Hyvä vs. Luma — Why 2026 Is the Year to Migrate Your Magento Frontend
I measured the mobile performance of 28 Magento shops in Tirol. The average score: 55 out of 100.
Google considers anything below 50 "poor" and anything between 50 and 89 "needs improvement." The shops that had migrated to Hyvä? They were scoring 90+. Some were at 93. One was at 100.
That’s not a marginal difference. That’s a different league.
What is Hyvä?
Hyvä (Finnish for "good") is an alternative frontend for Magento 2, built from scratch with performance as the primary constraint.
The default Magento 2 frontend, Luma, was designed in 2015. It relies on RequireJS, KnockoutJS, and jQuery, a stack that made sense at the time but delivers 200-500KB of JavaScript before the page becomes interactive. On desktop with a fast connection, most users don’t notice. On mobile, which is where 60-70% of e-commerce traffic actually happens, it’s a problem.
Hyvä replaces all of that with Alpine.js and Tailwind CSS. The result is roughly 80-90% less JavaScript. Pages load faster because there’s simply less to load.
What the Tirol audit found
82% of the Magento shops I measured scored below 60 on mobile. Only 4 shops, 14%, scored above 80. The ones scoring 90+ were either on Hyvä or running very lean, simple storefronts.
The worst performers were striking: one shop scored 9 on mobile. Several others were in the 30-35 range. These shops aren’t broken; they load, they function. But they’re losing visitors on mobile before the page even finishes rendering.
The most instructive comparison came from a shop I have direct visibility into: Naturabiomat. Their production site runs Luma and scores 46 on mobile. Their staging environment, same content, same products, migrated to Hyvä, scores 93. That’s not a performance optimisation. That’s a frontend replacement.
Same shop. Same content. 47-point difference.
Why this matters beyond the score
The PageSpeed score isn’t a vanity metric. Three things connect it directly to revenue.
First, user behaviour. Google’s own research shows 53% of mobile visitors abandon a page that takes more than 3 seconds to load. A mobile PSI of 55 typically means load times of 4-8 seconds. You’re losing more than half your mobile visitors before they see your products.
Second, Google rankings. Core Web Vitals have been a search ranking signal since 2021. Slower sites rank lower. Lower ranking means less organic traffic, which means higher customer acquisition costs or simply fewer customers. The shops scoring 9 and 35 on mobile aren’t just frustrating users; they’re invisible in search.
Third, conversion rate. Deloitte and Google published research showing every 100ms improvement in mobile load time increases retail conversion by roughly 8%. A jump from 50 to 90 on PSI typically represents 2-4 seconds of improvement. That compounds.
The back-of-napkin calculation
For a shop doing €500,000 a year online, at a 2% conversion rate:
- A conservative 15% conversion improvement from performance work brings the rate to 2.3%
- That’s €37,500 in additional annual revenue
- A Hyvä migration typically costs €12,000-25,000 depending on the complexity of the current theme and custom extensions
- Payback period: 4-8 months
These are conservative numbers. Some shops see 30-50% conversion improvements after significant performance work. The bigger the shop, the more expensive every slow second is.
What Luma is actually costing you
Luma isn’t just slower. It’s architecturally slow. The framework requires loading and executing hundreds of JavaScript files before the page becomes interactive. No amount of caching or CDN configuration fully compensates for this. The optimisations available within Luma have real ceilings.
Hyvä removes the ceiling. It’s not a faster Luma; it’s a different approach built for how browsers and devices work today, not in 2015.
The 18 Luma shops I identified in Tirol during the audit are sitting on the wrong side of that gap. So is any Magento 2 shop that hasn’t specifically moved to Hyvä or another modern frontend.
And most of them are missing critical security protections as well. Performance isn’t the only open question.
What a migration actually looks like
A Hyvä migration is not a redesign. The design stays. The content stays. What changes is the underlying frontend technology.
In practical terms: the Hyvä theme is implemented, existing customisations are adapted rather than rebuilt from scratch, extension compatibility is checked, and everything is tested against the production environment before go-live. For a shop without heavily custom frontend work, 2-4 weeks is a realistic timeline.
The migration is accelerated further by AI-assisted code analysis, tools that can assess existing customisations, flag compatibility issues, and handle the bulk transformation of templating code. I’ll cover that workflow separately. The short version is that it’s faster and cheaper than it used to be.
Google’s PageSpeed Insights will give you your mobile score in 30 seconds. If you want to understand what it means in practice, or what a Hyvä migration would realistically look like for your specific setup, I’m happy to take a look.
The next article pulls back further, looking at all 303 shops listed on shop.tirol and what the data says about the state of e-commerce in Tirol broadly.