A grocery store is a good place to think about design. Everything has to arrive, move, be stored, be found, be restocked, and eventually leave. The checkout screen is one part of that operation. By the time a customer gets to the register, an enormous number of decisions have already been made on their behalf.

Designers study the shopping journey. But the loading dock, the scheduling system, the warehouse worker, the purchasing agreement, and the rules about what gets stocked determine much of what happens before a customer arrives.

Those things have a direct effect on the experience. A store can have a wonderful app and still run out of food. A delivery service can have a good tracking interface and still schedule its drivers badly. A government agency can replace every form on its website and leave the actual process untouched.

I've been thinking about this as Interaction Logistics: bringing operational and supply-chain thinking into the design of interactions. Logistics asks concrete questions about movement, dependencies, timing, capacity, handoffs, and the consequences of a failure at any point in the chain.

The frontstage/backstage distinction can help document a service. But if the design team never gets access to the place where staffing, inventory, and scheduling decisions are made, the diagram will only tell them so much. If the people building the interface cannot alter the staffing plan, the work queue, the inventory process, or the handoff, they are designing around decisions somebody else has already made.

This matters outside retail. The same questions apply to a school bus arriving on time, an ambulance reaching a patient, a public-benefits determination, or a worker trying to find the current version of an operating procedure. Each involves people, places, materials, information, and obligations moving through an organization.

It also changes the skills we should expect from people who design these systems. They need to observe work in its setting. They need to understand constraints that cannot be solved with a better screen. They should know enough about operations to recognize when a user-interface problem is actually a capacity problem or a bad assignment rule.

A logistics perspective also makes it harder to hide the costs of convenience. An improved customer experience may mean an unpredictable shift for a driver, new unpaid coordination work for a store employee, or an increasingly fragile chain of contractors. Every streamlined interaction has an operating arrangement underneath it.

I want designers to have a hand in that operating arrangement. If a pickup fails because of a warehouse schedule, changing the checkout screen won't fix it.