Ecommerce Data Migration Cost in 2026: Tools, DIY and Custom Quotes Compared
What ecommerce data migration costs: real tool prices, the true hours in a DIY CSV move, seven drivers of a custom quote, and the costs nobody lists.
Long Nguyen
Lập trình viên Fullstack · Kỹ sư AI · Nhà nghiên cứu
What ecommerce data migration actually costs
The honest answer is that the same catalog can cost nothing or five figures to move, and the deciding variable is almost never how much data you have. It is how much of your data has no native home on the destination platform.
Ten thousand plain products with two variants each is a cheap migration. Eight hundred products with customer-specific pricing, three custom fields, and five years of order history is an expensive one. Vendors who quote from your SKU count alone are quoting the wrong number.
There are four cost models, and they are not tiers of the same thing — they are different work:
| Approach | What you actually pay | Fits when |
|---|---|---|
| DIY CSV export and import | No licence fee; your hours instead | Small catalog, no order history, one currency, no custom fields |
| Self-serve migration app | A one-time or per-period fee, scaled by record counts | Both platforms are supported and your data is standard-shaped |
| Tool vendor runs it for you | Tool fee plus a service upgrade | Same data profile, but you do not want to own the process |
| Custom migration by a developer | Engineer-days multiplied by a day rate | Custom fields, B2B pricing, subscriptions, ERP sync, unsupported source |
Before you compare prices, work out which column you are in. Paying agency rates for a job a $50 app does is waste; buying a $59 app for a store with tiered B2B pricing buys you a partial import and a month of manual repair.
What the migration tools actually charge
Published tool pricing is the only part of a migration budget with firm numbers in it, so start there.
Matrixify, the import/export app used on Shopify, publishes flat plans billed per 30-day period, with per-job record limits rather than a monthly quota:
| Plan | Price | Products per job | Customers | Orders | Redirects |
|---|---|---|---|---|---|
| Demo | Free | 10 | 10 | 10 | 10 |
| Basic | $20 / 30 days | 5,000 | 2,000 | 1,000 | 10,000 |
| Big | $50 / 30 days | 50,000 | 20,000 | 10,000 | 100,000 |
| Enterprise | $200 / 30 days | Unlimited | Unlimited | Unlimited | Unlimited |
LitExtension prices from $59 and scales with the number of entities you move — products, customers, orders and blog posts counted together — with a self-service tier and an assisted tier where their team runs the migration. Cart2Cart publishes no floor price at all; its cost comes out of an on-page estimator fed by the same entity counts.
Three things about entity pricing catch people out, and they are worth more than the headline number:
- Orders dominate the count, not products. A five-year-old store with 1,200 SKUs may carry 40,000 orders. Budgeting from the product count understates the quote by an order of magnitude. Decide how far back you actually need order history before you ask for a price — capping it at 24 months is the single cheapest decision in most migrations.
- Per-job limits are not per-month limits. Matrixify's tiers cap rows in one import run. A store that needs five passes — test, fix, re-run — stays on one plan, but a store whose single file exceeds a tier's row cap jumps a tier even for one run.
- The advertised price is the floor, and add-ons sit above it. Customer password migration, SEO URL migration and 301 redirect generation are commonly priced as options rather than included. Those are exactly the items you cannot skip.
Is a DIY CSV migration free?
The licence is free. The hours are not, and the hours are where DIY migrations go over budget.
The work is not the import click. It is everything around it:
- Field mapping. The destination's CSV schema will not match your source's column names. Every column is a decision, and the ones with no obvious target are the ones that become metafields or get silently dropped.
- Variant and option restructuring. Source platforms model options differently. Flattening or expanding them into the destination's shape is spreadsheet surgery, not a find-and-replace.
- Image handling. Most importers fetch images from URLs at import time, so your old images have to stay publicly reachable throughout. That constraint alone can block an early DNS cutover.
- The error loop. A 50,000-row file that fails on row 48,213 means fixing one cell and re-running the whole file. Three or four of these loops is normal, and each one costs the full import runtime.
- Validation. Counting rows is not validation. Spot-checking prices, tax settings, inventory and descriptions against the source is.
For planning, assume a day or two for a tidy catalog under 1,000 SKUs with no order history, and a week or more once you pass a few thousand products or need order data. Price that against your own hourly rate and the comparison with a $50 app usually answers itself.
The DIY cost that nobody counts is the silent one. An import that reports success can still have truncated long descriptions, dropped tags, lost SEO titles or collapsed metafields. You will find those weeks later, from a customer or from a traffic drop, and fixing them then costs more than doing it right would have.
Seven things that move a custom migration quote
When data does not fit the destination's standard shape, cost stops tracking volume and starts tracking exceptions. These are the seven that show up in nearly every scope conversation:
| Driver | Why it costs | Cheap signal / expensive signal |
|---|---|---|
| Order history depth | Orders carry line items, refunds, taxes, fulfilments and customer links; most platforms accept them only through the API, not CSV | No history needed / full history with refunds and partial fulfilments |
| Variants and options | Destination platforms cap options per product and variants per product; anything above the cap must be re-modelled as separate products | Under three options / configurable products with dozens of combinations |
| Custom fields | Every non-standard attribute needs a target definition, a type, and a template change to display it | None / dozens, used in filters and templates |
| B2B pricing | Customer-group pricing, tiered quantity breaks and net terms rarely map one-to-one; usually rebuilt, not moved | One public price list / per-customer price lists |
| Subscriptions | Recurring billing tokens are tied to the old payment gateway and generally cannot transfer; subscribers are re-created and often re-authorised | None / an active subscriber base |
| Multi-store, currency, language | Each locale multiplies the content to map and the URL set to redirect | One store, one currency / several storefronts and translated catalogs |
| URL structure difference | Destination platforms impose fixed URL patterns, so the redirect map is one row per indexed URL, not a wildcard rule | A few hundred URLs / tens of thousands plus paginated and filtered URLs |
Read that table as a quoting tool. If you can answer all seven with the cheap signal, you are a tool customer. If three or more land on the expensive side, a tool will get you a partial result and you will pay a developer to finish it anyway — twice.
When a tool is cheaper — and when it stops being cheaper
We sell custom migrations, so take this with that in mind: for a standard catalog moving between two supported platforms, a $20 to $200 app is the right answer and no agency should talk you out of it. Paying for engineering time to do what an app does well is money set on fire.
The crossover is not a store size. It is the first piece of data with no destination field:
| Use a tool when | Scope a custom migration when |
|---|---|
| Both platforms appear on the tool's supported list | Your source is a custom build, an ERP, or a platform no tool supports |
| Products, customers and orders cover your data | Custom fields, B2B price lists or subscriptions are in scope |
| You can accept the destination's native data model | Your catalog has to be re-modelled to fit it |
| A partial result is recoverable by hand | A partial result means weeks of manual repair |
The expensive mistake is running a tool first on complex data, accepting the partial result, and hiring a developer afterwards to reconcile two half-correct datasets. Reconciling is more work than migrating cleanly once. If your answers in the previous section leaned expensive, get the migration scoped properly before you buy a tool licence.
Estimate your own migration before you ask anyone for a price
Every quote you receive will be a guess until you hand over the counts. Pull these ten numbers from your current store first — it takes an hour and it is the difference between comparable quotes and theatre:
- Product count, and variant count (they are different numbers, and the second one prices the job)
- Maximum options on any single product
- Customer count
- Order count, plus the date range you genuinely need — decide this before asking
- A list of every custom field or attribute, and whether anything depends on it
- Indexed URL count, taken from your sitemap, not your page count
- Currencies, languages and storefronts
- Every app or extension in use, and whether the destination has an equivalent
- Live integrations: ERP, accounting, 3PL, marketplace feeds
- Whether subscriptions or B2B price lists exist at all
Send that same list to every vendor. Anyone who quotes without asking for it is pricing a different project from the one you have, and the gap turns up as a change order later.
If you would rather have the counts read for you and a scoped figure back, tell us what you are moving and from where and we will tell you whether you need us or a $50 app.
CÂU HỎI THƯỜNG GẶP
Câu hỏi thường gặp
How much does it cost to migrate an ecommerce store?
It depends on which of four cost models you fall into, not on how big your catalog is. A DIY CSV move costs only your hours. A self-serve migration app costs roughly $20 to $200 depending on record counts. A tool vendor running the migration for you costs the tool price plus a service upgrade. A custom migration is priced as engineer-days multiplied by a day rate, and lands there whenever your data does not fit the destination platform's standard shape.
Why do migration tools quote more than I expected?
Because entity-priced tools count orders, customers and blog posts alongside products, and in an established store the order count is usually far larger than the product count. A store with 1,200 products and five years of trading can carry tens of thousands of orders. Deciding how far back your order history genuinely needs to go is normally the cheapest cost reduction available.
Is a free CSV migration really free?
The licence is, the hours are not. The cost sits in field mapping, variant restructuring, image URL handling, repeated import error loops on large files, and validation. Plan on a day or two for a tidy catalog under 1,000 SKUs with no order history, and a week or more beyond a few thousand products. The expensive part is the silent failures: fields that are truncated or dropped by an import that still reports success.
What makes a migration expensive rather than cheap?
Exceptions, not volume. Deep order history, products above the destination's variant or option caps, custom fields, B2B customer-specific pricing, active subscriptions, multiple currencies or storefronts, and a large redirect map. Each one requires work that has no equivalent in a standard import, so cost tracks the number of exceptions rather than the number of rows.
What costs are usually missing from a migration quote?
Four of them. The organic traffic dip while search engines recrawl the new URLs, which can run for weeks or longer on a large site. App or extension subscriptions on the new platform that replace functionality you previously owned, which is a permanent recurring cost. Running both stores in parallel for the overlap period. And the fixes that surface in the first weeks after cutover, once real customers hit the new store.
Should I use a migration tool or hire a developer?
Use a tool when both platforms are on its supported list, your data is products, customers and orders, and you can accept the destination's native data model. Scope a custom migration when your source is a custom build or unsupported platform, or when custom fields, B2B price lists or subscriptions are in scope. The expensive mistake is running a tool on complex data, accepting the partial result, and then paying someone to reconcile two half-correct datasets.
How long should I keep 301 redirects after migrating?
At least a year, and longer where you can. Google's site move guidance states that permanent redirects do not cost you PageRank, but every old and new URL has to be crawled at least once before the move is treated as complete, which takes weeks or more on larger sites. Each old URL should point straight to its final destination rather than through a chain or to a catch-all home page.