Hydrogen and Oxygen are mature now, the developer experience is good, and there are plenty of case studies. We build Hydrogen storefronts when the case is real. Most brands who ask us about headless should still say no, and this is how we work out which side of that line a brand sits on.
Most of the cost comes later
The build is more expensive, by something like 50 to 80 percent. The bigger number is what you spend over the next three years, because going headless means you now own:
- The whole storefront stack. Hydrogen is a Remix app you maintain: updates, security patches, framework upgrades.
- Edge caching. What gets cached, for how long, with what invalidation. Liquid handles this for you.
- Image optimisation. Built into Liquid; configured by you in Hydrogen.
- A/B testing. Liquid plugs into Shopify's native experiments. In Hydrogen you bring your own.
- App fit. Most apps drop a script tag into Liquid and work. Hydrogen needs an explicit integration for each one, so "install Yotpo" becomes "build the Yotpo integration".
If you have the team and budget for that, fine. Most brands do not, and they feel it in year two.
When Hydrogen is the right answer
We see four cases where it is.
1. Merchandising Liquid cannot express
Composable bundles assembled at request time, product pages managed from a separate CMS, personalisation tied to a logged-in user's state. Liquid's section and block model cannot do these well. Hydrogen can.
Picture a build-your-own-scent flow with 12 selection steps, pricing that depends on the combination, and a recommendation layer calling a separate ML service. Liquid would collapse under it. Hydrogen handles it cleanly.
2. Content and commerce together
If half the storefront is content and that content lives outside Shopify (Sanity, Contentful, a custom CMS), Hydrogen's data fetching makes the composition clean. Pulling from two GraphQL endpoints on one page is the natural Hydrogen pattern.
3. Internationalisation beyond Markets
Markets covers a lot. If you need country-specific catalogs that are not subsets, region-specific flows, or content that varies by region beyond translation, Hydrogen gives you the routing and data fetching to do it.
4. Performance at the far end
Hydrogen's edge runtime can beat Liquid on TTFB for certain storefront shapes, especially product pages with heavy dynamic content.
If your Liquid TTFB is already under 600 ms, though, you will gain more conversion by removing three marketing tags than by going headless.
When Liquid is the right answer
For most brands the right answer is a clean Online Store 2.0 build on Liquid. It fits when:
- Your marketing team ships campaigns weekly. Online Store 2.0's section blocks are the difference between marketing filing a ticket and marketing shipping.
- You lean on Shopify-native tooling. Markets, Audiences, Functions and Analytics all integrate more cleanly with Liquid.
- You use off-the-shelf apps. Most are built for Liquid first. Headless integrations exist but vary.
- Your dev team is small. A Hydrogen app is a permanent maintenance cost. If you are not staffed for it, the cost shows up as a storefront that feels stale and a redesign nobody knows how to ship.
The trade-off for most brands
Most brands doing $5M to $50M a year on Shopify Plus do not need headless. They need a clean Online Store 2.0 theme, a section library their marketers can ship from, a performance budget enforced in CI, and the right integration layer. That is a six to ten week build for most stores, and it produces a storefront that ships campaigns weekly without dev tickets. Headless gets you a more flexible storefront and a permanent engineering tax.
What we ask on the call
When a brand asks whether to go headless, we ask four things:
- What specifically can you not do today that you want to? If they cannot name three things, the answer is no.
- Who maintains the storefront in two years? "Our agency" is fine. "Us, but we are not sure how" is a no.
- Is your performance ceiling limiting growth, or has someone told you it is? We want real numbers.
- Are you using Shopify-native tools you would lose? Markets, Audiences and native A/B testing all have headless paths, but none of them is free.
If headless still looks right after those four, we go headless. If not, the conversation turns to a clean Online Store 2.0 build that does 90 percent of what they wanted at half the cost. They can go headless later if the case shows up.
What would change this advice
Two things would push more brands to Hydrogen. Shopify keeps investing in headless tooling, and the gap between installing an app on Liquid and integrating it on Hydrogen closes every quarter. And if AI-driven personalisation becomes standard, the composition flexibility of headless wins once half the product page is dynamic.
We watch both. For now our default is Liquid, and we say yes to Hydrogen when the case is real.