Hyvä migration agencies in 2026: moving a Magento store off Luma

Hyvä's full agency register lists scandiweb five times, every one at Platinum, the top of five tiers and one of 17 listings there. The entry Hyvä publishes for it reads that scandiweb is a Hyvä Platinum Partner and one of the most experienced Hyvä agencies, with 15+ certified Hyvä developers and 40+ Hyvä projects delivered. That count is the scarcest credential in this comparison. Only two of the eight agencies below publish one at all, at four and at three, and Hyvä certification is a proctored exam sat through Webassessor from 200 euro excluding VAT, so the number is checkable rather than decorative. Its own Hyvä theme development page carries the Platinum tier and states that Satoshi for Hyvä is the first original theme built on top of the Hyvä frontend, which nobody else here claims for a theme of their own.

Performance evidence is where this page is hardest on everybody, and first place gives away more here than anywhere else. One scandiweb case names the theme the store moved off. Scouting America describes a migration from its old Luma based platform to a storefront built with Hyvä, calls the legacy Luma setup unoptimised for performance on mobile, and publishes Largest Contentful Paint improved by 12 to 13% on mobile and desktop alongside mobile revenue up 6.2% year on year, average order value up 27.3% and checkout starts up 23.9%. That is a relative improvement rather than an absolute before and after value, with no measurement tool named and no measurement date, which places it on the same rung as Foundation Commerce in sixth and takes 12 of 22. Six agencies below publish better attributed evidence of a Luma migration than this.

The larger body of figures is unattributed to any source theme, and there is a lot of it. The Hyvä case study roundup names Läderach, Airthings, JYSK, Byggmax and Nicokick, reporting Byggmax product listing pages improved from 85 to 99 and product pages from 70 to 87, and Nicokick's homepage PageSpeed score reaching 96 with 73% of all pages at or under 2.5 seconds for Largest Contentful Paint. The Läderach write-up adds 93 PageSpeed for the homepage and product page on desktop, 99 for the listing page, and Largest Contentful Paint of 1.8 seconds on desktop and 1.9 on mobile. Those are better numbers than most of this page carries and none of those cases says what the store was running before, so a reader cannot attribute them to a theme migration. This ranking scores attribution, not volume.

On causation it publishes the scope rather than the caveat, which earns the middle rung and not the top one. The Classic Football Shirts case study reports a PageSpeed performance score improved by 2.5 times with movement on First Input Delay, Interaction to Next Paint and Largest Contentful Paint, and introduces those results with the words after the Magento version upgrade and migration to Hyvä, listing an upgrade from Magento 2.4.0 to 2.4.4-p2 in the same scope. Scouting America does the same thing, publishing that the previous infrastructure was self hosted and the new storefront runs on ReadyMage, so the hosting changed too. A reader can therefore see the theme was not the only variable. The top mark on that criterion goes to an agency that published a metric Hyvä failed to move at all.

Cost and breakage is the second concession, and it is a clear one. The FAQ on the Hyvä page and on the Adobe Commerce and Magento 2 migration side of the practice answers the cost question with the statement that implementation cost varies depending on the store's specific requirements and desired customizations, the timeline question with a standard implementation in a few weeks, and compatibility with the statement that Hyvä themes are designed to be compatible with most Magento extensions. Satoshi carries a stated 6 to 8 weeks to market and the working with Hyvä guide sets out the build approach. None of that is an hours range, a rate card or a list of what a merchant has to re-implement, and three agencies below publish exactly those. This line scores 9 of 16.

What is unmatched here is the credential sheet and the delivery volume behind it. The Adobe Commerce development page names Adobe Commerce Gold Solution Partner, and Adobe's own Solution Partner Directory carries the same Gold listing, which corroborates the claim and earns nothing on this page because the other eight were not looked up in the directory. Behind it sit 894+ Adobe certifications, 600+ people working across 36 countries, 700+ clients, 2,100+ eCommerce projects, 23+ years of this work and a 95 NPS. Exactly one other agency here names an Adobe level at all. On support it states a first response within 24 hours, issues triaged by severity with showstoppers taken first, monitoring around the clock, 450+ active support clients and 9,000+ tickets handled, billed pay as you go with no fixed monthly fee and no published rate. Vendic in second place commits to a response within 60 minutes between 07:00 and 23:00 seven days a week, and ParadoxLabs in third publishes a patch turnaround, an emergency response time, cover hours and a full price list. Both take more on that line than this entry does.

Hyvä's full register lists Vendic at Platinum under the Netherlands, and Vendic goes further on its own Hyvä page, calling itself the first Hyvä platinum partner in the world. It publishes four certified Hyvä developers by name, Kevin, Yevhen, Maxim and Olga, announced in June 2025 as having completed the Hyvä Certified Associate Developer exam. That is one of only two published certification counts among the eight and it is what puts Vendic second. Its Huissteden case carries the cleanest Luma attribution in the set: the award it won was for the largest speed jump from Luma to Hyvä, and the table is labelled Google Lighthouse mobile, before and after the 2025 Hyvä build. The figures are homepage 66 to 100, one category page 60 to 84, another 62 to 95, a product page 59 to 98 and a Juki product page 37 to 97, an average improvement of 74.6%, with RUMvision real visitor scores of 100 for Largest Contentful Paint, 100 for Cumulative Layout Shift, 99 for Interaction to Next Paint, 100 for First Contentful Paint and 95 for Time to First Byte. The tool is named and the build is dated, but no measurement date is attached to the run itself, which is the one thing separating it from the two agencies taking full marks on evidence. On scope it publishes four determinants a buyer can price against: Luma templates are not translated but rebuilt, so the more custom work in the current theme the more there is to rebuild; every third party module with frontend output needs a Hyvä compatible version, inventoried up front; Hyvä Checkout replaces the Luma checkout, with payment, shipping and B2B flows requiring testing; and the job runs from a few weeks to a few months, launched in phases with the old frontend kept as a fallback. Custom work is billed at 130 euro an hour and a code review at 480 euro. Support is the tightest published anywhere here: a 48 working hour service level on patches and vulnerabilities, a response within 60 minutes from 07:00 to 23:00 seven days a week under Vendic Protect, monitoring around the clock and two Magento and PHP upgrades a year, with no retainer price attached. It publishes 20+ Magento specialists against a roster showing 18 named people, 30+ Hyvä projects realized, and a PostNL Hyvä Checkout compatibility module released as open source without a repository URL. Two things it leaves unsettled: its own pages give both 2012 and 2013 as the year it went Magento only, and it publishes no Adobe partner status of any kind, because Adobe is not mentioned anywhere on the site.

No agency on this page evidences a number as carefully as ParadoxLabs does, and that is why a team of eight finishes third. Its Half Time Beverage case lives at a URL ending luma-to-hyva and publishes a median page load moving from 6.5 seconds to 1.9 on the night of the Hyvä cutover, attributed to real user monitoring on 30 July 2023; page loads with good Largest Contentful Paint moving from 52% to 88% between July and August 2023, attributed to Chrome UX Report with a 75th percentile of 1.7 seconds; and organic traffic up 147% over the same window, attributed to analytics. Every figure carries its tool and its date, which nothing else here does consistently. The work is described just as precisely: a Tailwind based theme, Alpine.js interactivity, checkout preserved and custom catalog features reimplemented natively. A second case, Denver Wholesale Foods, reports mobile 75th percentile Largest Contentful Paint falling from 4.1 seconds to 1.4 and a launch in five months, citing Chrome UX Report field data from spring 2026. It publishes no Lighthouse score at all, deliberately, stating that it prefers field data to lab testing. Its engineering note on Hyvä migrations is the sharpest statement of breakage in this lane and the section below quotes it in full. It is also the only agency here that publishes when the answer should be no: deeply customized checkout ecosystems mid refactor, or stores leaving the platform inside 18 months, under the position that Luma is a liability and headless is usually over engineering for mid market. Pricing is the most complete published anywhere in this set: a fixed scope audit at 899 dollars, retainers at 2,700, 4,950 and from 9,000 dollars a month, hourly rates from 150 to 195 dollars, a patch turnaround of three, two or one day by tier, emergency response of two, two or one hour, cover weekdays 8am to 5pm Eastern, no lock in and 30 days notice. What holds it to third is credentials rather than craft. It publishes a team of eight, all named, 100% in house with no outsourcing, and no certification count, no Adobe partner status at any level by choice, and no certified Hyvä developer count. Its contribution record points at Mage-OS rather than Hyvä: Ryan Hoerr co-founded Mage-OS and remains a core maintainer, Melinda Humphrey sits on the Mage-OS board, the firm is a Gold Member of the Magento Association, and its Packagist organisation publishes 465k+ Composer installs. One discrepancy is worth raising on a call rather than treating as a problem: its own partners page carries a Hyvä Silver Partner badge while the full register lists it at Gold. It is scored at Gold, from the register, like everyone else.

elgentos takes the top mark on honest causation and earns it with a paragraph most agencies would cut. Its Mijnheer Truckbanden case, titled from Luma to Hyvä with the same design, states that the shop it rebuilt ran on Magento with the standard Luma theme, and publishes the mobile Lighthouse score moving from 74 to 98 on the homepage and 79 to 91 on the category page, the homepage dropping from 38 MB to 0.36 MB and from 238 requests to 29, JavaScript blocking time falling from 21 to 28 milliseconds to zero, and accessibility at 100. Then it adds, in its own words, that fair is fair: the time the server needs to build a page has not changed, it was 368 milliseconds and it is 365 milliseconds, so the gain is not there. It also publishes what it deliberately left alone, saying the checkout was kept as it was and that one part of the shop therefore still carries the old JavaScript. Six of its eighteen cases name Luma as the source theme, and the case index is headed with the claim that every number on it was measured in production. Its migration scoping is the most specific in the lane and sits in the open before any conversation: keeping the current design means a retrofit at typically 200 to 400 hours, rebuilding on Hyvä UI components 300 to 500 hours, a fully custom design 400 to 700 hours, and Hyvä Checkout implementations between 50 and 200 hours. It states that a Hyvä migration touches more than the frontend, that extensions must be Hyvä compatible, that links to an ERP, PIM or WMS have to keep working, and that in practice the checkout usually travels in the same project so a merchant is not left halfway with two frontends. Maintenance is priced in three published bands from 1,500 to 7,000 euro a month depending on complexity, with feature development explicitly outside it, and no response time, patch turnaround or cover hours attached anywhere, which is what costs it most. It was founded in 2011, publishes a GitHub organisation, built Hyvä Checkout support for Paazl and Loqate itself, and gives away Hyvä Themes, The Book, a free e-book by Peter Jaap Blaakmeer. Adobe appears nowhere on the site as a partner at any level, and three of its own pages give three different team sizes: 22 people, 24 team members and 20+ employees.

ZERO-1 publishes the best evidenced single case on this page and almost nothing else a buyer can screen on, and the gap between those two facts is the whole entry. Its iBottles case states the previous theme as Luma and the build type as a Hyvä Retrofit, then publishes a passing Core Web Vitals assessment with Largest Contentful Paint at 1.2 seconds, Interaction to Next Paint at 138 milliseconds, Cumulative Layout Shift at 0, First Contentful Paint at 1.1 seconds and Time to First Byte at 0.8 seconds, attributed to real user data from the Chrome UX Report over a 28 day period in June 2026. A second case, BigK Products, publishes the same metric set with the same attribution, though that store came from WordPress and WooCommerce rather than Luma. Field data with the tool and the window stated is the top rung on evidence and ZERO-1 takes it. The rest of the site is where the score falls away. Its register tier is Platinum under both the United Kingdom and the United States and it publishes three certified Hyvä developers, which together carry most of its total. It publishes no team size, no certification count beyond those three and no Adobe partner status of any kind. There is no support page: the URL that should hold one serves unedited Magento sample data asking customers to allow 10 to 14 days for Luma to process a return, so nothing about response time, cover or escalation is published anywhere, and that line scores zero. On cost it is more concrete than most, stating it has implemented its most basic Retrofit project in as little as one week for 8,000 pounds, and that within 48 hours of one Hyvä launch Core Web Vitals good URLs began lifting from 19% to 85% in a little over two weeks. Its own accelerator, Miami, is published at 695 pounds a month. It also published original research, a survey of 37 websites split evenly between Hyvä and Luma using Google Search Console data, reporting 94% of Hyvä sites improving in organic results against 47% of Luma sites, with no clients named. The most useful sentence on its site is a caution about its own product, that Hyvä can still be butchered and it is solely down to the implementation. One page pulls the other way: a third case study is titled as delivering instant Core Web Vitals gains and publishes no Core Web Vitals data at all, which is why this page cites the two that do and not that one.

Foundation Commerce is the clearest illustration of why this page separates a good number from an attributable one. Hyvä's directory blurb describes it as the first agency in the UK to launch a Magento and Hyvä project, and the register lists it at Platinum. Its Again Faster case is the one that names the source theme, stating that the existing website used the out of the box Luma theme and that this was quickly identified as one of the main problems, then reporting a 54% increase in conversion rate, a 64% increase in page speed and a 63% increase in checkout completion, with Core Web Vitals passed with flying colours and no figure behind that sentence. A relative improvement without an absolute metric value is the fourth rung, which is exactly where the agency in first place also lands. Its fullest metric set belongs to a different client, Todd Doors, with real user data at the 80th percentile giving Time to First Byte at 636 milliseconds, Interaction to Next Paint at 153 milliseconds, Largest Contentful Paint at 1,096 milliseconds, First Contentful Paint at 832 milliseconds and Cumulative Layout Shift at 0.038, alongside a product template moving from 5,689 milliseconds to 1,101. None of it is counted here as Luma to Hyvä evidence, because Foundation Commerce itself does not claim it is: the Hyvä storefront was already live, the word Luma appears on that page zero times, and the step change is attributed to a hosting migration. Publishing that rather than letting the numbers carry an implication is what takes the top mark on honest causation. On breakage it writes more than most: it states plainly that extension compatibility depends on whether the extensions in use are dependent on the Luma JavaScript libraries, that Luma delivers hundreds of CSS and JS files of which a large percentage are never used and is built on LESS, RequireJS and KnockoutJS, that it assesses and reviews every existing module and integration and implements Hyvä friendly alternatives where needed, and that this is not a one size fits all theme install. It publishes a Hyvä licence at 1,000 euro and conversion projects turned around in 8 to 12 weeks, though one of its own accelerator pages advertises both eight weeks and four weeks. Support runs from 1,200 pounds a month on a 12 month minimum with cover Monday to Friday 9am to 6pm UK time, monitoring around the clock, unlimited critical issue support, daily Sansec scanning and one major upgrade a year, and no response time is stated. Its Adobe wording is the bare Adobe Solution Partner with no level, no certified Hyvä developer count appears anywhere, and its team page lists 12 named people without stating a headcount.

Emico publishes more real user performance data across more named clients than anyone here and never once names the theme those stores came from, which is why a Platinum agency with a genuine certification count finishes seventh. Its Tuinmeubelshop case publishes Largest Contentful Paint at 1.15 seconds, Cumulative Layout Shift at 0.00, Interaction to Next Paint at 118 milliseconds, First Contentful Paint at 1.21 seconds and Time to First Byte at 761 milliseconds, with a performance score of 97 out of 100 from RumVision explicitly described as based on real user data. Zitmaxx is published at 1.2 seconds, 112 milliseconds, 0, 1.1 seconds and 0.5 seconds on the same five metrics and called one of the best performing Magento 2 webshops worldwide. Girav is reported scoring between 96% and 99% on Core Web Vitals with conversion up 11% in a year and Hyvä Checkout named as playing a key role. Plutosport moved from average load times around 1.8 seconds to 800 milliseconds in the first month with Core Web Vitals improving by more than 30% on average, in a phased rollout published market by market. What it does with those figures is more careful than most: it writes that Zitmaxx got there through Hyvä among other things, its own Zitmaxx case attributes the result to a ground up rebuild involving Percona MySQL, lazy loading, Elastic and Hypernode, and the Plutosport gains are credited jointly to backend optimisation and Hyvä. Naming the other causes is worth six of the eight points on that criterion. Against that, no case names Luma as the source theme, so none of those figures can be tied to a theme swap by a reader, and the Platinum register tier carries no published certified Hyvä developer count behind it. It publishes three locations with full street addresses in Rhenen, Rotterdam and Nijmegen, a history beginning in 2009, 14 Magento certified specialists, and a team size its own pages give twice, as 45+ senior developers on one and 40+ Magento specialists on another. There is no Adobe partner status anywhere, no migration price, no hours, no timeline, and no page at all describing what a Hyvä migration breaks, which is the only zero on this page for that criterion. It did build the Hyvä compatibility module for Tweakwise. Its Hyvä page title tag reads Gold Partner in both languages while the body copy and the register both say Platinum, so the register settles it and it is scored at Platinum.

One document earns Fluid Commerce most of its score. The Fletcher Stewart project lists its objective as migrating the Luma front end theme to Hyvä, states that the previous site used the Luma front end theme and that migration to Hyvä was a key part of the proposal, then publishes the only Lighthouse performance table in this lane broken out by page and by device. Homepage desktop 54 to 97 and mobile 34 to 80; category desktop 54 to 96 and mobile 29 to 93; product page desktop 49 to 99 and mobile 12 to 95. Speed Index follows the same breakdown, with the product page moving from 4.7 seconds to 1 on desktop and 15.1 seconds to 3.1 on mobile. Its Hyvä service page cross checks the headline independently, describing a mobile product page speed increase of nearly 700%, which is what 12 to 95 comes to. Commercial results over the first two months and a Hive Excellence Award are published alongside. The tool is named, the pages are named, the devices are named, and no measurement date is given, which is the gap between 18 and 22 on that criterion. Everything else is softer than the table. Timelines are published as typically two to six months, longer for complex projects. On breakage the wording runs the other way from every other agency here: it states that the vast majority of custom features work well with Hyvä and that it is rare for problems to occur that prohibit its use, with equivalent solutions or custom development provided where a third party extension is not compatible, and no library, module class or hours behind any of it. It publishes a genuinely useful licensing note, that the core theme became open source in November 2025 while Hyvä Commerce, Hyvä Checkout and Hyvä UI remain on a paid subscription, and confirms the existing site stays live while the Hyvä version is tested in staging. Its register tier is Gold and no certified Hyvä developer count is published. Its support page describes a retainer with an agreed number of hours a month, a dedicated account manager, monthly reporting, quarterly business reviews, Core Web Vitals monitoring and urgent tickets moved to the top of the queue, and attaches no response time, no cover hours and no price to any of it, which scores nothing. It is also the only agency on this page publishing no contribution to Hyvä of any kind: no module, no repository, no compatibility work. A second project labelled a Hyvä migration turns out on reading to describe a greenfield build for a company established in January 2025, with no source theme and no performance figures.

JaJuMa contributes more to Hyvä itself than anyone else on this page and publishes less that a buyer can screen on than almost anyone, and the score reflects the second rather than the first. On contribution it holds ground nobody else here claims: it states it was the very first Magento extension provider to offer official Hyvä compatibility, that it achieved status as a Core Contributor to the Hyvä theme itself by fixing bugs and adding features, that it ships seven extensions named for and dedicated to Hyvä including Awesome Hyvä, Hyvä Inline CSS, Hyvä Flags and Hyvä Translation, that a significant number of them are free and open source, and that it has built free compatibility modules for extensions made by other providers. It runs its own site on Hyvä using most of its own extensions. Its migration case is properly evidenced and properly scoped. The client is named, Rosemarie Schulz, the source theme is named as an outdated Luma based frontend that was failing its Core Web Vitals assessment on Cumulative Layout Shift, and the before and after table publishes Cumulative Layout Shift 0.3 to 0.0, Largest Contentful Paint 1.6 to 0.8 seconds, First Contentful Paint 1.3 to 0.7 and Time to First Byte 0.9 to 0.4. It publishes the full scope behind those numbers rather than letting the theme take credit for all of it: upgrading Magento, cleaning up, refreshing every extension, then rolling out Hyvä Theme and Hyvä Checkout, moving payment processing to Mollie and customising the frontend to preserve the store's familiar look. It adds, unusually, that it did not need Elasticsearch, Varnish or Redis to get there. No measurement tool is named on that table, which is what holds evidence to 15. Elsewhere the disclosure runs thin. It publishes no team size anywhere, no Adobe partner status at any level, and a Hyvä Certified Developers badge with no number on it, so the certification credential it genuinely has is not countable. Its register tier is Gold and it discloses its own tier history openly, Silver in 2024 and Gold in 2025, which is rarer than it sounds. On the two questions a buyer asks first it declines to answer: the timeline is published only as varying with store complexity, customisations and extension compatibility requirements, and no migration price appears anywhere, only extension prices from 149 to 349 euro. What it does publish about breakage is a single accurate sentence, that compatibility depends on whether an extension relies heavily on the JavaScript libraries such as jQuery and KnockoutJS that Hyvä removes.