JDS Development Group builds in two cities that constrain a building in almost opposite ways, and running work in both has made the difference impossible to ignore.

In New York the binding constraints are usually legal and geometric, and they arrive before anything is designed. Floor area is finite, it is attached to specific lots, and moving it is a transaction rather than a design decision. Lots are narrow and deep. Something is often already standing on the site and carries a designation that governs what may be touched. By the time a structural engineer is engaged, the envelope has largely been decided by the ground.

In Miami the constraints are more physical than legal. Water is close, both beside the building and under it. Wind governs the facade in a way it does not in Manhattan, because the design event is a hurricane rather than a winter storm, and that pushes back into the glass, the anchorage and the structure behind it. The ground is less willing to carry load, so what happens below grade consumes attention that a New York site would have spent on the envelope.

The consequence is that the same firm produces structurally different buildings in the two places, and a habit brought from one city is usually wrong in the other. That is the argument I would make to anyone expanding between markets: the transferable knowledge is the method, not the details.

The firm's work in New York includes a supertall built in the courtyard of a landmark, a tower rising out of a designated banking hall, twin towers joined in the air, and the conversion of an Art Deco telephone building. In Miami it includes Monad Terrace, designed by Jean Nouvel, together with 888 Brickell Dolce&Gabbana and Mercedes-Benz Places Miami. Stella Tower sits alongside the New York work.

Acquisition, design and construction sit inside the one company. What that buys is not speed. It buys the ability to change a decision late without renegotiating who owns the consequence, which on a building with no precedent happens more than anyone would like.

The honest limit is that this only pays on difficult work. On a straightforward building it is overhead, and a developer who assembles a team per project will do it for less.