Link reclamation after a replatform, rebrand or purge

On a fashion or beauty site that has replatformed in the last five years, I usually find that 8 to 15 per cent of historic external links point at URLs that now return 404, redirect to the homepage, or pass through three hops before resolving. Those links were already earned and paid for. Recovering them is cheaper than earning replacements, and it is the first thing I look at before anyone commissions new outreach. What the audit reads: Referring domain exports from Ahrefs, the Search Console links report, Bing Webmaster Tools, and 90 days of raw server logs. The logs matter because they catch referrers no backlink crawler has indexed, which in this category is often a print title's web edition or a closed forum. Every dead target is then pulled from the Wayback Machine, so a redirect can go to a genuinely equivalent page rather than to whatever is nearest. The decisions, which are the actual work: A blanket redirect to the homepage is treated as a soft 404 and passes nothing, so it is never the answer. Each dead URL gets a decision: redirect to a real equivalent, restore the page, or serve a permanent archive page that states what the product was and what replaced it. That last pattern is what discontinued SKUs need. Fashion sites delete a season and then wonder where the links went; a colourway that a magazine linked to in 2021 should still resolve to something honest in 2026. Collection URLs that died with a season get the same treatment. Redirect chains are flattened at the edge so every hop becomes one hop, and the rules are handed over in a format your platform can actually ingest rather than as a spreadsheet somebody has to retype. Old domains from a rebrand or an acquisition get checked individually. It is common to find a previous brand domain holding several hundred referring domains and serving a parking page, or expiring in four months with nobody watching the renewal. Unlinked mentions: Monitoring covers brand name, founder name, product and collection names, and reverse image search of the press kit photography. A picture desk that used your image and credited you in text converts at roughly one ask in three, against about one in five for a plain text mention, because there is an obligation attached and the editor knows it. The ask names the exact URL and says what the reader gets from following it. One follow-up, then the record is closed. The disavow question: No disavow file unless you have a manual action in Search Console. Google's own position is that most sites never need one, and an uploaded file removes links you cannot get back if you were wrong about them. If a previous vendor bought placements and you want them gone, that is a removal conversation with named risk attached, and I will give you my view in writing rather than quietly filing a text file. Not included: No outreach to buy links back. No mass removal requests to webmasters, which is a treadmill with no end. This is not a technical SEO audit, though you get every status code and chain I encountered, and you should hand those to whoever owns that work. Who does not need this: Sites under three years old on their original platform with no rebrand behind them. There is usually nothing to recover, and I will tell you that from a 30-minute look rather than sell you a report about it. The same applies if your referring domain count is small: the arithmetic only works once there is history to reclaim.