From Framer to Next.js (Webflow, WordPress, and Ghost Too)

I open the network tab and throttle the connection to a slow mobile signal. The founder's site is live. The main image takes close to ten seconds to appear. Visitors are waiting that long right now, on every page load.

Francois Brill

Francois Brill

Designer + Builder

Nov 1, 2026

Last updated

A migration should leave every URL, analytics event, and form working while the renderer changes underneath. I baseline the live site first, rebuild in Next.js or Astro, cut over with one redirect map, then rerun the same test and publish the diff. This page leads with the Framer to Next.js path. Webflow, WordPress, and Ghost follow the same method.

The proof: Hire Team Up

hireteamup.com ran on Webflow. I rebuilt it in Next.js. It went live on October 19, 2026. The baseline test ran on the live Webflow site on October 13, using the same tool and settings I'd use after launch.

Mobile LCP (the main image) dropped from 9.7 seconds to 2.7 seconds. Desktop went from 1.6 to 0.6 seconds. The homepage went from about 1.4 MB to 0.44 MB. The about page went from 3.1 MB to 0.46 MB. Mobile first paint on the homepage went from 5.2 seconds to 1.2 seconds. The mobile performance score moved from 36 to 96.

One thing to be clear about: part of that score jump comes from how the CRM tracking script loads. It now waits for a tap, a scroll, or a few idle seconds before it fires. A lab test does not see it, but real visitors still download it. I'll publish 30-day field data from real phones once there's enough to show.

We tell the full story in our TeamUp case study.

This was a Webflow site, not a Framer site. The method is the same either way.

I have also shipped custom marketing surfaces from Ghost (churnkey.co/blog) and WordPress (suggestionox.com). That is platform range, not a performance brag on those domains.

Still choosing a builder? See Framer vs Webflow. Still deciding whether to leave Framer for code? See Framer vs custom.

Scoped on a discovery call. See how we work

Who this is for (and who should stay in the builder)

This is for Series A to growth SaaS teams, and specialist B2B firms that sell like SaaS, when the marketing site argues with the product. Two design systems. Experiments that are just another script tag. Editors who can ship a page but cannot own performance, redirects, or SEO. A CMS shape that breaks on calculators, portals, or localized routes. Or a stack the product team can actually call when marketing needs to share components or auth.

Say it plainly: this is not for teams that only need a prettier marketing page and are happy in the builder. If you are still in that decision, start with Framer vs Webflow for SaaS marketing sites (who edits after launch) or Framer vs custom development (builder vs Next.js). Moving between Framer and Webflow instead? Start here. If you have already decided to leave, this page is the measured cutover offer.

Rebuild vs measured migration

Typical "rebuild"Measured migration (this engagement)
New URLs when convenient

URL and canonical plan first; keep URLs that already earn traffic unless there is a reason not to

Analytics wired at the endBaseline plus events at QA, then a 30-day diff on what you already track
Builder JS patterns copied into ReactCritical path owned; fonts and LCP images fixed; tags re-justified
Fresh marketing design system by defaultLand in the same system as the product when that is the point
Launch screenshot

Same lab fields compared: LCP, CLS, JS weight, TTFB; GSC and field CWV when connected

Why not just vibe-code it yourself

A first deploy can look right in a weekend. The hard part is everything that has to stay true after cutover. We are not anti-AI. We are pro-ownership of the boring jobs that decide whether marketing still works in month two.

Deploys. Preview is not production. Env vars, DNS, and rollback should not mean hoping the old host is still up.

Analytics. Events that still fire after cutover. Consent behavior. A rebuilt tag plan, not a pasted GTM container.

Page speed. LCP image, fonts, third parties, Core Web Vitals. Local "feels fast" is not the mobile lab.

Forms. Validation, spam, CRM handoff, thank-you URLs that should not rank.

Security. Headers, dependency updates, secrets out of the repo, gated tools not left open.

SEO continuity. Apex vs www decided once. One canonical. Redirect map. Sitemap. Real schema. Keep URLs that already have impressions.

Content ops after launch. Who edits a page, what is noindexed, what breaks when a collection changes. After cutover to Next.js or Astro, many teams let marketing ship through AI agents and pull requests instead of reopening a CMS.

Astro vs Next.js — how we choose the stack

Same engagement either way. The stack is a recommendation, not the product.

Choose Astro when the site is content-first, islands are enough, marketing stays separate from the product app, and Core Web Vitals on organic pages are the brief.

Choose Next.js when marketing needs to share product UI, auth, design-system packages, middleware, or app-like tools (calculators, gated flows, experiments) with the product.

The usual SERP advice ("content site → Astro, app-adjacent → Next.js") is directionally right. Our addition: we pick after the baseline and sitemap, not before a stakeholder workshop picks a logo font.

Start with a diagnostic Sprint

The first step is a two-week diagnostic Sprint. I run the baseline on the live site, write the redirect map, and pick the stack (Astro or Next.js). Follow-on Sprints cover the build itself. Most sites need one or two, depending on how many templates are in play. After the cutover, ongoing work runs on the Standard subscription.

What the engagement includes

  1. Discovery + baseline — Sitemap, templates, forms, third parties, redirect intent. Lab performance on the homepage (mobile and desktop) and two to three real templates, same method every time.

  2. Design / system carryover — What stays pixel-close, what moves into the product design system, what gets deleted. CMS collection inventory, not a clone of every builder interaction.

  3. Custom rebuild (Astro or Next.js) — Pages that actually rank or convert. Shared layout. Fonts and images done properly: no lazy-loaded LCP image, no builder JS on the critical path.

  4. Content / CMS migration — Map collections to MDX, a headless CMS, or both. Keep URLs that have impressions unless there is a deliberate reason not to.

  5. Redirects — Apex vs www once. One redirect map. Thank-you and thin archives get deliberate index or noindex.

  6. QA — Forms, analytics events, consent, 404s, canonicals, schema as valid JSON-LD. GTM rebuilt, not pasted.

  7. Cutover — DNS, cache, first-hour monitoring. Old host as rollback only, not a second canonical.

  8. 30-day post-live comparison — Same URLs, same lab method, same fields, plus Search Console and analytics once access exists. A short written diff, including anything that got worse.

How we measure (the proof, not a promise)

Before cutover (public): lab LCP, CLS, FCP, Speed Index, TTI, performance score; mobile and desktop home; mobile on main templates; TBT as a lab INP proxy when field INP is missing; JS weight, transfer, and requests; document TTFB and cache headers; third-party list.

After go-live: identical URLs and method; both JSON runs kept on file.

With access (ask day one, do not block the build): Search Console clicks, impressions, indexing, field CWV; analytics conversion events you already care about.

We publish numbers only from those runs. If a tool fails, we say which tool and what replaced it.

Proof set (what we can say today)

PropertyFromStatusWhat we cite
churnkey.co/blogGhostDone (custom)Platform range — Ghost off the list
suggestionox.comWordPressDone (custom)Platform range — WordPress off the list
hireteamup.comWebflowLive Oct 19, 2026 (Next.js)

Same-method before/after: LCP, page weight, requests; what a lab test can't see

Hire Team Up is the first before/after on this page. We publish the numbers it actually produced, including the parts a lab test can't see, and add the 30-day field data once Search Console has enough of it.

Package shapes (and how pricing works)

We do not publish fixed dollar amounts for a one-time migration package on this page. Shapes only:

Standard rebuild — Known template set (home, a few landers, blog/resources, legal, forms). Baseline → rebuild → content move → redirects → QA → cutover → 30-day lab comparison. Design carried over, not a new brand.

Advanced — Standard plus real design-system handoff (tokens and components the product team can import) and portal or app hooks: logged-in entry, a calculator or tool that should not be a builder embed, experiments, content synced with the product. Still not a full product build.

After either shape, many teams want the site to keep moving the month the project ends. That is a design-build subscription or broader product design subscription beside engineering — the design partner for SaaS model. Clearly's public monthly rates (Standard scoped on a discovery call · Advanced scoped on a discovery call) apply to that ongoing delivery; confirm inclusions on pricing. See the pricing page to learn how the ongoing partnership scoped on a discovery call.

When you should not leave yet

Stay in Webflow, Framer, WordPress, or Ghost when the site is small, launch speed beats deep control, marketing needs day-to-day visual editing, and SEO and ops requirements are still straightforward. Framer vs Webflow and Framer vs custom cover stay-or-leave while you are still in the builder. CMS choice and migration are related but different jobs; we are not rebuilding a full CMS chooser essay here.

Still deciding whether to leave Framer for code? Read Framer vs custom development before you book. Ready to baseline the live site? Book a discovery call from the homepage, or review pricing for how the ongoing partnership scoped on a discovery call.

Frequently asked questions

How long does a Framer to Next.js migration take?
It starts with a diagnostic Sprint: two weeks to baseline the live site, map the redirects, and settle on the stack. The migration itself usually runs in one or two follow-on Sprints, depending on template count, CMS collections, and how involved the redirect map is. There is no fixed project price. I scope the work after the baseline, so you know what you're committing to before the build starts.
What does a measured migration off Webflow, Framer, WordPress, or Ghost include?
Leaving Webflow, Framer, WordPress, or Ghost for custom Next.js or Astro means baseline the live site first (lab LCP, CLS, and transfer on the real templates), rebuild the pages that rank or convert, keep the URLs and events that matter, cut over with one redirect map, then run the same lab method after go-live and write the diff — including anything that got worse. Discovery covers forms, third parties, and redirect intent. QA covers analytics, consent, canonicals, and schema. The offer is the method, not a promise of invented before/after percentages.
How is a measured migration different from a pretty rebuild or a vibe-coded Next.js site?
A pretty rebuild often ships new URLs because they are convenient, pastes the old GTM container, and calls launch done when the homepage screenshot looks right. A vibe-coded first deploy can look correct locally and still miss deploy rollback, mobile LCP, form handoffs, security headers, and SEO continuity. A measured migration decides apex vs www once, rebuilds the tag plan, owns the critical path without builder JS, and compares the same lab fields before and after cutover.
When should we choose Astro vs Next.js for the new marketing site?
Choose Astro when the marketing site is content-first, islands are enough, and the site stays separate from the product app — organic pages and Core Web Vitals are the brief. Choose Next.js when marketing must share UI, auth, design-system packages, middleware, or app-like tools (calculators, gated flows, experiments) with the product. Clearly picks the stack after the baseline and sitemap, not before. The engagement shape is the same either way.
How do you keep SEO, analytics, forms, and Core Web Vitals intact through cutover?
SEO: one canonical host, a written redirect map, sitemap, and JSON-LD validated before DNS moves. Analytics: baseline events at QA, consent behavior checked, GTM rebuilt rather than pasted. Forms: validation, spam handling, CRM handoff, and thank-you URLs that should not rank. CWV: fix LCP images, fonts, and third parties on the templates that matter, then re-run the same lab URLs after go-live. Search Console and field data enter the 30-day comparison when you grant access; they do not block the build.
What proof does Clearly have today, and what do you not claim yet?
I have shipped custom marketing sites from Ghost (churnkey.co/blog) and WordPress (suggestionox.com) — platform range, not performance claims on those properties. Hire Team Up moved from Webflow to Next.js in October 2026; the before/after numbers and the lab-vs-field caveat are in the proof section on this page. I do not publish invented migration package prices, customer counts, or before/after case-study deltas.
When should you not leave Webflow, Framer, WordPress, or Ghost yet?
Stay when the site is small, launch speed matters more than deep control, marketing needs day-to-day visual editing, and SEO and ops requirements are still straightforward. If you are still choosing between builders, read Framer vs Webflow at https://clearly.design/resources/webflow-vs-framer-saas-marketing-site. If you are deciding whether to leave Framer for code at all, read https://clearly.design/resources/framer-vs-custom-development-saas. CMS choice and migration are related jobs; this page is the leave path once you have decided.

Related Decision Guides

Framer to Webflow or Webflow to Framer? Migrate or Skip Both

Switching builders is a full rebuild. Here's what moves, what gets rebuilt, how to keep your rankings, and when a repo site beats a second builder.

Framer vs Webflow for SaaS marketing sites: who edits after launch

Framer vs Webflow for a SaaS marketing site: Framer if design owns edits, Webflow if marketing ops and the CMS do, custom when it must match product.

Framer or Custom for SaaS? Pick Your Marketing Site Path (2026)

Framer vs custom development for SaaS marketing sites: when the builder wins, when Next.js + shared tokens win, and how Clearly helps after the choice.

Design-Only vs Design+Build Subscription for SaaS (2026)

Design-only vs design+build subscription: when prototypes are enough, and when your SaaS needs Framer, Webflow, or React shipped in the same retainer.

Design Partner for SaaS Startups

Looking for a design partner for your SaaS, not a request queue? Prototype-first embedded design, from a 2-week Sprint to monthly plans.

Product Design Subscription for B2B SaaS: Ownership, Not a Queue

A product design subscription for B2B SaaS should own whole features from rough intent to developer-ready flows, not just clear a request queue.

Keep Your Marketing Site Consistent with Product UI

Keep your marketing site consistent with product UI through a shared design system. Tokens and invariants stay shared, while density and scale can diverge.

GEO Retainer Cost and Structure for B2B SaaS

GEO retainers bundle site work and off-site citations on one invoice. See what they cost, what moves the price, and how to separate the two jobs.

Ready to leave the builder without guessing at performance?

I'll walk your sitemap, baseline the live site, and tell you honestly whether Astro or Next.js fits — or whether you should stay put another quarter. See the pricing page to learn how the Standard plan keeps your site moving after launch.