Case study: projecting a bakery's loyalty program into a second context

The client, a regional bakery with nine locations, ran a loyalty program that worked well in its original stores and poorly in its newer ones. They asked us whether to redesign it.

We proposed a projection first: move the program, unchanged, into other contexts and observe what holds its shape.

Projection one: a different customer

We modeled the program for commuters rather than neighborhood regulars. The points structure survived. The reward did not: commuters did not want a free loaf, because they could not carry it on the train. This explained most of the gap at the newer stores, which sit near stations.

Projection two: a different scale

We modeled the program at forty locations. It held, with one change: the cards were stamped by hand, and at forty locations the stamps would need to match. This was cheap to fix and would have been expensive to discover later.

Projection three: a different dimension

We attempted, as we sometimes do, to project the program into a third dimension, to see how it behaved with depth as well as frequency and spend. It did not survive. The program has no depth; a customer can visit often and spend a great deal and still not be any further into it. We noted this as a finding, though the client did not ask about depth, and we were unable to explain it to them in the time remaining.

Outcome

The bakery kept the points structure, changed the reward at station locations to a coffee, and standardized the stamps. Repeat visits at the newer stores rose over the following two quarters. The third projection is in the appendix.