You score a page migrate, rewrite, or kill by asking three questions in order: does it earn traffic or citations, is anything valuable pointing at it, and would you be proud to see it on the new site. A page that passes all three you migrate as-is. A page that matters but reads like a 2014 press release you rewrite. Everything else you kill. And you can decide all of it in a spreadsheet, weeks before the new platform has a single template built.
The adjuster and the wreck
Imagine this: someone rear-ends you at a stoplight, and now your car is on a flatbed heading to a lot you'll never visit. A few days later a claims adjuster walks around it with a clipboard. He isn't emotional about it. He isn't asking whether you loved that car or whether it got you through college. He's asking one cold question: is what this thing is worth bigger than what it costs to make it whole again?
If yes, they fix it and hand you the keys. If the frame is bent but the car's worth saving, they rebuild it. If the repair costs more than the car will ever be worth, they total it, cut you a check, and it goes to the crusher.
That is exactly the posture you want walking into a content inventory. Not sentimental. Not "but I wrote that one." An adjuster with a clipboard, deciding what's worth moving to the new site before anyone has poured a foundation.
Three questions, asked in order
Order matters, because the first question does most of the sorting for free. Ahrefs studied over a billion pages and found roughly 96 percent of them get zero organic search traffic from Google. Ninety-six percent. On a real corporate site the number is kinder, but not by as much as anyone hopes. So the first thing you check for every URL is whether it earned anything in the last twelve months: sessions, conversions, or the newer prize, getting quoted by an AI answer engine.
Second question: is anyone pointing at it? External backlinks, internal links from pages you're keeping, referral traffic, an inbound redirect chain. A page with no traffic but a strong backlink is not dead. It's a mailbox other people still write to, and if you kill it you throw away link equity you'll spend a year trying to rebuild.
Third, and only third: is it any good? Accurate, on-brand, not duplicated four times across old campaign folders. This question is last on purpose, because a beautifully written page nobody reads and nobody links to is still a page nobody reads and nobody links to.
Turning three answers into one verdict
Line the answers up and the verdict falls out almost mechanically.
| Traffic or citations? | Links or authority? | Quality? | Verdict |
|---|---|---|---|
| Yes | Yes | Good | Migrate as-is |
| Yes | Yes | Weak | Rewrite (keep the URL) |
| No | Yes | Any | Rewrite or consolidate, never kill |
| Yes | No | Weak | Rewrite |
| No | No | Any | Kill and redirect the URL |
Notice the one row that never says kill: anything with links attached. That's the mailbox rule. You can gut the house behind the door, but you keep the address, because the mail is still coming.
Migrate: the ones you drive off the lot
These are your cornerstone pages, your best-converting product pages, the evergreen guide that has quietly ranked for six years. Migrate means you carry the content, the URL, the metadata, and the internal links across intact. The temptation on a shiny new platform is to "improve" these while you're in there. Resist. You don't repaint a car that's running fine just because you're already at the shop. Move it, verify it renders, leave it alone.
Rewrite: bent, not broken
The rewrite pile is where migrations quietly blow their budget, because it's real work masquerading as a checkbox. These pages have earned something worth keeping (a URL, a ranking, a backlink) but the actual words are dated, thin, or duplicated. The frame is fine. The body needs panel work. Keep the URL, keep the redirects pointing at it, and schedule the writing as its own line item with its own hours. If you fold rewrites into "content migration" and assume a copy-paste, you will discover in week ten that half your inventory needs an author, not a script.
Kill: the write-off nobody mourns
Expired promotions. The 2019 webinar landing page. Four near-identical "about our values" pages from four reorganizations. Killing them is the point of the whole exercise, because every dead page you carry over is weight the new site has to drag, crawl, and secure forever. But killing a URL is not deleting it into the void. You total the content and redirect the address to the nearest living relative, so the mail still gets forwarded. A kill without a redirect is just a future 404 with a delay timer.
Do the verdicts before the platform exists
Here's why the clipboard beats the sandbox. If you wait until the new CMS is stood up to decide what deserves a home in it, you have already paid to model, template, and QA thousands of pages you were going to kill anyway. The scoring is a pure content-and-analytics exercise: it needs your traffic data, your backlink data, and a duplicate check, none of which care what platform you're on. Crawling the old site and joining those signals into one migrate/rewrite/kill column per URL is precisely the pre-cutover pass a tool like Replatform Radar is built to run, but you can absolutely start it by hand in a sheet this afternoon.
So here's my actual position, no hedging. Score every page on paper before the new build starts, and be ruthless in the kill column. The single most expensive mistake I see is teams migrating everything "to be safe" and promising to clean up later. Later never comes, and the clutter you carried over is now baked into the platform you're stuck with for the next five years. Be the adjuster. Total the wreck. Nobody has ever stood in a crisp, fast new site and wished they'd kept the 2019 webinar page.