220 likes | 330 Views
Rigorous Support for Flexible Planning of Product Releases — A Stakeholder-Centric Approach and its Initial Evaluation. Ville Heikkilä Anas Jadallah Kristian Rautiainen Günther Ruhe. Ville Heikkilä, Kristian Rautiainen Aalto University School of Science and Technology
E N D
Rigorous Support for Flexible Planning of Product Releases —A Stakeholder-Centric Approach and its Initial Evaluation Ville Heikkilä Anas Jadallah Kristian Rautiainen Günther Ruhe
Ville Heikkilä, Kristian Rautiainen Aalto University School of Science and Technology Software Process Research Group Anas Jadallah, Günther Ruhe University of Calgary SE Decision Support Laboratory Ville Heikkilä
Contents • Introduction and research problem • Process description and case study • Key findings and discussion Ville Heikkilä
Introduction • Release planning by PM or PO • Too many stakeholders for F2F • Stakeholder CentricRelease Planning (SCERP) Ville Heikkilä
Objective function F(x) = k=1…K(k) [n: x(n)=k WAP(n)] Releaseplanningproblem f(2) f(3) f(N) Features . . . k(1) k(2) Releases (1) = 9 (2) = 5 x(1)=1, x(2)=1 x(3)=2, x(N)=2 . . . S(1) S(2) S(q) Stakeholders Cap(k,r) f(1) λ(1) = 7 λ(2) = 2 λ(q) = 4 Resources r(1) r(2) r(3) Ville Heikkilä
Case study www.agilefant.org www.releaseplanner.com Roadmap Release Iteration Hartbeat Ville Heikkilä
SCERP Step 1: Selection of criticalstakeholders and pre-selectionof candidate features Step 2: Prioritization of features Step 3: Collective effort estimation Step 4: Calculation of optimizedrelease plan alternatives Step 5: Prioritization of alternativeplans 1 2 3 4 5 Ville Heikkilä
SCERP Selection of critical stakeholders and pre-selection of candidate features 1 2 3 4 5 Ville Heikkilä
Case study Selection of critical stakeholders and pre-selection of candidate features 1 2 3 4 5 • Done by the product owner • Two most important companies using Agilefant • 10 of 73 backlog items selected Ville Heikkilä
SCERP Prioritization of features 1 2 3 4 5 Ville Heikkilä
Case study Prioritization of features 1 2 3 4 5 • Stakeholder invitation • 19 of 33 participated in the end • Cumulative voting • Criteria: value, urgency and dissatisfaction • Voting done in ReleasePlanner • Time taken by voting wasrecorded Ville Heikkilä
SCERP Collective effort estimation 1 2 3 4 5 Ville Heikkilä
Case study Collective effort estimation 1 2 3 4 5 • Only 2 active developers • Collaborative effort estimation • Full time equivalent (FTE) developer Ville Heikkilä
SCERP Calculation of optimized release plan alternatives 1 2 3 4 5 Ville Heikkilä
Case study Calculation of optimized release plan alternatives 1 2 3 4 5 • Two next releases • Release weights 9 and 5 • Capacity 22 FTE-days • ReleasePlanner created 5 optimized release plan alternatives • Product Owner created a manual plan Ville Heikkilä
SCERP Prioritization of alternative plans 1 2 3 4 5 Ville Heikkilä
Case study Prioritization of alternative plans 1 2 3 4 5 • 10 stakeholders • 5 plans + manual plan • Match score • “Perfect match” (9) • “No match” (1) Ville Heikkilä
Plan optimization output Ville Heikkilä
Survey • 6-choice attitude scale • “I understood feature X” • Attitudes towards method • Free text field Ville Heikkilä
Key findings • Stakeholders understood the features and their priorities varied considerably • Stakeholders understood and liked the method • SCERP is time-efficient for the stakeholders • Optimized plans were more acceptable than the manual plan • List of candidate features was insufficient • Not enough difference between criteria Ville Heikkilä
Discussion • Does SCERP work? • Does SCERP scale? • Was the case success? Ville Heikkilä
Questions? Ville Heikkilä