Hire a CMS migration service when the cost of the site being wrong is larger than the cost of the quote, and do it yourself when it is not. That is the whole decision, and it turns on three things: how many posts you have, how much of your content is platform-specific rather than plain text, and whether anyone in your operation can spend two uninterrupted weeks on it. A small archive of clean posts moving between two mainstream platforms is genuinely a do-it-yourself job. A large archive full of custom fields, shortcodes and one-off layouts is not, regardless of budget.
Whichever way you decide, the backup rule does not change. Take a full backup of the existing site and verify that you can restore it before anyone starts, including an agency. Ask a prospective vendor what their rollback plan is and what state your old site will be left in, and treat a vague answer as disqualifying. See the backup and rollback plan for what “verified” means here.
This site does not sell migration services and takes no referral fee for sending you to one. That matters for this particular question, because almost nothing else you will read about it is written by a party without a stake.
Why you cannot trust the prices you find
Search for what a CMS migration costs and you will get numbers. Read the bylines on the pages carrying those numbers and a pattern appears immediately: they are published by migration agencies, by hosting companies whose migration tooling is part of their sales pitch, and by CMS vendors who want you on their platform. The results for this query are close to unanimous on that.
Those pages are not lying. They are anchoring. A cost guide published by a company that sells the work exists to set your expectation before the quote arrives, and a range wide enough to cover everything from a small blog to an enterprise replatform tells you almost nothing about your own job.
So this article is not going to add another set of numbers to that pile. Prices move, they vary enormously by region and by who you ask, and a figure repeated here without a source you can check is worth less than nothing. What is stable, and what nobody selling the work has much incentive to give you, is the list of variables that set the price. Fill those in and you can read any quote critically, which is more useful than knowing what a stranger charged someone else.
The variables that set the price
Work through these before you talk to anyone. Each one moves the cost, and the last two move it most.
| Variable | Cheap end | Expensive end |
|---|---|---|
| Volume | Under a hundred posts | Thousands, especially with an inconsistent history |
| Platform pair | Two mainstream platforms with a documented path between them | An obscure, custom or end-of-life source system |
| Content uniformity | Plain posts, one layout, one author | Custom fields, shortcodes, page builders, mixed post types |
| Media volume | A featured image per post | Large libraries, galleries, video, documents |
| Structural change | Same shape of content on both sides | A new content model that the archive cannot populate |
| URL preservation | New site, no history to protect | An indexed archive whose URLs must survive |
| Integrations | None | Memberships, paid subscriptions, forms, e-commerce, email |
| Design | Content only | A redesign happening at the same time |
| Deadline | Whenever it is right | A fixed date somebody has already announced |
The last row deserves particular attention. Urgency is the single easiest way to make a migration expensive, and it is almost always self-imposed.
Two of these rows also decide whether the job is DIY-able at all. Content uniformity is the real difficulty gradient: a thousand plain posts are easier to move than a hundred posts built with a page builder, because the plain ones are the same problem a thousand times and the others are a hundred different problems. And integrations are where migrations stop being about content. If people pay you money through the site, the risk profile changes entirely and the case for hiring gets much stronger.
What you are actually buying
It is not skill you could never acquire. The mechanics of most migrations are documented and learnable, and much of the work is unglamorous checking.
What a competent migration service sells is three things:
Calendar time. Somebody else’s two weeks instead of yours. If your own two weeks are worth more than the quote, that alone can settle it.
Risk transfer, partially. They have hit the failure modes before and recognize them early. This is real value, and it is also the most oversold part of the pitch, because the consequences of a bad migration still land on you and your readers, not on the vendor.
A finish date. Self-run migrations stall. They get to eighty percent, the site goes half-live, and the remaining twenty percent sits for months. A contract with a deadline is often what gets that last stretch done.
What you are not buying, and this is the important part
Nobody can make the content decisions for you. This is the cost that surprises people who hired specifically to avoid work.
A migration is full of judgment calls that only the owner can make. Which of your three overlapping categories survives. Whether the 2019 series that nobody reads comes across at all. What the new required fields should contain for old posts. Which of the twelve shortcode formats in your archive matter and which can become plain text. Who the author is on the forty posts published under a shared account.
An agency will ask you all of these and wait. Every hour of waiting is on your invoice if you are being billed for time, and on your timeline regardless. Budget your own hours for the reconciliation work even in a fully paid migration, because the transfer is the cheap part and the decisions are not.
There is also a fourth thing you are not buying: certainty that the result is correct. Verification is yours. You are the only person who knows what the site was supposed to contain, which is why the post-migration audit is your job whoever ran the move.
A test that answers this in an afternoon
Rather than agonizing, run the experiment. Take twenty posts and migrate them yourself, choosing the twenty deliberately: your oldest post, your longest post, the one with a gallery, the one with an embed, the one with a table, the one with three authors, the one you have edited eleven times.
If those twenty come across recognizably in an afternoon, your archive is uniform and you can do this. Multiply the time by your post count, add half again, and that is roughly your DIY cost in hours.
If those twenty produce a mess you cannot diagnose, you have learned something far more valuable than a price range: you have learned which specific things break in your content. Take that list to a quote conversation and you will get comparable, honest answers instead of a number invented from an assumption about a typical site.
That sample list, plus the variable table above, is also the closest thing to a request document you can send several vendors at once. It is what makes quotes comparable, and comparable quotes are the only kind worth collecting.
Questions to ask before you sign anything
Independent of who you hire, and without naming or recommending anyone:
- What exactly is in scope? Posts, pages, media, comments, users, taxonomies, redirects, forms. Get the exclusions in writing, because the exclusions are the argument later.
- What is the rollback plan, and how long does the old site stay running? The correct answer is weeks, not hours, and it should be their suggestion rather than yours.
- Will you do a sample migration first, and can I inspect it? A vendor who declines this is telling you something.
- Who verifies completeness, and against what? If the answer is “you”, that is fine and normal, but it needs to be said out loud before the invoice.
- What happens to content that cannot be converted automatically? Silently dropped, flagged, or left as-is are three very different answers.
- What is the change process if we find problems after handover? Included for a period, billed hourly, or finished is finished.
Anyone who has actually done this work will answer all six quickly. The answers also tell you whether they have read what breaks in every CMS migration or are quoting from a template.
The honest summary
Most small publishers moving a straightforward archive between mainstream platforms can do this themselves, and the main thing standing in the way is uninterrupted time rather than expertise. Most operations with paid memberships, heavy custom formatting, or an archive whose URLs carry real traffic should get help, and should get it before making the announcement rather than after.
And if the migration fails halfway either way, the procedure is the same, and it is worth knowing in advance: stop the import and roll back the destination rather than pushing forward.
FAQ
How much does a CMS migration cost?
It depends on variables that are specific to your archive: post count, content uniformity, media volume, integrations, whether URLs must be preserved, and whether a redesign is bundled in. Published ranges are almost all written by companies selling migration work, so use the variable list above to price your own job rather than adopting someone else’s number.
Is a CMS migration hard enough to need a professional?
The mechanics are learnable and well documented. The difficulty comes from content that is not uniform, from integrations that handle money, and from archives large enough that manual repair is impractical. Uniformity matters more than size.
Can I do part of it myself and hire out the rest?
Yes, and it is often the best value. Content decisions, cleanup and verification are yours anyway. Paying for the technical transfer and the failure diagnosis while doing the editorial reconciliation in-house splits the work along its natural seam.
What should I have ready before asking for a quote?
A post and page count, a media count, a list of integrations, a note of whether URLs must be preserved, and the results of a twenty-post test migration. That package gets you comparable quotes instead of a wide range.
Does hiring someone remove my risk?
Only partly. The consequences of a bad migration land on your site and your readers regardless of who performed it, so a verified backup, a real rollback plan and your own post-migration audit stay necessary either way.
Written July 2026. No price, quote or job duration is stated here because none could be verified against a source independent of the companies selling migration work. Get quotes for your own job instead.
