Horizon is new enough that genuine experience varies sharply between developers. That's the main thing to screen for.
Ask how many Horizon builds they've shipped, and when. Horizon launched in 2025 and updates weekly. A developer who worked on it during beta and has shipped several stores since knows things that someone reading the docs this month does not — particularly around app compatibility and the settings with undocumented behaviour. Ask for live Horizon stores with dates.
If you're migrating from Dawn, get the scope right before you commit. There is no automatic migration path. Your current theme's customisations, custom sections, and app integrations all have to be rebuilt rather than ported, commonly 30-60 hours for a moderately customised store. Ask for a written inventory of what's being rebuilt before you agree a price, because "migrate us to Horizon" quoted without that inventory will change.
Have your app list ready. App compatibility is the most common source of unexpected cost. Older apps that inject into the variant picker, gallery, or cart drawer often break because those are now web components. Send your installed app list with the brief — a developer who knows Horizon can flag likely problems before starting rather than discovering them mid-build.
Ask about mobile performance specifically. Horizon's mobile PageSpeed has trailed Dawn, and some of the gap is configuration rather than architecture. A developer who has optimised Horizon on mobile will have specific opinions about animation settings and which choices cost what.
Consider whether you should migrate at all. If you're on a mature Dawn store with a custom design that performs well, the case for moving is weaker than the marketing suggests — you'd be paying for a full rebuild to reach roughly where you already are. Horizon is clearly where Shopify is investing, so the gap in what's buildable will widen, but that's an argument for planning a migration rather than rushing one. A good developer will say this.