Competitor Link Gap Analysis, Classified Row by Row

A link intersect across four competitors usually returns between 3,000 and 6,000 referring domains. Somewhere between 60 and 120 of them are worth a decision. This engagement is the work of deciding which, and stating what each of those would cost you in money, in relationships, or in time you do not have. Why the export is not the answer: Gap tools compare domains, and the domain is the wrong unit. A link on a page with no rankings, no search traffic and 140 outbound links is not the same asset as a link on a page that has held position for a query since 2019, whatever the two domain scores say. So the filtering happens at page level: does the linking page rank for anything, does it receive search traffic, is it indexed, how many external links does it carry, and is it one of 300 copies of the same wire story being counted 300 times. What the classification does: Every surviving row gets one of five labels. Obtainable by reporting, where the page cites a study or dataset you could plausibly beat. Obtainable by relationship, where a person made that decision and can be reached. Bought, where the footprint says so and I attach an estimated price from what that publisher's inventory sells for. Unobtainable, meaning a partnership, an investor, an acquisition, a founder's old university. Artefact, meaning a scrape, a mirror, an expired domain, or a link that is no longer on the page. Bought and artefact rows together usually account for a third of the export, and knowing which third is what stops a team funding a chase. What else is in the file, which is not a gap list: A read on which competitor's profile is a liability. In games and sport there is nearly always one competitor whose numbers look strong because of a paid network that will not age well, and teams copy them because nobody told them what they were copying. I name that competitor and show the evidence. And the queries where the advantage is not links at all. Roughly half the ranking gaps I examine close with a page that answers the query in the format the results page is rewarding, and no volume of link acquisition fixes a page of the wrong type. I say so in writing, even though I sell the other thing. Scope boundary: This is analysis and nothing else. No outreach is sent, no placements are made, and the engagement is not a lead-in to one. Several clients have taken the sheet to their own team and run it themselves, which is a good outcome and I would rather it stayed available to them. Inputs: Ahrefs and Majestic taken as a union, Semrush as the third view, deduplicated and then re-crawled with Screaming Frog to confirm each link is live and that the rel attribute is what the tool claimed. Your Search Console link export as ground truth for your own side. In Nordic verticals I read local results pages and Sistrix rather than trusting a global visibility score to describe a Danish or Swedish market. Delivered as: One sheet with the classified rows and the filters left in place so your team can re-run them, a two-page summary stating three moves in order of return, and a 60-minute call with whoever owns the budget. No slide deck. Who this is not for: Anyone who needs the conclusion to be that they need more links. Sometimes it is. Often it is not, and if the decision has already been made upstairs, this sheet will read as an obstacle.