There are many reasons why CMMI is difficult to implement in small organizations and small projects -- the fixed costs of establishing the necessary infrastructure; the large number of roles which must be filled by a limited number of people; the quantity of information that must be absorbed to properly interpret the model. Similar problems are experienced when applying the CMMI to short duration projects.
This tutorial will outline the challenges in applying CMMI in small settings, and present practical strategies for overcoming them. Specific techniques for infrastructure, adoption, and appraisals will be presented. In addition, a guide for interpreting each CMMI practice in small settings will be provided.
Applicability of CMMI for Small to Medium Enterprises
1. 3 CMMI® Views Rick Hefner Director, Process Assurance Northrop Grumman Corporation Southern California SPIN 4 June 2010
2. Background Many published results show improved cost and schedule performance from adopting CMMI® Despite these results, there is still community debate over the value of CMMI®, and whether CMMI® ratings provide sufficient guarantees of program performance. This program will explore three factors contributing to the confusion: Inaccurate CMMI® ratings Over-estimating the benefits that CMMI® provides a customer Contractors not living up to their CMMI® rating 2
3. Agenda Underlying CMMI® Principles CMMI®relationship to productivity, predictability and speed Does CMMI Benefit the Customer? How Projects Fail How to Get Contractors to Live Up to Their CMMI Ratings 3 SM SCAMPI, SCAMPI Lead Appraiser, and SEI are service marks of Carnegie Mellon University. ® Capability Maturity Model Integration and CMMI® are registered in the U.S. Patent & Trademark Office.
7. Many Approaches to Solving the Problem Which weaknesses are causing my problems? Which strengths may mitigate my problems? Which improvement investments offer the best return? People Business Environment Management Structure Product Tools Methods One solution! Technology Process 5
8. Approaches to Process Improvement Data-Driven (e.g., Six Sigma, Lean) Clarify what your customer wants (Voice of Customer) Critical to Quality (CTQs) Determine what your processes can do (Voice of Process) Statistical Process Control Identify and prioritize improvement opportunities Causal analysis of data Determine where your customers/competitors are going (Voice of Business) Design for Six Sigma Model-Driven (e.g., CMM®, CMMI®) Determine the industry best practice Benchmarking, models Compare your current practices to the model Appraisal, education Identify and prioritize improvement opportunities Implementation Institutionalization Look for ways to optimize the processes 6
9. What Is the CMMI® Trying to Achieve? A model is a simplified representation of the world. Capability Maturity Models (CMM®s) contain the essential elements of effective processes for one or more bodies of knowledge. These elements are based on the concepts developed by Crosby, Deming, Juran, and Humphrey. -Introduction, CMMI® CMMI® provides a model of industry best practices Following these practices has shown to produce software and systems faster, better, and cheaper, when properly applied The main benefits cited by CMMI® users are: More predictable adherence to budgets and schedules Reduced re-work (which can reduce cost and schedule) Reduced risk 7
12. Agenda Underlying CMMI® Principles Does CMMI Benefit the Customer? Cost of implementing CMMI-compliant processes Timelines for impacting program performance Practical tips and techniques for realizing the benefits How Projects Fail How to Get Contractors to Live Up to Their CMMI Ratings 9 SM SCAMPI, SCAMPI Lead Appraiser, and SEI are service marks of Carnegie Mellon University. ® Capability Maturity Model Integration and CMMI® are registered in the U.S. Patent & Trademark Office.
13. CMMI® Provides Several Related Benefits 10 Project Performance Organizational Performance Institutionalization Quality/Rework Rick Hefner, “Achieving the Promised Benefits of CMMI,” CMMI Technology Conference & User Group, Denver, CO, 14-17 Nov 2005
14.
15. Project managers (not organizational staff) must be responsible for implementing the improved processes
25. QA can be the organization’s strongest tool – if they are focused!14
26. Benefits The typical benefits are: Reduced cost Faster schedules Greater productivity Higher quality Increased customer satisfaction Over 40 published studies on the benefits of SW-CMM® DoD DACS website: http://www.thedacs.com/databases/roi/ Similar results starting to be seen for CMMI® “Demonstrating the Impact and Benefits of CMMI: An Update and Preliminary Results,” Software Engineering Institute, CMU/SEI-2003-SR-009, Oct 2003 http://www.sei.cmu.edu/cmmi/results/results-by-category.html 15
27. Typical CMMI Benefits Cited in Literature Reduced Costs 33% decrease in the average cost to fix a defect (Boeing) 20% reduction in unit software costs (Lockheed Martin) Reduced cost of poor quality from over 45 percent to under 30 percent over a three year period (Siemens) 10% decrease in overall cost per maturity level (Northrop Grumman) Faster Schedules 50% reduction in release turnaround time (Boeing) 60% reduction in re-work following test (Boeing) Increase from 50% to 95% the number of milestones met (General Motors) Greater Productivity 25-30% increase in productivity within 3 years (Lockheed Martin, Harris, Siemens) Higher Quality 50% reduction of software defects (Lockheed Martin) Customer Satisfaction 55% increase in award fees (Lockheed Martin) 16
28. Cost vs. Benefit Both theoretical models and industry data suggests that CMMI-compliant projects achieve a cost reduction of 10% per level, i.e., Level 3 is 20% cheaper than Level 1 The key is reducing rework Knox Model – Theoretical Benefits 17 COCOMO predicts similar benefits based on current industry data
29. When Good Organizations Go Bad Some organizations are driven to achieve a maturity level only for it’s marketing value Focus on passing the appraisal, not understanding and deciding among possible interpretations Improvement goals are not set realistically (“Level 5 in ’05”) Practitioners/customers perceive CMMI as more expensive Only some of the projects participate in the improvement effort The remaining projects don’t implement Only some of the projects get appraised People don’t learn or become proficient in the new behaviors Insufficient resources (e.g., training, QA, metrics, consultants) Benefits are not realized because projects do not start up effectively Management doesn’t enforce using processes on new programs Rick Hefner, “CMMI Horror Stories: When Good Projects Go Bad,” Software Engineering Process Group Conference , 6-9 March 2006 18
30.
31. How Does Level 4 & 5 Benefit the Customer? Better Products and Services Produced Faster And Cheaper Rick Hefner, “How Does High Maturity Benefit the Customer?,” Systems & Software Technology Conference, 18-22 April 2005 20
32. The Project Manager’s Dilemma at Level 3 I want to use the organization’s standard process, but… … Does it’s performance and quality meet my customer’s expectations? … If not, how should I tailor the process? 21
33. Understanding the Process 22 Managing by Variation How many errors are typically found in reviewing an interface specification? Useful in evaluating future reviews Was the review effective? Was the process different? Is the product different? Average Expected Variation Corrective and preventative actions
34. Typical Choices in Industry Most customers care about: Delivered defects Cost and schedule So organizations try to predict: Defects found throughout the lifecycle Effectiveness of peer reviews, testing Cost achieved/actual (Cost Performance Index – CPI) Schedule achieved/actual (Schedule Performance Index – SPI) 23
35. What Can a Level 4 Organization Do? Determine whether processes are behaving consistently or have stable trends (i.e., are predictable) Identify processes where the performance is within natural bounds that are consistent across process implementation teams Establish criteria for identifying whether a process or process element should be statistically managed, and determine pertinent measures and analytic techniques to be used in such management Identify processes that show unusual (e.g., sporadic or unpredictable) behavior Identify any aspects of the processes that can be improved in the organization's set of standard processes Identify the implementation of a process which performs best 24
36. Lessons Learned Based on over 20 Northrop Grumman CMMI Level 5 organizations Six Sigma is an enabler for higher maturity Focus on data, measurement systems, process improvement Tying improvements to business goals Tools and methods support the Level 4/5 analysis tasks Level 3 metrics, measurement processes, and goal setting are generally inadequate for Levels 4 and 5 Better definitions of the measures Lower level metrics of lower level subprocesses Having all the tools at Level 5 gives you the insight to manage each project the way the customer needs it to be managed 25
37. Agenda Underlying CMMI® Principles Does CMMI Benefit the Customer? How Projects Fail Start up problems Appraisal inaccuracies How to Get Contractors to Live Up to Their CMMI Ratings 26
38. Where Could Problems Arise? The projects within the organization may not live up to the capability Start-up problems with planning, subcontractors, and infrastructure Problems with staffing, either as the prime or with subcontractors Differences in domain experience Back-sliding The appraisal results may not be an accurate reflection of the organization’s capability Sampling bias Appraisal inaccuracies Organization’s inability to immediately apply their appraised processes Note that all of these problems are equally possible with both the staged and continuous representations 27
39. The First Three Months: Essential Project Start-Up Activities Many process-related problems arise in the first few months of a project New relationships are established Personnel changes and shortfalls Pressure to produce quickly Gaps between the planned processes and what was bid If a project is going to live up to the organization’s process capability, it is essential to fully implement the processes from the beginning Processes should be defined during the proposal, by tailoring the organization’s standard process Estimates should be based on historical data from the organization’s measurement repository Process assets (e.g., templates) should support detailed planning to ensure consistency with the organization’s best practices Evidence reviews should be used to ensure CMMI compliance 28
40. Preventing Back-Sliding 29 The CMMI generic practices ensure that processes are institutionalized – sustained over time The approach for implementing the generic practices must reflect: Efficiency Effectiveness Applicability to ALL projects Frequent appraisals can be used to assess the effectiveness of the institutionalization Geoff Draper and Rick Hefner, “Applying CMMI Generic Practices with Good Judgment,” SEPG Conference, 2004.
41. Sampling Bias The size and number of instantiations investigated should be selected to form a valid sample of the organizational unit to which the results will be attributed. - SCAMPI A Method Description Document The Lead Appraiser is permitted to select sample projects as “representative” of the organization as a whole Little guidance in the MDD Wide variation among Lead Appraisers If an organization is only interested in a good appraisal result, they will appraise large organizations with a handful of samples, and/or exclude/hide inferior projects This potential abuse exist with both staged and continuous representations The remaining projects don’t implement Only some of the projects get appraised 30
42. Organizational Sampling An organization with 50 projects at multiple sites may select 4-5 sample projects Are the appraisal results representative of the organization? Sampled projects 31
43. Appraisal Inaccuracies Methodology SCAMPI A appraisals are the only approach that provides benchmark quality appraisal results SCAMPI B, C, and other appraisal methods may be useful, but they are not designed to provide the same accuracy Appraiser Skill There is wide variation in appraiser skill, experience and insight Although appraisal experience is a crucial contributor to accuracy, the appraisal methods do little to ensure sufficient experience There is also wide variation in how the model is interpreted Appraiser Independence Appraiser independence in needed to ensure unbiased results It is difficult to establish a completely independent situation 32
44. Fiction or Non-Fiction: How to Read Appraisal Results for Fun and Profit The ADS is a summary statement describing the appraisal results that includes the conditions and constraints under which the appraisal was performed. It contains information considered essential to adequately interpret the meaning of assigned maturity level or capability level ratings. - SCAMPI A Method Description Document The Appraisal Disclosure Statement (ADS) provides keys to assessing an appraisal’s accuracy Organizational unit appraised (the unit to which the ratings are applicable and the domains examined) Appraisal team leader and appraisal team members and their organizational affiliations Process areas rated and process areas not rated Dates of on-site activity Not included - sampling approach or percentage of projects sampled 33
45. How to Write a Better RFP Acquirers seeking to ensure that the proposed project will implement mature practices should request the following: SCAMPI A Appraisal Disclosure Statement Organizational unit appraised Appraisal team leader affiliation Process areas rated and not rated Dates of on-site activity Explanation of sampling approach used in appraisal Approach to be used to ensure proper project start-up Data to demonstrate the speed with which new projects adopt and execute the organization’s processes Approach to be used to prevent back-sliding 34
46. Agenda Underlying CMMI® Principles Does CMMI Benefit the Customer? How Projects Fail How to Get Contractors to Live Up to Their CMMI Ratings Contenders and Pretenders 35
47. Background There is a marked difference between organizations that truly want to implement CMMI®, and those who simply want a “certificate” Contenders invest time and energy on understanding the industry best practices in the model, fitting them to their projects and organization, and improving their effectiveness and efficiency Pretenders simply do enough to convince an appraiser to give them the maturity level -- along the way, they de-motivate their staff with bureaucratic processes, disappoint their customers with inconsistent performance, and generally give the model a bad name 36
48. Where Could Problems Arise? Assuming the contractor’s CMMI® rating is accurate, and applicable to the team doing the work, where could problems arise? Areas outside of the CMMI® Start-up problems Back-sliding 37
49. Areas Outside of the CMMI® 38 Process Technology Domain-specific Maturity Tools People Domain knowledge Sufficient quantity Motivation
50. Top Five System Engineering Issues Lack of awareness of the importance, value, timing, accountability, and organizational structure of SE on programs Adequate, qualified resources are generally not available within Government and industry for allocation on major programs Insufficient SE tools and environments to effectively execute SE on programs Requirements definition, development and management is not applied consistently and effectively Poor initial program formulation 39 “Top Five Systems Engineering Issues In Defense Industry”, NDIA Systems Engineering Division Task Group Report, Jan, 2003
51. Top Software Engineering Issues The impact of requirements upon software is not consistently quantified and managed in development or sustainment Fundamental system engineering decisions are made without full participation of software engineering. Software life-cycle planning and management by acquirers and suppliers is ineffective. The quantity and quality of software engineering expertise is insufficient to meet the demands of government and the defense industry. Traditional software verification techniques are costly and ineffective for dealing with the scale and complexity of modern systems. There is a failure to assure correct, predictable, safe, secure execution of complex software in distributed environments. Inadequate attention is given to total lifecycle issues for COTS/NDI impacts on lifecycle cost and risk. 40 “Top Software Engineering Issues In Defense Industry”, NDIA Systems Engineering Division and Software Committee, Sep 2006
52. Start-Up Issues Project Planning starts in the proposal phase, is refreshed at contract start, and re-occurs throughout the project lifecycle Contenders extend their CMMI practices to proposal teams and re-planning efforts Pretenders focus on contract start Costs and schedules defined at proposals may be immature and overly-aggressive Re-planning may be ad hoc Mature estimates may also be overruled by business interests 41
56. Keys to Success Ask suppliers to show how they extend the CMMI practices to proposal activities Request planning documents with the proposal During re-planning, ask suppliers to show how they performed the CMMI practices 45
57. Back-Sliding: A Failure of Institutionalization Institutionalization: The ingrained way of doing business that an organization follows routinely as part of its corporate culture. - CMMI-DEV v1.2 GG 2 Institutionalize a Managed Process GP 2.1 Establish an Organizational Policy GP 2.2 Plan the Process GP 2.3 Provide Resources GP 2.4 Assign Responsibility GP 2.5 Train People GP 2.6 Manage Configurations GP 2.7 Identify and Involve Relevant Stakeholders GP 2.8 Monitor and Control the Process GP 2.9 Objectively Evaluate Adherence GP 2.10 Review Status with Higher Level Management GG 3 Institutionalize a Defined Process GP 3.1 Establish a Defined Process GP 3.2 Collect Improvement Information When mentioned in the generic goal and generic practice descriptions, institutionalization implies that the process is ingrained in the way the work is performed and there is commitment and consistency to performing the process. An institutionalized process is more likely to be retained during times of stress. 46
59. Organizational Support Contenders Fully support the CMMI ® -based improvement program by providing training, templates, tools, process assets libraries, measurement repositories and other work aids focused on improving the ability of practitioners to competently adopt the model Pretenders Largely ignore organizational support, often to save money Where required by the model, they establish process asset libraries and measurement repositories, but they are largely shelfware 48
60. 49 Organizational Infrastructure Required for CMMI® Level 3 Policies, Processes,Templates & Tools Process Group Process Improvement Training Program Best-Practice Libraries Audits & Appraisals Measurement Repositories Predictive Modeling Communications Developing and maintaining mature processes requires significant time and investment in infrastructure
61. Organizational Culture A pattern of shared basic assumptions that the group learned as it solved its problems of external adaptation and internal integration, that has worked well enough to be considered valid and, therefore, to be taught to new members as the correct way you perceive, think, and feel in relation to those problems. Artifacts The practices that can be observed in such areas as dress code, leadership style, communication processes Espoused values The elements the organization says it believes in, the factors that it says influence the practices in which it engages Basic underlying assumptions Unstated beliefs the organization has come to accept and abide by Organizational Culture & Leadership, Edgar H Schein, used with permission 50
62. Management Commitment and Support Understands the key messages Is willing to take actions to reinforce them Provides resources to support/sustain process improvement efforts Sets expectations that essential project functions will be funded and processes will be followed Project planning, estimation, tailoring, CM, QA, etc. Supports process improvement and sustainment, rather than passing appraisals Rewards mature processes development and sustainment rather than individual heroics 51 Rick Hefner, “Sustaining CMMI Compliance,” 2006 CMMI Technology Conference and User Group
63. Keys to Success Ask suppliers to show how they perform the CMMI generic practices When problems occur, ask why the CMMI practices were not effective in sustaining the desired behavior, and what will be done to prevent future problems 52
64. Summary There is a marked difference between organizations that truly want to implement CMMI®, and those who are simply try to get a “certificate” By discussing the differences, we hope to help the CMMI® community the true value of CMMI® 53