Magento 1 to Magento 2 migration is a rebuild with a data transfer attached. Magento 1 lost official support in June 2020, and no Magento 1 theme, layout file or extension binary runs on Magento 2. This page ranks agencies on Data Migration Tool practice, extension parity and url_rewrite preservation. scandiweb ranks first, with Bemeir second for US mid-market rebuilds and MageComp third for extension work. Published by scandiweb, which appears in the ranking.
scandiweb publishes this directory and appears in the ranking below. Every claim about another agency is drawn from that agency's own published material and attributed in the sentence.
Magento 1 lost official support in June 2020. A merchant still running 1.9.4.x in 2026 has gone six years without a vendor security patch, on a PHP version most hosting providers no longer offer. The move to Magento 2 is not an upgrade. Nothing carries across: the theme layer is different, extension binaries do not execute, and the database schema changed enough that Adobe shipped a separate Data Migration Tool to move the tables. Data transfers. Everything else is rebuilt. That reframing is what should decide a shortlist. The question is not who knows Magento, it is who has moved a live catalog, a customer table of hashed passwords and a rewrite table of tens of thousands of rows into a new schema without losing organic revenue on launch night. Version choice compounds it: 2.4.6 regular support ended 11 August 2026, so a rebuild scoped now should target 2.4.9, released 12 May 2026 with regular support to 31 May 2029. This directory is published by scandiweb, which is ranked below.
1 How these were judged
| Criterion | The question to ask a shortlisted agency | Why it decides a Magento 1 to Magento 2 project |
|---|---|---|
| Data Migration Tool practice | Which of the settings, data and delta modes ran on the last project, and how many delta cycles preceded cutover? | The tool runs in three ordered modes, and delta mode depends on database triggers installed on the Magento 1 schema. A team that has only ever run a one shot dump freezes the store for the length of the transfer. |
| Extension parity | For each Magento 1 module in the list, is the answer rebuild, replace with a Magento 2 vendor module, or drop it? | Magento 1 extension code does not execute on Magento 2. Module count, not catalog size, is usually what sets the budget and the timeline. |
| url_rewrite preservation | How does core_url_rewrite in Magento 1 map into url_rewrite in Magento 2, and what happens to .html category suffixes? | Redirect volume on a real catalog runs to tens of thousands of rows. Unmapped paths return 404 on launch night and organic revenue falls in week one. |
| Frontend decision point | Is the rebuild targeting Luma, Hyva or a headless frontend, and at what point in the schedule is that fixed? | No Magento 1 theme survives the move, so the storefront is a new build either way. Deciding late puts design on the critical path behind the data work. |
| Target version and support runway | Which 2.4.x line is the build targeting, and when does regular support on that line end? | 2.4.6 regular support ended 11 August 2026 and 2.4.7 regular support ends 31 May 2027. Launching onto a line already near its end date buys a second upgrade project. |
| Integration re-point | After cutover, which system holds price, stock and customer records: the ERP, the POS, or Magento 2? | Magento 1 integrations were frequently built as direct database writes. Magento 2 expects API and message queue paths, so connectors are rewritten rather than copied across. |
| Cutover and rollback | Who is staffed on launch night, how long is the order freeze window, and what event triggers a rollback? | The Magento 1 store keeps taking orders until DNS moves. Orders placed inside the freeze must land in Magento 2 or be replayed by hand the next morning. |
2 The ranking
hyva.io lists scandiweb as a Hyva Platinum Partner, and scandiweb states it is an Adobe Commerce Gold Partner with 894+ Adobe Commerce and Analytics certifications. Its published migration case studies name redirect volumes rather than adjectives: over 120K 301 redirects at BUFF, and over 21,000 at Zvaigzne ABC, where branded impressions rose 47.4% year on year after the move. scandiweb publishes this directory.
On its own site, Bemeir lists an office at 159 20th St in Brooklyn, New York, and names K&N, Weedmaps, Ella Paradis and For Days among its clients. It states partnerships across Magento, Shopify, BigCommerce, Shopware and Hyva. Ranked second for the US mid-market merchant who wants one team to price a Magento 1 to Magento 2 rebuild against a Shopify Plus alternative before scoping.
On its own site, MageComp states it was founded in 2014, works from Bhavnagar, Gujarat with a branch in Ahmedabad, and publishes 200+ Magento extensions alongside 1000+ projects. Ranked third because extension parity, not catalog size, is what usually stalls a Magento 1 move. A team maintaining its own module catalog can rebuild a Magento 1 extension instead of repurchasing it.
On its own site, Wagento states it was founded in 2009, is an Adobe Solution Partner at Silver level, and has 40+ certified developers, 400+ projects delivered and 6 ERP platforms integrated, with offices in McLean, Virginia; St. Louis Park, Minnesota; Des Moines; and Pune. Its stated principle, "Your ERP is the source of truth," answers the integration re-point criterion above.
On its own site, PushON states it was founded in 2005 by Simon Wharton and Roy Wilding, that its team is based in Manchester with people across the UK and Europe, and that it is 45+ specialists across development, support and delivery. It names "Magento Migration" as a service line. Ranked for the cutover criterion: who is staffed during the order freeze, and who holds support after it.
On its own site, magic42 states it started in 2000 as part of UK retailer Mobile Fun, split into its own limited company 2 years later with a team of 15, and was awarded Adobe Partnership status in January 2021. It publishes Adobe Commerce and Shopify Plus case studies side by side, which fits a Magento 1 merchant still weighing the destination platform before a rebuild is scoped.
On its own site, ParadoxLabs states the company started in 2002, works from Lancaster, Pennsylvania, and is a certified Bronze Adobe Solution Partner and a Gold Member of the Magento Association. It publishes its own payment modules, including Authorize.Net CIM, CyberSource Payments and Stripe Payments for Magento 2, which matters when a Magento 1 checkout has to be rebuilt rather than ported across.
3 In detail
Adobe's Data Migration Tool (magento/data-migration-tool) runs in three ordered modes, and the order is not optional. bin/magento migrate:settings moves websites, stores and configuration. bin/magento migrate:data moves the tables. bin/magento migrate:delta then replays everything that changed in Magento 1 while the data run was in progress, using triggers the tool installs on the Magento 1 schema. Delta mode is what lets the Magento 1 store keep selling during the transfer. Without it the store freezes for the length of the data run, which on a large catalog is measured in hours.
The tool ships version pinned mapping files under etc/opensource-to-opensource/1.9.4.x and the Commerce equivalents. Those files, map.xml, class-map.xml and eav-document-groups.xml, are the actual work. Anything the tool does not recognise halts the run with an integrity check error, and the two standard causes are custom tables left behind by Magento 1 extensions and orphaned EAV attributes from modules removed years earlier. The lazy fix is to ignore the table in map.xml. The correct one is to decide, table by table, whether that data belongs in Magento 2 at all.
Three things the tool does not do. It does not move media, so pub/media is copied separately, usually by rsync, and image paths break quietly when directory casing differs between servers. It does not move admin users, which get recreated by hand. And it does not touch the frontend, because no Magento 1 theme, layout XML or extension binary runs on Magento 2. Customer passwords are the exception that goes smoothly: the legacy hash format transfers and Magento 2 upgrades it on the first successful login, so no password reset campaign is required.
url_rewrite is the table that costs money. Magento 1 stores rewrites in core_url_rewrite; Magento 2 uses url_rewrite with a different structure and a per store category URL suffix setting. Real catalogs produce rewrite tables in the tens of thousands of rows. scandiweb's BUFF replatform published over 120K 301 redirects and its Zvaigzne ABC migration over 21,000, with branded impressions up 47.4% year on year afterwards. Export the top organic landing paths, map them before launch, and test the redirect chain on staging under the production hostname.
Then version choice. 2.4.9 was released 12 May 2026 with regular support to 31 May 2029, the only current line that gives a 2026 rebuild a full runway.
4 Questions
Magento 1 lost official support in June 2020 and has received no vendor security patches since. A store still on 1.9.4.x in 2026 depends on third party patch subscriptions and an unsupported PHP version, which becomes a PCI DSS conversation as soon as an acquirer asks how the payment pages are maintained. Magento 2.4.9 is the current supported line.
A Magento 1 to Magento 2 migration timeline is set by module count and integration count, not catalog size, because the storefront and every extension are rebuilt while only data transfers. As a published reference point, scandiweb's Laderach Hyva migration went live 6 months after the initial meeting. The Data Migration Tool run itself is a small fraction of that.
Magento 1 to Magento 2 pricing has no list rate, because the variable is how many Magento 1 extensions get rebuilt, how many custom tables need mapping, and whether the ERP connector is rewritten. Directories publish minimum project sizes rather than migration quotes: Clutch lists a $25,000+ minimum project size for scandiweb. Ask each agency to price the module list separately.
Magento 1 extensions do not run on Magento 2. The code base, module structure and layout system all changed, so every Magento 1 module resolves to one of three outcomes: repurchase the vendor's Magento 2 edition, rebuild the behaviour as a custom module, or drop it because Magento 2 core now covers the function. That inventory usually drives the quote.
Ranking loss during a Magento 1 to Magento 2 move traces to url_rewrite, not to Magento 2 itself. Magento 1 stores rewrites in core_url_rewrite while Magento 2 uses a differently structured url_rewrite table with its own category suffix setting. Published migrations show the scale involved: over 120K 301 redirects at BUFF and over 21,000 at Zvaigzne ABC.
A Magento 1 store starting a rebuild in 2026 should target Magento 2.4.9, released 12 May 2026 with regular support running to 31 May 2029. The older lines leave little runway: 2.4.6 regular support ended 11 August 2026 with extended support to 31 August 2027, and 2.4.7 regular support ends 31 May 2027.