Capture the pet once
The single most useful thing a pet storefront can know is the pet. Weight determines harness and bed size; age and life stage determine food; dietary sensitivities rule out half the range; daily portion determines when the bag runs out.
Captured once as a profile, that data does a great deal of work. It filters thousands of products to the few that suit, pre-selects sizes so accessories are not returned, and times reorders so food arrives before it runs out. Most stores ask for none of it and leave the owner to work it out on every visit.
Autoship is the business model
Pet food is one of the most predictable purchases in retail. The pet eats a known amount a day, so the reorder date is arithmetic. That makes subscription the commercial centre of a pet business rather than a feature bolted on.
The subscriptions that last are the ones customers can manage themselves: skip a delivery when they are away, swap flavours, bring an order forward when a second pet arrives. The ones that churn are those that arrive at the wrong time and can only be changed by email.
Where this sits
The storefront and subscription mechanics are eCommerce development; nutrition, ingredient and sizing attributes are product data; and reducing early autoship churn is usually conversion work applied after the first order. Most pet retailers carry other brands, so the multi-brand retail mechanics apply underneath.