Most Shopify setups assume products are independent.
They’re not.
The moment you’re dealing with gear, components, or anything that has to work together, the platform starts to show its limits fast.
I approached this in two layers:
1. PDP Compatibility (validation layer)
Before anything gets added to cart, the system filters out invalid combinations. Not recommendations - hard constraints.
2. Structured Builder (assembly layer)
Instead of browsing and guessing, users move through a guided flow that resolves a complete, valid setup across multiple product types.
Not bundles. Not “frequently bought together.” Not AI suggestions.
A few things that became very clear:
- Product-level logic is useless - this has to operate at the variant level
- Letting users make bad combinations is a bigger problem than limiting choice
- Once you constrain the pool, ranking logic becomes the real battleground
- The highest intent shows up when users are allowed to modify within constraints (not start from scratch)
Everything is deterministic. Because in this context, “close” is still wrong.
What’s interesting is the behavior shift - structured flows are producing significantly stronger intent signals than standard browse → click → hope-for-the-best patterns.
Curious who else has run into this wall with Shopify - and what you did about it. Because most of what I see in the ecosystem doesn’t actually solve the problem.