Jordan BennettJordan Bennett

MassArt Kennedy Starbucks: a service audit

The Kennedy building entrance on Huntington Avenue, with an angled green canopy over the glass doors
Kennedy Hall on Huntington Avenue.

Field notes from one observation session, 9:00 am. 

Honest limits: n = 3 timed orders plus a 15-batch count, only one visit. 

1. What is easy to see going wrong

2. Why these exist (hypotheses)

The store is a campus licensee run by Sodexo, not a corporate Starbucks, so it inherits a shared dining operator's schedules and layout. Students arrive in waves between classes, with two other Starbucks within a short walk, so one slow wave costs a customer. Work seems to be released in batches rather than as items finish, and nothing shows order status. I have not verified any of this from inside the kitchen.

3. The evidence

Food, order 13:59Food, order 24:25Drink, order 15:15Drink, order 25:38
Seconds waited after ordering (observed). Blue is drink, navy is food. Food took about 4 minutes though a muffin needs no making.
Food batchesDrink batches
Each dot is one item released together. Food: 8, 3, 3, 2, 1. Drinks, 15 releases: 4, 2, 3, 4, 3, 3, 2, 1, 1, 1, 2, 1, 3, 2, 2 (assumed one continuous run; the second row of notes may be a separate window).
5:38 Eight hands, one cup.The minutes pool at the counterwhile the tables sit empty, listening.
Poetic graphic: eight staff around one waiting cup.
Finished drinks and mobile orders lined up at the pickup counter
Photo 4: the pickup counter — finished drinks and mobile orders pile up in one spot while ordering keeps happening beside it.
Open tables in the foreground while a queue of customers waits beside the counter
Photo 5: the open tables — seating sits empty in the foreground while the queue runs the length of the counter.
People choosing to stand and wait, tables open…Observer's field note
How much food comes out at once?Observer's field note
8 people working, someone back from break.Observer's field note

4. Why I chose this problem

I picked a place I use between studio classes, where a 10-minute break turns into a 6-minute wait. It is small enough to observe in one sitting and has a clear timing signal.

5. Assumptions and biases

I am a regular, so I notice delays more than a first-time visitor. I assumed the staff were not the problem, which my count supports but one visit cannot prove. I timed only my own order, and I expected mobile ordering to be faster.

6. Recommendations for the owner

Scores are 1 to 5: desirable (do students want it), feasible (can the store do it), viable (does it earn or save money). They are my judgment from this evidence, not tested numbers.

RecommendationDesirableFeasibleViableTotal
Split food from the drink line
One person pulls and hands over food the moment it is ready; the 8-item pile never forms. Uses existing 7 to 8 staff; pastries took ~4 min.
55414/15
Stagger breaks around the peak
Never run 2 on break at the busiest minutes; confirm the peak with a 3-day tally. Costs nothing; schedule change only.
35311/15
Show order status on a pickup screen
Name and ready-state on a small display so people can sit at the open tables. Standing and waiting beside empty tables is the clearest cue.
44311/15
Class-change mobile order windows
Push "order for the 11:50 break" so mobile orders are made before the rush. Matches 10-minute between-class demand; mobile saved only ~30 sec.
43512/15
Pre-pack cold pastries at the register
Muffins and cold items go in a grab-and-go case: no wait at all. Removes items from the queue entirely.
45413/15
Pickup lane away from the queue
Mark a separate pickup spot with a floor and wall cue so ordering and waiting stop colliding. Needs a layout check on site.
43310/15