers have said publicly. It's the best picture we can assemble, but it's still a picture assembled from the outside.

The old model was relatively straightforward. More links meant more authority. Higher PageRank on the linking page meant the link was worth more. The anchor text of the link told Google what the target page was about. Done. You could almost calculate it on a spreadsheet.

The new model — or what we think the new model looks like — is layered with context signals that would have been impossible to process at scale fifteen years ago. Google almost certainly evaluates the topical relevance of the linking page now, not just its authority. A link from a highly authoritative page about cooking probably doesn't do much for your page about car insurance. The topical distance matters. A link from a moderately authoritative page about auto insurance might be worth more than a link from a massively authoritative page about an unrelated topic.

There's evidence that Google evaluates the placement of the link within the page. A link embedded naturally in the body content of an article likely carries more weight than a link stuck in a footer or sidebar. This aligns with the "reasonable surfer" patent Google filed years ago, which describes a model where links that users are more likely to click receive more weight. A link buried in a list of a hundred other links at the bottom of a page? Probably not getting clicked much. A link in the first paragraph of a relevant article? Much more likely to get a click, and therefore — if the patent reflects reality — much more likely to pass significant ranking value.

User behavior signals might modify link value too, though this is where I'm least certain. There's a theory that if a link drives actual traffic — actual clicks from real users — Google treats it as a stronger signal than a link nobody ever clicks. This would make sense. A link that people interact with is a stronger endorsement than a link that exists on a page nobody visits. But proving this is hard. Google has denied using click data as a direct ranking factor, though the DOJ antitrust trial surfaced internal documents suggesting it plays a larger role than Google has publicly admitted.

Temporal signals are another dimension. A link from a page published yesterday might be weighted differently than a link from a page published five years ago. Fresh links could signal that a page is currently relevant and being discussed, while older links might represent established authority. Google probably uses both, but the balance between them likely shifts depending on the query. A search for "best phones 2024" should favor recent signals. A search for "how does photosynthesis work" probably doesn't need to prioritize freshness. For the full picture, read What Are Backlinks and Why Do They Matter for SEO.

I've also noticed — and this is purely anecdotal, so take it accordingly — that links from pages that themselves rank well for relevant queries seem to carry disproportionate weight. It's like Google is saying: "I trust this page enough to rank it, so I'll trust its editorial judgment about what it links to." This creates an interesting dynamic where getting a link from a page ranking on page one for a related keyword might be worth as much as dozens of links from pages Google doesn't rank for anything.

The Relevance Question

This is a tangent, but it matters. Ten years ago, SEOs debated whether relevance mattered for link building or whether raw authority was all that counted. That debate is pretty much settled now. Relevance matters enormously.

I've seen sites build hundreds of links from high-authority domains in completely unrelated niches and see almost no ranking improvement. And I've seen sites build a handful of links from topically relevant, moderate-authority sites and jump several positions. The anecdotal evidence is overwhelming, and it aligns with what we know about how Google processes entity relationships and topical clusters.

Google has gotten really good at understanding what a page is about, what a site is about, and whether a link between two pages makes contextual sense. A link from a gardening blog to your gardening tools page fits. A link from a gardening blog to your cryptocurrency exchange page doesn't, and Google likely discounts or ignores it. This is a massive shift from the early days when any link from any page with high PageRank was gold.

The practical implication is that a targeted approach to earning or building links now matters more than it ever did. Blasting links across random high-authority sites is a waste of time and potentially a risk. Focusing on getting links from sites and pages that are topically aligned with your content? That's where the ranking value lives.

Anchor Text Still Matters, But Less Than You'd Think

Anchor text used to be the primary way Google understood what a linked page was about. If fifty links all pointed to your page with the anchor text "best running shoes," Google took that as a very strong signal that your page was relevant for the query "best running shoes." This led to predictably abusive behavior — people building thousands of exact-match anchor text links to manipulate rankings.

Google cracked down hard on this with the Penguin algorithm update in 2012, and subsequent updates have made anchor text manipulation even riskier. But anchor text itself hasn't become irrelevant. It's still a signal. Google still looks at it. The difference is that Google now considers anchor text as one of many signals rather than the dominant one, and it's gotten much better at detecting when anchor text patterns look unnatural.

The natural distribution of anchor text is something Google understands well. For most pages, the majority of inbound anchors are either branded (the site or company name), generic ("click here," "this article," "learn more"), or naked URLs. Only a small percentage — maybe 5 to 15% in most cases — use keyword-rich anchor text. When Google sees a page where 60% of inbound anchors are exact-match keyword phrases, that's a red flag. It doesn't look like organic linking behavior because it isn't.

What seems to work well now is a natural-looking distribution. Let the anchor text be whatever the linking person naturally writes. If you're doing outreach or guest posting, vary your anchors. Use branded text sometimes, descriptive phrases other times, and keyword-rich anchors sparingly. The days of controlling every anchor text for maximum keyword relevance are over, and trying to do it will hurt you more than help. There's plenty more to go through, and 10 Proven Link Building Strategies That Work is a great place to start.

The Follow vs. Nofollow Saga

Quick detour here because this comes up constantly. The nofollow attribute was introduced in 2005 to give webmasters a way to link to something without passing ranking value. It was Google's solution to the comment spam problem — if links in blog comments were nofollowed, there'd be no SEO incentive to spam them.

For years, nofollow meant "this link passes no PageRank." Full stop. But in 2019, Google changed the rules. They introduced two new attributes — sponsored and ugc — and reclassified nofollow as a "hint" rather than a directive. Meaning Google now decides whether to count a nofollowed link, rather than automatically ignoring it.

What this means in practice is murky. Some SEOs have run experiments suggesting nofollowed links from high-authority sources do move the needle. Others have found no effect. My best guess? Google selectively passes value through nofollow links when it trusts the context, but not in a way that's consistent or predictable enough to base a strategy on. Don't build your link strategy around nofollow links. But don't panic if your best links happen to be nofollowed — they probably still have some value, direct or indirect.

Internal Links Aren't an Afterthought

Most of this article has been about external links — links from other websites to yours. But internal links — links between pages on your own site — also affect rankings, and they're criminally underrated.

Google crawls your site by following links. If an important page is buried five clicks deep with only one internal link pointing to it, Google might not realize how important it is. If that same page is linked from your homepage, from related blog posts, from your main navigation — Google gets a clear signal that this page matters to you and should be given more attention.

Internal linking also distributes PageRank within your site. If your homepage has the most external links (which is true for most sites), those links are essentially pooling authority at the homepage. Internal links allow you to channel some of that authority to deeper pages — product pages, blog posts, category pages — that might not attract many external links on their own but are important for your business.

I've seen cases where simply adding internal links to a page — no new external links, no content changes, just better internal linking — led to a noticeable rankings bump within weeks. It's one of the most controllable, lowest-risk things you can do for SEO, and it's amazing how many sites do it poorly or not at all.

The Quality vs. Quantity Debate

This is one of those things where the "right" answer is obvious but the practical reality is messy. Yes, quality matters more than quantity. One link from the New York Times is worth more than a thousand links from random, no-traffic blogs. Everyone knows this.

But nobody wants to say this out loud: quantity still matters. Not in the sense that you should build a million garbage links. But in the sense that a site with 500 referring domains from decent, relevant sources is almost always going to outperform a site with 10 links from amazing sources. The correlation data supports this. And it makes intuitive sense — breadth of endorsement signals broad trust, not just the opinion of a few elite sources. If this is new to you, How Many Backlinks Do You Need to Rank on Google? breaks it down step by step.

The practical reality for most sites is that you're not choosing between quality and quantity. You're working with limited resources and trying to get the best links you can at the scale you can manage. A healthy link profile has a mix. A few high-authority links at the top, a solid middle of decent-quality relevant links, and a natural-looking tail of lower-quality links that inevitably accumulate as your site grows and gets mentioned in various places.

Where people get into trouble is when the profile is heavily skewed in one direction. All low-quality links? Looks spammy. All high-quality links and nothing else? Actually, that can look unnatural too, though it's a much better problem to have. The most natural-looking profiles have variety — variety in authority, in relevance, in link type, in anchor text, in follow status. That variety is hard to manufacture artificially, which is partly why Google uses it as a trust signal.

What About User Signals Replacing Links?

There's a persistent argument in SEO that user behavior signals — click-through rate, dwell time, bounce rate, pogo-sticking — are gradually replacing links as ranking factors. The logic is that Google has access to massive amounts of user behavior data (through Chrome, through Android, through search itself) and that this data is a more direct measure of content quality than links are.

I don't think this is wrong exactly, but I think it overstates how far along this transition is. User signals are noisy. They vary wildly by query type, by device, by time of day, by user intent. A short dwell time on a recipe page might mean the user found what they needed quickly — that's a success, not a failure. A long dwell time on a product page might mean the user was confused, not engaged. Google has to interpret these signals carefully, and I suspect the interpretation is still imperfect.

Links, for all their flaws, have a clarity that user signals lack. A link is a deliberate editorial decision. Someone chose to reference your content from theirs. That's a signal that's hard to fake at scale (Google's gotten pretty good at detecting link schemes) and relatively easy to evaluate. User signals are easier to spoof (click bots exist) and harder to interpret correctly.

My read on the situation? User signals are a growing factor, and they'll probably continue to gain weight relative to links over time. But links aren't going away. The shift is additive, not substitutive. Google is adding more signals to the mix, which dilutes the relative weight of links somewhat, but that's not the same as links becoming unimportant. A page with great user signals and terrible links probably won't rank well. A page with strong links and mediocre user signals has a much better shot.

Link Velocity and Patterns

Something that doesn't get talked about enough is the pace at which links are acquired. Google doesn't just look at your total backlink count — it looks at how that count changes over time. A natural link profile grows gradually, with occasional spikes when a piece of content goes viral or gets picked up by a major publication. An unnatural link profile might show a sudden influx of hundreds of links in a week, followed by nothing, followed by another spike. That pattern screams manipulation.

Link velocity matters in another way too. If your competitors are acquiring links faster than you are, your relative position erodes even if you're still building links. Rankings aren't static. Google is continuously recalculating who deserves to rank where, and the rate of new link acquisition is part of that calculation. You don't just need links. You need to keep earning them at a pace that's competitive for your niche.

This is why one-off link building campaigns have limited long-term value. You build 50 links in a month, your rankings improve, and then you stop. Meanwhile, your competitors keep building. Within six months to a year, the gains have eroded. Sustainable link acquisition — whether through ongoing content marketing, continuous outreach, or building a brand that naturally attracts links — is the only approach that produces durable results. If you want to go further, What Is Domain Authority and How to Improve It has you covered.

There's also the question of link decay. Links disappear. Pages get taken down. Sites go offline. Domains expire. One study estimated that about 5 to 10% of links disappear each year. That means if you're not acquiring new links, your effective link count is shrinking. Another reason why treating link building as a one-time project doesn't work.

Where Are Links Heading?

I genuinely don't know. And I'm suspicious of anyone who claims to.

The trends are visible. Google is investing heavily in AI, in understanding natural language, in evaluating content quality through means other than links. The Knowledge Graph, the neural matching systems, the multimodal understanding capabilities — all of this points toward a future where Google can evaluate content quality more directly, without needing the proxy signal that links provide.

But we've been hearing some version of "links are dying" for at least a decade, and it hasn't happened yet. Maybe it's because links still work too well as a signal. Maybe it's because the alternative signals aren't reliable enough yet. Maybe it's because when you think about it, the concept of citation — one source endorsing another — is just a really good way to evaluate trust on the internet, and no amount of AI can fully replace it.

Will links be the biggest ranking factor five years from now? I honestly couldn't say. Probably not the biggest in every context. Google is clearly moving toward a more nuanced, query-specific approach where different signals matter more or less depending on the type of search. For brand queries, links probably don't matter much at all. For competitive commercial queries, I'd bet they still matter enormously. For informational queries, maybe content quality and freshness and E-E-A-T signals will overtake links. The future is probably not "links matter" or "links don't matter" but "links matter this much for this type of query," with the weights constantly shifting.

What I do know is this: right now, in the current environment, ignoring links is a mistake. You can have the best content on the internet, and if nobody links to it, it probably won't rank for anything competitive. The sites winning in search almost always have strong link profiles. That's not a coincidence, and it's not going to change overnight. How much longer links stay this important is the open question. But it's a question, not a settled answer — and I think being honest about that uncertainty is more useful than pretending we know where this is heading.