The business is decided in thirty seconds at the desk
A salon can have excellent stylists, a full Saturday and a steady stream of new faces, and still be flat year after year, because clients leave without a next appointment and drift away without anybody noticing. Salon reporting is mostly about that moment at the desk and the weeks that follow it.
Rebooking rate at checkout
Share of visits that ended with the next appointment booked, by stylist and service. It decides desk coaching. A bad number is a wide spread between stylists, which is the good case because it means the habit is learnable. The action is a short script, a diary open at the desk, and the number published weekly so the team can see it move.
Column utilisation by stylist and slot
Booked hours against available hours, by stylist, day and time band. It decides the rota and where offers go. A bad number is empty weekday mornings while weekends turn people away. The action is a targeted offer into those specific slots rather than a general discount that mostly gives away hours you were going to fill anyway.
Client lapse by service interval
Clients past the normal interval for their service, with last visit and usual stylist. It decides the weekly contact list. A bad number is a long overdue list that has never been worked. The action is to assign it in small batches to the stylist who knows them, because a message from a familiar name performs far better than a salon wide campaign.
First to second visit retention
New clients who returned, by stylist and by source. It decides whether acquisition is worth what it costs. A bad number is strong new client volume with weak second visits, particularly from discount platforms. The action is a deliberate rebooking conversation during the first visit rather than more acquisition spend.
Retail attach and no show patterns
Retail sold per visit by stylist, plus no shows by slot, service and client. Together they decide training and policy. A bad number is retail concentrated with two stylists, or no shows clustered in specific slots that a confirmation routine would largely solve.
