System Stability and Sustainability Data for 295 Agile Systems

Purpose: This data set is a collection of 295 "Agile Systems" performed by an European mid-tier software development consulting firm between June 2008 and August 2023.  The systems were a mixture of Scrum, Kanban, LeSS, SAFe and other Agile-type approaches to work, including Agile In Name Only (AINO). The data set is the output from a script that analysed the raw data extracted from Jira with each Agile System corresponding to a Jira "Project".

The script performed the following actions:

1. Read the raw data for each system. (this is not included in this dataset)

2. Remove Epics and Sub-tasks so all PBIs are approximately the same size.

3. Remove any cancelled (and similar) PBIs so only PBIs analysed were done by a team.

4A. Identify the Arrival rate (average rate at which PBIs were created over the duration of the system, Lambda) and the Service rate (average rate at which PBIs were moved to Done over the duration of the system, Mu). Also calculate the system size (total arrivals versus total services) at the end of the duration of the system. 

4B. Calcuate the Stability Metric (Mu / Lambda) and Inventory Days (Total System Size / Mu). 

4C. Classify the system to a Strategy:

If Stability Metric < 1 and Inventory Days < 30, Strategy : Start-Up 

If Stability Metric >= 1 and Inventory Days < 30, Strategy : Plan-Up

If Stability Metric >= 1 and Inventory Days >= 30, Strategy : Catch-Up

If Stability Metric < 1 and Inventory Days >= 30, Strategy : Scale-Up

4D. Identify when the first arrival, first service, last arrival and last service.

4E. Calculate t0 - the time required for 5 PBIs to be completed. 

4F. Calculate the number of PBIs that occurred outside 6am to 6pm Mon-Fri. These arrivals and services were for our research purposes "Unsustainable".

4G. This data is under the Timeset = "All" for each system.

5. Repeat step 4 but for just for the time period leading up to t0. This data is under the Timeset = "0" for each system.

6. Recursively repeat step 4 by incrementing the period under investigation after t0 by 7 days until the higher of last arrival and last service is reached. Each period is given an incremented timestep, 1,2,3 and so on. This data is under a numbered Timeset each system. 

To use the data set: 

We recommend importing this data into a tool that faciliates pivot tables.

It is possible to compare across systems at each timepoint or review time histories across systems.