Using Scrum to support in progress experiments

[Robert Freedman]

We are trying to use a scrum board to support a set of in progress experiments across a range of teams. The need:

- There are 24 experiments, spread over 9 sprints (1-2 weeks per sprint)

- Each team of 3-4 people completes a given experiment and rates it a status of 0, 1, 2, or 3 (a given individual on the team is the assignee for an experiment).

- Then a coach reviews the experiment and either accepts the team's rating or works with the team to adjust the rating (up or down).

- At any time, a team may revisit an experiment, do further work, and change the rating. Again, with review/approval/adjustment by a coach.

We've set up Jira as follows:

- Custom issue scheme to support our Experiment/Sub-experiment issue types

- Custom statuses and workflow scheme to support our 0, 1, 2, 3 ratings, with the ability to transition an issue between any of the statuses.

Then, when setting up a project, we use Software-Scrum, and set up the board with five columns + backlog: Backlog, ToDo, 0, 1, 2, 3. Because nothing is ever "done", but we need a done column, we let 3 be Done. So Status 1=Workflow Step 1=ScrumBoard Column 1, and so on for all four statuses.

What we want this to yield is a board at the end of ~12 weeks with 24 issues spread amongst the columns (hopefully more in the 2 and 3 columns than than the 0 and 1 columns). But Jira doesn't really like In Progress issues, so we can't seem to get the sprints or the boards to behave: issues that haven't been started don't always appear in the ToDo column at the start of the sprint, and issues that have moved to one of the In Progress columns fall into the next sprint when you complete the prior sprint, since technically those are still in progress.

Help? Is there any way to achieve the above, perhaps using a different configuration or different kind of board? We need the issues to be persistent in their assigned columns so that the teams get a visual representation of where they are as they perform their experiments, and so that we can report across teams and see how they are all doing.