Applying Kanban to IT support work

Ask anyone who has a glancing familiarity with Kanban what they know of it, and one of the (if not the) first things they will mention, is the use of a kanban board. This is true; the kanban board, whether physical or digital, is one of the most visible parts of the Kanban method. It is an eminently visual way to represent WIP. So, how do you implement a kanban board in IT support?

Chances are, you already have (and if you haven’t you really need to go back to the drawing board – you’re doing it wrong). What I mean by that, is that any IT department with any respect for themselves and their trade has implemented some sort of ticket tracking system. That, right there, is (or more accurately can be) your kanban board.

The board, however, is simply a prop. The board is less important than what it represents: the flow of the tasks you (the collective You, the support department) have to contend with. The idea, as I understand it, is to limit the number of tasks each person is allowed to have assigned to them at any given time. Now, the queue of incoming tickets remains, but resolution of tasks (as in their being assigned to someone) is only begun when there is available capacity to do so.

It is a common requirement in IT service departments – certainly in all I have worked – for incoming tickets to be classified and assigned as soon as possible, and these two are usually seen as two parts of the same action. With the WIP restrictions of Kanban, however, that changes. Tickets do need to be classified as soon as possible as their impact on the organisation, the urgency of getting a system back up and running, and the resulting priority is key to prioritising the tickets.

Assigning them, however, is not immediately needed. Most systems have system users to hold unassigned tickets. Use these as the queue, and stagger the release of tickets to technicians as they resolve other tickets, releasing capacity for another ticket. Doing this will allow technicians to concentrate on a smaller number of tasks at a time, devoting more attention to them, and achieving reduced batch size with an eye towards single piece flow.