Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
SEPG Conference Amsterdam 2006 Developing Metrics for agile projects compatible with CMMI Graham Collins, UCL [email protected] Research supported by Deutsche Bank Introduction UCL’s research and projects The problems and requests EV (Earned Value) Agile practices Combining EV and agile is it possible? Developing suitable metrics compatible with CMMI What works, and what reduces the overheads Is this approach likely to lead to CMMI Level 5? Key changes to projects Background - Requests We would like to use earned value ‘We would like to predict the outcome of project end date, cost and value’ As a project manager I cannot be there all the time What is the best way to measure progress with earned value? ‘As developers we would like to experiment with some agile approaches’ Your team will be working on other projects as well We are hoping to achieve metrics suitable for CMMI level 3 and higher… The Project Style Quadrant TheDeath Death March March Project Style Quadrant high Kamikaze Mission Impossible Suicide Ugly Happiness low low Chance of success high Edward Yourdon, Death March:The complete Software Developer’s guide to surviving ‘Mission Impossible’ projects, 1997 Prentice Hall Earned Value - Definition ‘The value of the useful work done at any given point in a project. The value of completed work expressed in terms of the budget assigned to that work. A measure of project progress. Note: The budget may be expressed in cost or labour hours’ APM (Association of Project Management) BoK 2006 Earned Value – Graphical Representation Planned Actual Cost or value Earned Value Time Planned (BCWS = Budgeted Cost of Work Scheduled) Actual (ACWP = Actual Cost of Work Performed) Earned Value (BCWP = Budgeted Cost of Work Performed) CPI = Cost Performance Index = BCWP/ACWP (or Earned/Actual) SPI = Schedule Performance Index = BCWP/BCWS (or Earned/Planned) note this is SPIc i.e. cost. Use of Variance Variance (Monthly) 1.20 1.10 1.00 Value 0.90 0.80 cpi 0.70 spi 0.60 0.50 0.40 0.30 1 2 3 4 Months 5 6 spic is used in this situation Earned Value compared to Agile Process Planning Earned Value Agile Development Based on predictive planning Estimates effort, cost and completion date End-to-end value tracking Adaptive planning Iteration to iteration tracking Predication of the next iterations effort Schedule most of the activities Adaptation to unpredictable events is problematic. Changes may require the planned to be revised or baselined Near the beginning, it is not always possible to schedule. Time based iterations allow initial estimate of duration which can be revised through the adaptive driven build-feedback cycle Estimates based on past performance Estimates are based on progress being made (velocity) Change rates often low Unpredictable change the norm Small variations in early measurements of cost and time at the start the project give wide variation in forward predications Unknown team development rates Some organisations do not chart progress for an initial period Progress is tracked immediately Earned value well established Prioritization of the value of user stories No earned value approaches in methods CMMI- Process Areas Category Process Area Project Management Project Planning Project Monitoring and Control Quantitative Project Management Support Process and Product Quality Assurance Causal Analysis & Resolution ‘Project planning is a necessity for success, Yet it still ranks on most surveys within the top three or four problem areas leading to failure.’ Tony Ciorra, Planner’s Progress, Project, APM May 2006 CMMI Comparative Advantages Continuous Representation Staged Representation Grants explicit freedom to select the order of improvement that best meets the organization’s business objectives and mitigates the organisation’s areas of risk Enables organisations to have a predefined path Enables increased visibility of the capability achieved in each individual process area Focuses on a set of processes that provide an organization with a specific capability that is characterized by each maturity level Provides a capability-level rating that is used primarily for improvement in an organisation and is rarely communicated externally Provides a maturity-level rating that is often used in internal management communication, statements external to the organization, and during acquisitions as a means to qualify bidders Allows improvements of different processes to be performed at different rates Summarizes process-improvement results in a simple form – a single maturity-level number Reflects a newer approach that does not yet have the data to demonstrate its ties to return on investment Builds on a relatively long history of use that includes case studies and data that demonstrate proved return on investment Agile Manifesto Individuals and interactions over processes and tools Working software over comprehensive documentation Customer collaboration over contract negotiation Responding to change over following a plan That is, while there is value in the items on the right, we value the items on the left more Several agile projects have achieved CMMI level 3, example David Anderson, Stretching Agile to fit CMMI Level 3, Agile Conference 2005 The Agile Principles www.agilealliance.com Our highest priority is to satisfy the customer through early and continuous delivery of valuable software Agile processes promote sustainable development Welcome changing requirements, even late in development. Agile processes harness change for the customer competitive advantage The sponsors, developer, and users should be able to maintain a constant pace indefinitely Deliver working software frequently, from a couple of weeks to a couple of months, with a preference to the shorter time scale Continuous attention to technical excellence and good design enhances agility Business people and developers must work together daily throughout the project Simplicity - the art of maximizing the amount of work done – is essential Build projects around motivated individuals. Give them the environment and support they need, and trust them to get the job done. The best architectures, requirements, and designs emerge from self-organizing teams The most efficient and effective method of conveying information to and within a development team is face-to-face conversation At regular intervals the team reflects on how to become more effective, then tunes and adjusts its behaviour accordingly Working software is the primary measure of progress Iterative Development (Bittner-Spence) 1. 2. 3. 4. 5. 6. Agree with the team the objectives for the iteration, including evaluation criteria, timescales, and constraints Agree on a plan for how the team will achieve the objectives Execute the plan Assess the achievements of the team against the initial set of objectives and evaluation criteria Assess the impact of the iteration’s results on the project as a whole Start the next iteration. What is iterative development? Part 3: The management perspective 15 May 2005 www-128.ibm.com/developerworks/rational/library/may05/bittnerspence/index.html Fundamental shift in measurement 100% Progress ( % complete measured in scenarios 0% Iteration 1 coded 2 3 tested 4 Tested & Passed Developer Perspective Developers are less interested in the business value, benefits realization and return on investment They work on a small number of requirements or change requests from their list of outstanding work They anticipate a decreasing number of requirements and change requests as the product is developed Outstanding requirements and change requests is termed the product backlog The developer will therefore be aware of progress via work completed, product backlog and new work allocated. User Satisfaction driving Development Release User satisfaction Release planning Iteration User satisfaction Iteration planning Development Increment Iteration Planning (Goal identification, story selection, task estimation, team commitment) Project teams need to adopt some attributes What is the purpose? Community of practice Project team To develop members’ capabilities; to build and exchange knowledge To accomplish a specified task What holds it together? Passion, commitment, and identification with the group’s expertise The project’s milestones and goals Adapted from: Communities of Practice: The organizational Frontier, Etienne C. Wenger and William M. Snyder, Harvard Business Review p139-145 Jan-Feb 2000 Rate of work - velocity Velocity 50 45 40 Story Points 35 30 25 20 15 10 5 24 /0 2/ 06 17 /0 2/ 06 10 /0 2/ 06 03 /0 2/ 06 27 /0 1/ 06 20 /0 1/ 06 13 /0 1/ 06 06 /0 1/ 06 0 Velocity, gives an indication of the average rate of work and also a comparison of planned against delivered, each iteration story points weeks mR 24/02/2006 20 17/02/2006 10/02/2006 03/02/2006 27/01/2006 20/01/2006 13/01/2006 06/01/2006 story points Individuals and Moving Range (XmR) Charts Velocity 60 50 UNPLx 40 30 UCLR 10 0 Control Limits for XmR Charts k sequential measurements provide k-1 =r (two-point) moving range values ith moving range = mRi = │Xi+1 – X i │where integer i is 1 ≤ i ≤ k – 1 __ i=r Individuals average moving range =mR = 1 ∑ mRi r i=1 _ __ _ __ Upper Natural Process Limit =UNPLx= X + 3mR = X + 2.660mR d2 _ i=k Centerline = CLx = X = 1 ∑ Xi (average of individual values) k i=1 _ __ _ __ Lower Natural Process Limit =LNPLx= X - 3mR = X - 2.660mR d2 ___ Centerline or average moving range =UCLR = mR __ __ Upper Control Limit for moving range =UCLR= D4mR = 3.268mR __ Sigma for individual values = sigmax (σ) = mR d2 When n=2 d2 =1.128and D4 =3.268 (from Dispersion and Bias factor tables) ‘Under Control’ Velocity 60 UNPLX story points 50 40 30 20 LNPLX 10 24/02/2006 17/02/2006 10/02/2006 03/02/2006 27/01/2006 20/01/2006 13/01/2006 06/01/2006 0 weeks story points Velocity measures of work rate are useful in that estimates of the next iteration can be planned in a rolling process The use of σ variation is supportive in this aim Automated colour coding (Red Amber Green) can be used to show condition requirements ‘Weighted average’ Story points Velocity 30 25 weighted average 20 15 Weighted points 0 1 0 5 2 10 18 3 54 14 4 56 24 5 120 27 6 162 14.66667 3.5 19.14286 UNPLx 35 story points count centreline (CLx) 10 5 0 1 2 3 4 5 6 iterations 50 45 40 35 30 25 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 24/02/2006 17/02/2006 10/02/2006 03/02/2006 27/01/2006 20/01/2006 13/01/2006 06/01/2006 Story Points ‘Burn-down’ Story Points Remaing 350 300 250 200 150 100 50 0 With the appropriate metrics we can improve Acceptance Test Data 100 90 80 ATs 70 60 Acceptance Tests 50 Failing ATs 40 Passing ATs 30 20 10 0 1 2 3 4 5 6 7 8 Iteration 9 10 11 12 Use of Multipliers Iterations Completed Low Multiplier High Multiplier 1 0.6 1.60 2 0.8 1.25 3 0.85 1.15 4 or more 0.90 1.10 Multipliers for estimating velocity based on number of iterations completed from Cohn 2006 Charts and Metrics Velocity and Burn-down Cumulative acceptance tests Inventory Failing Passing Cumulative Issue Charts Backlog - Active issues (which show inventory line) Resolved issues Closed issues Earned Value EV progress charts Performance via cpi and spi Cpi and spi combined with control charts Earned Value EV can be applied to estimates of agile projects - this is complex if more stories are added as the work progresses EV may need to be shown to senior managers - who are used to EV figures, or comparison to other projects where EV figures have been tracked EV estimates can be accurate - story points tend to remain static in an iteration when the process is understood by managers and developers. When additional stories are added, stories with lower business priority level may be dropped to compensate and keep the work load (story points) similar. Business Value More importantly business value (or contribution) should be considered and evaluated Often units of measure such as story points can be valued as 0.5 or 1.0 units The key issue in agile project management is to continually assess with the client the most important work that should be done. Story number ‘Business Value’ Story Points ‘Points earned’ Planned (developer hours) Actual (logged hours) EV (earned value) 1 3 10 10 100 120 100 2 2 8 8 60 60 60 3 2 4 4 60 80 60 4 1 2 0 20 0 0 total 8 24 22 240 260 220 Conclusions EV can be combined with agile reporting The most useful approach to achieve process improvement is via understanding of process control, the basis for CMMI The use of acceptance tests is the most useful approach for both methods and is the basis for metrics outlined EV remains a viable approach where story points work load is maintained EV is useful for reporting to senior managers and at a programme and project level, but even with major scope changes, re-planning can give some useful insights The process used in agile project management is critical, that this should be adhered to in the team. It needs to be explained to managers that the team are trying to achieve maximum business value per iteration, and that the final goal and plans evolve during the process Tracking suitable metrics and understanding variation is the key to better estimation and process improvement – CMMI assessments can help achieve this goal. Key Project Changes 1. Pair reporting - valuing individuals and team, moving towards self-determining teams 2. Acceptance testing - working software 3. Ensuring Business Value – continual prioritisation, estimation and understanding there is a cost to development 4. Measuring progress over shorter time periods meeting the needs of process improvement CMMI, velocity tracking in agile methods, and better EV planning Individuals and interactions Working software over processes and tools over comprehensive documentation Customer collaboration over contract negotiation Responding to change over following a plan