Seth Weissman makes the case:
The problem is that the desire for ground floor retail comes at a time when we already have too much retail, much of it sitting empty. If all of the ground floor retail that planners are wanting cannot be filled and developments end up with revolving door tenants and boarded up store fronts, the ground floor retail can financially undermine the mixed use projects we are trying to encourage.
So can streets be active, interesting places without ground floor retail?
Of course, people walk on streets where they feel safe and where there are interesting things to see or do. Residential uses with interesting architecture, multiple doors opening up to the street, street gardens, stoops and front porches make streets desirable places to walk (think of the Savannah historic district).
Most cities would be better off encouraging uses that are actually needed by the residents of these new urban areas including food stores, dry cleaners, hair salons and even some of the big box uses that planners are taught to dread.
When such retail uses are combined with street level residential and office uses, what you get may not be as glitzy as the fantasized urban areas, but will likely create urban areas that are more sustainable over the long term.
This reads like one of those pieces that comes from a personal experience and some frustration associated with it, but I do think there's a lot in here that Mr. Weissman is correct about. In the enthusiasm for walkable, mixed-use neighborhoods, some planners by extension believe that should mean ground-floor retail is needed on every project on every block. And, that's certainly not reality. Coffee shops won't work everywhere, nor will viable storefront retail. Residential and other uses (I prefer to think in terms of building types, not uses) can all be attractive and pedestrian-friendly. The hat tip to Savannah is noted, and very true. A great deal of Savannah's landmark historic district is residential-only, and no one would argue that it's beautiful to explore on foot. (In fact, I have a book about this coming out soon)
But I do want to throw one important condition into the discussion. Often this argument is presented by developers who a) don't really want to do mixed-use, walkable places but are forced to or b) have a user in mind to lease a space that doesn't care about walkability. As more and more places require active streets and better design, more conflicts are happening with builders and developers who've been doing it the "old" way for many years. In those cases, it's important for planners and designers to hold firm and demand something better than "good enough."
While there's no doubt that a successful, walkable city can contain everything from auto service to big-box stores to salons and much more to complement other retailing, it's also true that lowering standards of design is sure to kill any momentum. We're in a transitional time of re-learning how to make cities that people love, and ferreting out the good from the bad and the ugly is going to cause quite a bit of heartburn in the years ahead.