Competitor Backlink Analysis: Their Backlinks Are Not Your To-Do List
A competitor's backlink profile is a record of transactions that already happened. Most of them were not available to you then and are not available to you now.
Competitor backlink analysis is the most-performed and least-useful ritual in link building. You export the domains linking to three rivals, subtract the ones already linking to you, sort what is left by domain rating, and hand somebody a spreadsheet. The spreadsheet looks like a plan. It is not a plan.
It is a list of transactions that already closed. Each row records a decision some publisher made, at a particular moment, for a particular reason, in favour of a particular company that was not you. The export preserves the outcome and throws away the only field that matters: why.
Competitor backlink analysis, backlink gap analysis and link gap analysis all name the same export, and the industry uses the three terms interchangeably. The disagreement is not about what the tool produces. It is about whether the thing it produces is a plan.
Strip that field out and every row looks equally actionable. Put it back and most of the list disappears, because most competitor links are not available to anybody but the competitor. That is not a defeatist reading. It is the reading that turns a 400-row export into the fifteen rows you can actually work.
Competitor backlink analysis treats a moving target as a static list#
Start with the mechanical problem, before the interesting one.
Ahrefs took 10,000 non-branded keywords with monthly search volumes in the 2,000 to 5,000 range, pulled the top 20 ranking US pages for each, and measured how many followed referring domains each of those 200,000 pages had picked up in the previous three months. Their finding: "most of the #1 ranking pages tend to get followed backlinks (from new referring domains) at a pace of between +5% and +14.5% per month." For everything below the top spot, they "saw a more evenly-distributed performance between +0% and +14.5% per month," with the median pace falling as ranking position declines.
Ahrefs spell out what this does to the arithmetic everybody performs: if a top-ranking page has 55 referring domains and you decide you need 55 too, then by the time you have built them "the backlink profile for the current top-ranking page will have grown by between +17% and +50%, meaning that it'll now have links from between 64 and 83 referring domains."
There is a second, quieter finding in that study worth holding on to, because it undercuts the whole velocity-matching genre. When the same researchers looked for a relationship between how fast a top-three page acquired links and whether it held its position, they report that they "failed to distill any observable relationship between the pace of backlink acquisition for top-ranking pages and their respective ranking positions." Matching a competitor's link velocity is not a strategy anybody has shown to work. It is a number that feels like a target.
Two caveats worth stating out loud, because a lot of this piece leans on Ahrefs data. Both of their studies are correlational, and both are built on Ahrefs' own crawl of the web rather than on Google's index, so they describe what one link graph can see rather than what Google scores. Backlinko ran into the same wall analysing 11.8 million Google search results and said so plainly: "As this is a correlation study, it's impossible to determine the underlying reason behind this relationship from our data alone." That does not make any of these numbers useless. It makes them evidence about shape and scale rather than proof of mechanism, which is all the argument here needs them to be.
So the export is stale, and the obvious metric to chase is not load-bearing. That is the mechanical problem. The structural one is worse.
Every competitor backlink analysis needs a repeatability column#
Here is the reframe. In the link building pillar we argued that every link is a transaction, sorted by what the publisher receives: you fixed something, you supplied something, you already earned it, or you traded. Competitor backlink analysis is that model applied backwards. You are not choosing a transaction. You are reading the receipts of transactions somebody else closed.
Which means the only column worth adding to the export is repeatability: given what my company actually has, could this same transaction close for me?
Structural links are the ones worth most of your attention, because the publisher has already published their criterion. A resource page for "best project management tools" is telling you, in writing, what it takes to be listed. So is a comparison round-up, an industry directory, a university library guide. You do not need a relationship or a news hook. You need to meet a stated standard and tell somebody you meet it. This is the same inversion that makes broken link building work: the publisher's own page defines the gap.
Supply links are conditional. Your competitor got cited because they had a number nobody else had. You can get the same link if you produce the same category of material, which is a real project with a real budget, not a line on a prospect list. That is the whole subject of digital PR link building, and pretending a supply link is a structural one is how outreach campaigns end up pitching nothing.
Event links are the pure noise in every export. A competitor raised a Series A and got forty links in a week. Those forty domains are now sitting in your gap list, sorted by domain rating, looking exactly like opportunities. They are not opportunities. They are the residue of a thing that happened to somebody else.
Most competitor backlinks are relationships, and relationships do not export#
The fourth category is the one people underestimate most, and there is data on how common it is.
Ahrefs took a sample of around 10,000 non-branded queries with keyword difficulty between 40 and 60, pulled the top 10 ranking pages for each, and measured the overlap between the sites linking to a page and the sites that page links out to. The result: "26.4% of the 140,592 domains in our sample have no reciprocal links. 73.6% do." Nearly a third of the sites studied, "27.4%", have "at least a 15% overlap between the sites to which they link out and the sites that link to them." At page level, "43.7%, almost half, of the 112,440 top-ranking pages we studied have some reciprocal links."
Read that carefully, because it is easy to misread. Ahrefs are not saying three quarters of links are link exchanges. They are saying that mutual linking between sites that operate in the same space is ordinary, which is a different and more useful claim. It means a large share of any competitor profile is the visible footprint of a network the competitor is inside and you are not: the podcast they went on, the partner who wrote them up, the agency they work with, the newsletter written by a friend.
None of that is cheating. All of it is real. And none of it exports. If you put those domains on a prospect list and send a pitch, you are writing to a publisher whose reason for linking last time was "I know them", which is not a criterion you can satisfy in an email.
How to run a backlink gap analysis that produces a plan instead of a list#
The practical version of all this is a change of order of operations. Classify first, filter second, sort last. Most teams run a link gap analysis in exactly the reverse order and then wonder why the reply rate is bad.
| Step | What most teams do | What actually works |
|---|---|---|
| 1 | Export the gap list for three competitors | Same, but pull the linking page, not just the domain |
| 2 | Sort by domain rating | Open the linking page and record why the link exists |
| 3 | Assign the whole list to outreach | Drop every event and relationship row |
| 4 | Send a template | Split the rest into structural now and supply later |
| 5 | Report emails sent | Report structural candidates found per hour |
Two details carry most of the value.
Pull the linking page, not just the domain. The domain tells you nothing about the transaction. The page tells you everything: a resource list is structural, a news article about a funding round is an event, a "partners" page is a relationship, a post citing a chart is supply. You cannot classify from a domain column, which is exactly why the standard workflow skips classification.
And use the overlap intelligently. A domain linking to one competitor is weak evidence of anything. A domain linking to three or four of them, on separate pages, over separate years, is strong evidence of a structural criterion, because no publisher has a personal relationship with every company in a category. Multi-competitor overlap is the cheapest repeatability signal available, and it is the one output of competitor backlink analysis that genuinely deserves to be sorted to the top.
What you are left with after all that is short. A few dozen structural candidates, each of which still needs somebody to find the right person at that site, work out what the page is actually for, and write something true about it. That research is the real cost of link building, and it is the part GenLink runs end to end: it takes a classified target list, finds the contact, writes the pitch against what the page publishes, sends from a warmed mailbox and follows up. The classification is your judgement. The grind after it does not have to be.
If you want the mechanics of that stage on their own, link building prospecting covers building and qualifying the list, and link building outreach covers the message.
Domain rating is the wrong axis to sort a competitor link analysis by#
One more habit worth breaking, since it is welded onto every export.
Domain rating, domain authority and every similar score are third-party approximations built by SEO tools out of their own link graphs. This is not a contrarian reading, it is the vendors' own position. Moz's documentation on the metric it invented says it flatly: "Domain Authority is not a Google ranking factor and has no effect on the SERPs." It is, in their words, "a predictive metric developed by Moz and is not used by Google to determine search engine rankings."
The obvious objection is that authority scores do track with rankings. They do. Backlinko's analysis of 11.8 million search results found that "a site's overall link authority (as measured by Ahrefs Domain Rating) strongly correlates with higher rankings", and that a site's overall authority correlated more strongly than the authority of the individual page. But look at what that sentence actually measures: one vendor's score correlating with rankings, inside a study that says of itself it cannot establish cause. Correlating with rankings is a different property from being available to you, and availability is the only question a prospect list exists to answer.
So the scores are useful as a rough filter for obvious junk, and actively misleading as a priority order, because they rank rows by a proxy for strength while saying nothing about whether the transaction behind the link is one you could close.
A resource page at DR 42 that lists every tool in your category and has linked to four of your competitors is worth more of your week than a DR 88 news site that covered a funding round you did not have. The first has a criterion you can meet. The second has a story you do not own.
The same instinct to optimise the visible number shows up in anchor text optimization, where people reverse-engineer a competitor's anchor distribution and try to reproduce it. That is the same error in a different column: copying the output of a process instead of running a process. And before you spend anything on new acquisition, check what you already have with a backlink audit and a pass of link reclamation, because the cheapest links in any competitor comparison are usually the ones you already earned and quietly lost.
What competitor backlink analysis is actually good for#
Strip out the to-do list fantasy and the technique is still valuable, just for different things than people use it for.
It is excellent at finding structural pages. That is its highest and best use, and multi-competitor overlap is the filter that surfaces them.
It is excellent at diagnosing the shape of a market. If nearly every competitor link is an event link, you are in a category where press cycles drive links and no amount of outreach will substitute for having something to announce. If the profiles are thick with resource pages and comparison round-ups, you are in a category where a patient structural programme wins. That single read tells you which quadrant of the link building model you should be funding, which is a genuinely strategic output.
It is good at showing you what you cannot do yet. A competitor cited everywhere for a dataset is telling you the price of entry in their lane. That is a budget conversation, not a prospecting one.
And it is bad, in every instance, at producing a list of domains to email this week. The export is a record of other people's history. Your link building has to be a record of your own.
Frequently asked questions#
What is competitor backlink analysis?
Competitor backlink analysis, also called backlink gap analysis or link gap analysis, is the practice of examining the sites that link to your competitors in order to inform your own link building. The standard version exports the domains linking to rivals, subtracts the ones already linking to you, and treats the remainder as a prospect list. A more useful version adds a step: open the linking page, record why each link exists, and keep only the ones whose reason is a criterion you could also meet.
How do I do a backlink gap analysis?
Pull the referring pages, not just the referring domains, for three or four close competitors. Classify each linking page by the transaction behind it: structural, supply, event or relationship. Drop the event and relationship rows, because they were not available to you. Then prioritise what remains by how many different competitors the same page links to, since a page that has linked to several companies in a category is applying a published standard rather than favouring a friend.
What tools do I need for a competitor backlink analysis?
Any backlink index will do the export: Ahrefs, Semrush, Moz and Majestic all have a link gap or link intersect report, and the differences between their crawls matter far less than what you do with the output. The part no tool does for you is the classification, which needs a browser and a person reading the linking page. Budget accordingly. The licence is the cheap half of this exercise, and the hours spent opening pages are the half that decides whether the output is a prospect list or a spreadsheet.
How is competitor backlink analysis different from a backlink audit?
They point in opposite directions. A backlink audit looks inward at the links you already have, to find toxic patterns, lost links and pages that are wasting the authority pointed at them. Competitor backlink analysis looks outward at links somebody else has, to find publishers with a standard you could meet. Run the audit first: reclaiming a link you already earned is cheaper and faster than acquiring a new one, so there is no sense prospecting against a rival before you know what is already broken on your own side.
Should I copy my competitor's backlinks?
Only the repeatable ones. A resource page, a category round-up or a directory has an objective criterion you can meet, so those are worth pursuing. A link earned by a funding announcement, an acquisition or a personal relationship is not a target, because the thing that produced it was specific to them and no email will reproduce it. The practical test is to ask what the publisher received in exchange, and whether you have that thing to give.
How many backlinks do I need to catch a competitor?
There is no threshold, and the honest answer is that nobody has produced evidence for one. The count-matching exercise assumes referring domains are interchangeable units, which they are not, and it assumes the target sits still, which it does not. It also rests on correlation studies that explicitly decline to claim causation. A smaller number of links from pages whose subject genuinely overlaps yours will do more than a larger number harvested from a gap export, because relevance and the reason the link exists carry the weight that a raw total does not.
How often should I redo a competitor backlink analysis?
Less often than you think, and never as a reporting ritual. The classification work is slow and the structural pages it surfaces do not churn quickly. Re-running the export monthly mostly generates fresh event links from whatever your competitors announced that month, which is noise with a timestamp. Once or twice a year per competitor set is enough, plus an ad-hoc pull whenever you enter a genuinely new topic.
A competitor's backlink profile is a record of what worked for them, not a list of what is available to you. Open the pages, label the transactions, throw away the ones that were never yours, and what is left is a short list of publishers with a published standard you can meet. That is the honest output of competitor backlink analysis, and it is small enough to actually work through. The part that does not scale by hand is everything after it: finding the right person at each of those sites, reading what they publish, writing a pitch that is true about it, sending it from a warmed mailbox, following up, and triaging the replies so you only read the ones worth reading. That loop is what GenLink runs. You decide which links are worth having. The agent goes and has the conversations.
