Before you touch a single template, every page you own should already carry one of three verdicts: migrate it as-is, rewrite it before it moves, or kill it and redirect the URL. You make those calls on a spreadsheet, judging each page on the traffic it earns, the links pointing at it, whether a human or a machine can still read it, and whether anything else on the site already says the same thing. Do it before the build, not during, because the build is when it gets expensive.
Picture the last week before a house move. You know the one. The truck is booked, the movers are downstairs texting you, the meter is running, and you are standing in the kitchen holding a fondue set you have used exactly never. This is the worst possible moment to decide whether the fondue set sparks joy. You decide that weeks earlier, calmly, holding each thing in your hands, deciding: box it, fix it, or bin it.
A CMS migration is that move. Every page is an object in the apartment. And the number one way people blow the budget is by rebuilding templates first and sorting content later, which is exactly like paying movers by the hour to carry your fondue set up three flights of stairs so you can throw it out in the new place.
What the three verdicts actually mean
Migrate means the page is fine. It earns traffic, or it earns links, or it does a job (a pricing page nobody links to still has to exist). It moves with its URL intact, its content intact, and ideally nobody notices anything happened. Most of your effort should go toward not disturbing these pages.
Rewrite means the content is worth keeping but the wiring is shot. The topic is good, the URL has history, people find it. But the page is thin, or three years stale, or written for a search intent that has drifted, or it is one of nine near-identical pages that should be one. You keep the address and gut the interior.
Kill means the page goes to the curb and the URL gets a 301 to the nearest living relative. And before anyone panics: killing a page is not throwing away its authority. Google has said for years that a 301 redirect passes full PageRank. The link equity survives the move. The clutter does not.
The signals you score each page on
You are not judging on vibes. You are judging on a small set of things you can actually pull from a crawler, from analytics, and from your backlink tool. Here is the shape of it.
| Signal | Leans migrate | Leans rewrite | Leans kill |
|---|---|---|---|
| Organic traffic (last 12 mo) | Steady or growing | Some, but declining | Effectively zero |
| Referring domains | Real external links | A few, worth saving | None |
| Content quality | Current, accurate | Good topic, stale body | Thin or obsolete |
| Duplication | Unique | Overlaps, should merge | Full duplicate |
| Machine-readability | Clean, parseable | Fixable structure | Trapped in a PDF nobody opens |
| Business need | Required page | Useful, needs work | Nobody would miss it |
No single row decides anything. A page with zero traffic and forty referring domains is a rewrite, not a kill, because those links are real money and you would be an idiot to throw them out. A page with decent traffic that duplicates a better page is a merge, which is a fancy rewrite. Read the row as a whole, the way you would read a person, not a rule.
Why zero-traffic pages are the whole game
Here is the uncomfortable part. When Ahrefs studied around a billion pages, they found roughly 96.55 percent of them got no organic traffic from Google at all. Not low traffic. None. Your site is almost certainly no exception, and a migration is your one sanctioned excuse to admit it.
Every dead page you carry over is a page a crawler has to fetch, render, and decide about before it gets to the pages that matter. It is a template you have to test. It is a redirect you might have to map twice. Carrying it is not free, it just charges you later, quietly, in crawl budget and QA hours and the vague dread of a launch checklist that never gets shorter.
So kill with confidence. The point of scoring first is that by the time the truck arrives, you are not making decisions. You are executing them.
Do the sorting while it's still cheap
The whole argument is timing. A rewrite decided in a spreadsheet costs a writer an afternoon. The same rewrite discovered in staging costs a writer an afternoon plus a developer re-migrating the node, plus a QA pass, plus the meeting where everyone agrees it should have been caught earlier. Same fondue set. Third flight of stairs.
Pulling traffic, links, duplication, and readability signals for a whole site by hand is grim, which is the part a pre-migration scan like Replatform Radar exists to do for you: crawl everything, score each page, and hand you the migrate/rewrite/kill plan before anyone opens a template. But whether a tool builds the spreadsheet or you do, build it before the build.
So here is what I actually think, no hedging: if you have not scored your pages before your developers start the new templates, you are not early, you are behind. Sort the apartment before you book the truck. Your movers will thank you, or at least they will stop texting you from the stairwell.