The Exit Test: Could You Leave Your Platform Tomorrow?

Lock-in is a measurement, not an insult

“Vendor lock-in” gets used as an accusation, usually by someone selling the alternative. It’s more useful as a number.

Everyone has some. A WordPress owner who can’t read their own database is locked in. A Shopify merchant who can’t take their checkout is locked in. Someone with a bespoke site and one freelancer who knows how it works is extremely locked in, and usually doesn’t realise because nobody sold them a subscription.

The question isn’t whether you’re locked in. It’s whether you know your number, and whether it’s proportionate to what the site is worth.

Here’s how to find out. Ten checks, three minutes each, no technical knowledge required. Score each 0, 1 or 2 and total out of twenty.

The ten checks

1. Content

Test: try to export your blog posts, case studies or products right now, with their fields intact, in a format someone else could import.

2 — you get structured data (CSV, JSON, an API) with every field.
1 — you get some of it, or a static snapshot without the structure.
0 — you can’t, or you’d be copying and pasting.

Cheap fix: schedule a monthly export to storage you control. Most platforms have an API or a CSV export. This costs an hour to set up once.

Worth knowing: Webflow’s own engineering post concedes that while you can export your code, you couldn’t edit that export and bring it back. The export exists; the round trip doesn’t. Most platforms are the same and fewer are honest about it.

2. Design and structure

Test: if you moved, would any of the layout come with you in a usable form?

2 — you have components, a design system, or code you could hand to someone.
1 — you’d have visual reference but would rebuild.
0 — you’d start from a screenshot.

Realistically most people score 1 here and it’s fine. Rebuilding a front end is the cheapest part of moving now.

3. Hosting and DNS

Test: log in to your domain registrar. Right now.

2 — you know where it is, it’s in the business’s name, you could repoint it today.
1 — you have access but someone else set it up.
0 — you don’t know who holds it, or it’s in an agency’s account.

Cheap fix: this is the single highest-value hour you’ll spend. Get the registrar and DNS into an account the business owns, with billing on a company card. Nothing else on this list matters if someone else controls where your domain points.

4. URLs and redirects

Test: do you have a current list of every URL on your site?

2 — yes, maintained.
1 — you could generate one from a sitemap.
0 — no idea how many pages you have.

This is the item that turns a routine migration into a disaster. Every URL that changes without a redirect is a page that disappears from search and every link pointing at it that breaks. It’s also the cheapest thing on the list to sort out.

5. Form submissions

Test: can you retrieve every enquiry you’ve received in the last two years?

2 — they go to your CRM or inbox and you hold the history.
1 — they’re in the platform and exportable.
0 — they’re in the platform and you’ve never checked.

Cheap fix: route form submissions to somewhere you own as well as wherever they currently go.

6. Members and logins

Test: if you have gated content, who holds the user identities?

2 — a dedicated provider, or your own database.
1 — a third-party tool bolted to the site.
0 — a feature bundled into your website platform.

If you scored 0, note that this is exactly the position Webflow’s User Accounts customers were in before it was switched off on 29 January 2026. Not a hypothetical.

7. Automations

Test: what happens to your workflows if the tool running them stops?

2 — they’re in a dedicated automation platform you pay for separately.
1 — a mix.
0 — they’re built into your website platform.

Webflow’s Logic was disabled on 27 June 2025. Flows stopped. Forms connected to them reverted to ordinary forms.

8. Analytics history

Test: could you get three years of traffic data out?

2 — you run something independent and own the data.
1 — you have a platform-native tool and an independent one.
0 — platform-native only.

Native analytics almost never export. You lose your baseline exactly when you most need it — right after a migration, when you’re trying to work out whether it went well.

9. Access and bus factor

Test: how many people could publish a change to your site this afternoon?

2 — two or more, at least one on your payroll.
1 — one person, on your payroll.
0 — one person, not on your payroll.

Be honest here. This is the most common form of real lock-in I see, it has nothing to do with which platform you chose, and people who worry loudly about vendors often score zero on it.

10. Commercial

Test: what would it cost to stop paying next month?

2 — monthly, no notice, no seat commitment.
1 — annual billing you’d ride out.
0 — a contract with notice, or a seat count you’re locked into.

Scoring

16–20 — Portable. You could move if you needed to. You probably don’t need to. File this and get on with your week.

10–15 — Harden it. You have two or three specific dependencies worth removing. Not a migration — a Tuesday afternoon. Do those and re-score.

Under 10 — You have a tenancy, not a website. Worth taking seriously, though see the last section before you spend anything.

The three fixes that move the score most

If you do nothing else:

1. Registrar and DNS into a company-owned account. One hour. Moves check 3 and underwrites everything else.
2. A scheduled content export to storage you control. One hour to set up, then automatic. Moves check 1.
3. A written URL map and a one-page document listing your stack, logins and monthly costs. Two hours. Moves checks 4 and 9, and is the thing you’ll be most grateful for in a crisis.

Four hours, roughly, and a typical score moves five or six points. No migration, no new platform, no invoice.

What leaving actually costs

If you do decide to move, the cost sits in places people don’t expect. Design is cheap now. Content mapping and redirects are not, and they’re where migrations go wrong.

Rough bands, and treat these as orientation rather than quotes:

• A small brochure site — a few days. The kind of thing where rebuilding is genuinely faster than migrating.
• A content site with a hundred-plus pages — weeks, and most of that is content mapping, redirects and checking nothing dropped out of the index.
• A site with members, integrations or transactions — months, and the price reflects it. Specialist agencies publish bands for this work; one European firm quotes €15,000–€30,000 for a Webflow-to-headless engagement, which is the top of the market rather than the middle, but it tells you the shape.

Most SME sites sit well below any of those numbers. Get a quote before you assume.

When a low score is fine

I’d rather say this plainly than have you spend money on my account.

If your website generates no revenue, costs £30 a month, and would take a fortnight to rebuild, then portability is a rounding error. A score of six is completely rational. Don’t spend £5,000 insuring against a £2,000 risk — that’s just a worse version of the problem.

The score matters in proportion to what the site does. A site that carries your entire enquiry pipeline, holds customer data, or has ten years of accumulated search authority deserves a serious answer. A five-page brochure doesn’t.

Work out which one you’ve got first.

→ Back to Rented Ground
→ Next: Builder, Headless, or Bespoke