Can we adopt this gradually, starting with a single Sanity component?
Dev at a Swedish SaaS. We want to pilot personalization on the homepage hero only. Leadership is on board with an experiment but nobody wants a project that touches the whole site or restructures our Sanity dataset.
So the question is scope: can we map just one component, run the pilot, and leave literally everything else alone? And if the pilot does not earn its keep, how messy is the removal, I dont want to inherit a half-integrated system either way.
1 answer
One component is exactly how this is meant to be adopted. The integration works at component level: you map your hero component to a single Croct slot, and every other component on the site keeps rendering straight from Sanity with zero changes. There is no migration and your dataset is not restructured, the mapping is an overlay on top of what you have.
During the pilot, the Sanity hero content stays as the fallback for the slot. Visitors who match no experience, or any request where the fetch fails, get the regular hero exactly as today. That makes the pilot low-risk in the literal sense: the worst case is the current behavior.
Removal is equally contained. If the pilot does not pay off, you unmap the slot and delete the fetch from the hero component, and the site is back to pure Sanity. Nothing else references the integration, so there is no half-integrated residue to clean up.
We ran this exact play, hero first, and the useful side effect was that the pilot produced the internal evidence for whether to expand, instead of arguing about it in the abstract.