Before you touch a single template, you can decide every page's fate in a spreadsheet by grading it on three things: whether anyone actually finds it, whether the content is worth keeping, and whether moving it carries risk. Three green lights means migrate. Content worth keeping that nobody can find, or findable content that's embarrassing, means rewrite. No traffic, no links, no job means kill. The build comes after the cuts, not before.
Here's why I keep banging this drum. You've got the shiny new CMS humming in a demo tab. The design system looks fantastic. Somebody has already fallen in love with the hero component. And in that exact moment, the question nobody wants to ask out loud is this: are we really moving all nine thousand pages, including the 2016 press release about a parking lot reopening? (Yes. Yes, that is what the plan currently says.)
So let me pitch you a different week.
Watch the film before you run the scrimmage
Think of a coach the week before training camp. He doesn't line everyone up on the field, blow a whistle, and sort it out live. He sits in a dark room and watches tape. He grades every player on last season's film before anyone sweats a drop. By the time camp opens, he already knows who's a starter, who's a project, and who's getting the quiet conversation by the gate.
Your sandbox is the field. It's expensive, it's slow, and it's a terrible place to discover that a page was never worth suiting up. The tape is your crawl: a flat export of every URL with the numbers that actually predict whether a page earns its spot. You grade in the dark room. You build on the field. In that order.
And the lovely part about grading on tape is that it's cheap to change your mind. Reclassifying a page in a spreadsheet costs you a cell. Reclassifying it after it's been rebuilt, styled, and QA'd costs you a sprint.
The three grades
Every page gets one verdict. No maybes, no "revisit later" (later never comes, later is a myth).
| Grade | What it means | Rough test |
|---|---|---|
| Migrate | Starter. Moves as-is, keep the URL, keep the content. | Earns traffic or links, content still true, low move risk. |
| Rewrite | Project. The topic deserves a page; this execution doesn't. | Good demand but thin, dated, or duplicated content. |
| Kill | Cut. Doesn't make the roster. Redirect or let it 410. | No traffic, no links, no internal links, no reason it exists. |
Notice that "rewrite" is not a polite way of saying "migrate, but I feel guilty." A rewrite is a promise you're committing budget to. If you can't name who writes it and when, it's not a rewrite. It's a kill wearing a nicer jacket.
The measurements you actually tape up
You can't grade on vibes, and you definitely can't grade on "the stakeholder likes this one." Pull real columns: organic entrances over the last twelve months, search impressions, external backlinks, inbound internal links, last meaningful update date, word count, and whether the content is a near-duplicate of something else on the site. That last one matters more than people expect, because migrations love to quietly carry three versions of the same page across and let them fight for the same query on the new domain.
A page with zero entrances, zero impressions, zero links pointing in, and a near-duplicate twin is not a hard call. It's a cut, and the only thing standing between you and the cut is sentiment. A page with strong impressions but a two-minute time-on-page and a 2017 byline is your classic rewrite: the demand is real, the execution has aged like milk. And the quiet winners, the pages pulling steady organic traffic and a handful of honest backlinks, migrate untouched, URL and all, because the fastest way to break something is to improve it.
This is roughly the grading engine we built Replatform Radar to run across a whole site at once, because doing it by hand past a few hundred URLs is how audits die in a half-filled tab. But the logic above works in any spreadsheet. The tool just saves you the squinting.
Cutting hurts, and you do it anyway
The kill pile is where the project earns its keep, and it's where everyone loses their nerve. Someone will defend a page with no traffic and no links because "a customer might need it someday." Maybe. But a page nobody finds isn't serving that customer today, and hauling it to the new platform doesn't make it findable. It just makes it dead weight in a faster car.
Every page you kill is a page you don't redirect-map, don't QA, don't template, and don't explain to Google. The cull is not the sad part of the migration. The cull is the discount.
Here's where I land: grade the whole roster on paper before you build anything, and treat every verdict as final until a number proves it wrong. If the evidence is genuinely a coin flip, migrate and leave it alone, because moving a page you don't touch is cheap and safe. Everything else gets a decision and a date. The teams that skip this don't save time. They just move the argument to the most expensive possible room, with the template already built, the clock already running, and the coach already out of whistle.
Roster's due Friday. The parking-lot press release did not make the cut.