Starbucks: Handcrafted Sodas?

The Wall Street Journal reports that Starbucks is testing handcrafted sodas at locations in Atlanta, Georgia, and Austin, Texas.   The article states that the flavors of soda include spiced root beer and lemon ale.  The sodas are created using a carbonation machine in the Starbucks stores, and a grande size drink sells for $2.95 at these locations.

What do we make of this newest product line extension for Starbucks?  As the article explains, the move clearly constitutes another attempt to drive sales at off-hours, i.e. in the afternoon and evening.  Moroever, it offers another way to drive same-store sales growth, a key metric for any restaurant or retailer.   

The move does come with some risks.  Naturally, customers must enjoy the product, or it could harm the Starbucks brand.  In addition, continued menu extensions add complexity to the operations of a particular location.  Can Starbucks maintain speed and customer service as the menu expands?   With food items expanding as well, Starbucks must watch the issue of complexity carefully.  The good news:  Starbucks understands the power of experimentation.  It is testing the concept in two locations, and presumably, it's gathering a great deal of data.  That data include customer feedback, barista input, and information about the impact that the new products are having on the speed of service.   What happens if customers love the soda, but service slows?  That could be tricky.  At that point, Starbucks will work to streamline operations as much as possible to increase throughput.   They have done quite a bit of that type of work in the past few years.  Still, I notice considerable differences in speed across locations.  Part of the difference can be accounted for by the fact that some store layouts are clearly not as conducive to streamlined, efficient operations.   Part of that is due to local differences in the way locations are managed.  

One final point about the notion of testing and experimentation.  Sometimes, companies find that such small tests go quite well, but national roll-outs then falter.  Why?  The tests are not truly representative of what will happen during a full-scale expansion.  Senior managers supervise the tests very carefully. Extra resources are deployed.  The test becomes something more than a test... it's actually a "proving ground" or "demonstration" rather than a true controlled experiment.