Comments
Transcript
Paul Gerrard Managing Software Projects With g g j
Assurance with Intelligence Managing g g Software Projects j With Intelligence Paul Gerrard Gerrard Consulting 1 Old Forge Close Maidenhead Berkshire SL6 2RD UK e: [email protected] w: http://gerrardconsulting.com t: +44 (0)1628 639173 Slide 1 Assurance with Intelligence Paul Gerrard Paul is the founder and Principal of Gerrard Consulting, a services company focused on increasing the success rate of IT-based projects for clients. He has conducted assignments in all aspects of Software Testing and Quality Assurance Assurance. Previously, Previously he has worked as a developer, designer, project manager and consultant for small and large developments using all major technologies and is the webmaster of gerrardconsulting.com and several other websites. Paul has degrees from the Universities of Oxford and London, is Web Secretary for the BCS SIG in Software Testing (SIGIST), Founding Chair of the ISEB Tester Qualification Board and the host/organiser of the UK Test Management Forum conferences. He is a regular speaker at seminars and conferences in the UK, continental Europe and the USA and was recently awarded the “Best Best Presentation of the Year” prize by the BCS SIGIST. Paul has written many papers and articles, most of which are on the Gerrard website. With Neil Thompson, Paul wrote “Risk-Based E-Business Testing” – the standard text for risk-based testing. Slide 2 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 1 Assurance with Intelligence Agenda Why Projects Fail Introduction to Project Intelligence Results Chain Analysis PI Strategy Test Strategy Completion Goal, Risk and Requirements Based Testing Close Slide 3 Assurance with Intelligence Why Projects Fail © 2003-7 Gerrard Consulting Limited Version 1.0 Page 2 Assurance with Intelligence Why projects fail Won’t bore you with yet another survey of IT projects that fail - take it for granted, most do But why is this? Hundreds of modes of project failure – all very well documented But a large proportion of project failures can be traced to poor decision making: - Constraints to decision makingg - “you y are not allowed to decide” - Decisions already made cannot be changed - Decisions that are ‘too big’ to make now – it’s simply too late Sometimes, the worst decisions are those that are never made at all. Slide 5 Assurance with Intelligence Decision-making when things go wrong Imagine a world where the project plan is perfect, and everything goes well There are few decisions to make except exit/acceptance reviews - These would be easy to make, wouldn’t they? But plans aren’t perfect and things don’t go to plan Most decision-making is required when things don’t go to plan - Late delivery – we are late or expect to be late - Poor P quality lit or stability t bilit – deliverables d li bl nott meeting ti requirements, i t ffailil tto stabilise or are defective - Technical challenges that cannot be met – technology doesn’t work, can’t demonstrate working systems, cannot achieve objectives with proposed solution. Slide 6 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 3 Assurance with Intelligence Decision-making dependent on good information When called upon to make a call on delays, quality or solution viability… - Stakeholders and project managers usually lack the right information at the right time - Most of the information actually required is TEST EVIDENCE Early and continuous review and testing will highlight - Items not ready (we can’t test them) - Poor quality (they are buggy) - Developers p that cannot deliver I call this evidence Project Intelligence THE TEST STRATEGY IS A STRATEGY FOR GATHERING PROJECT INTELLIGENCE. Slide 7 Assurance with Intelligence Our project management approach is critically flawed The way we plan projects and monitor/manage progress is flawed Project plans set out a prediction of the future A network of interdependent activities aligned with a calendar We identify activities, activities costs and timelines We monitor activities, spend and the calendar We act when progress doesn’t meet the plan. Slide 8 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 4 Assurance with Intelligence But STOP! Imagine we planned a journey this way We monitor progress by checking… - The time – “it’s 12.45pm p – we must be half way” y - The fuel gauge – “we are nearly out of gas – we must be there!” This is nonsense of course We are monitoring INPUTS, NOT OUTPUTS We should be monitoring ACHIEVEMENTS, not ACTIVITY What are achievements? How do we monitor them? Slide 9 Assurance with Intelligence The measurement of achievement The achievements of most significance are But wait! The REAL measure of achievement is the acceptance decision, based on evidence Deliverables (on time and of good quality) Decisions made (on the acceptability of deliverables) Requirements Designs Code S Systems Etc… Testing provides that evidence. Slide 10 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 5 Assurance with Intelligence Pawnbroker model of IT development • Developers usually complain that requirements q change g • Were the users focused on the Required Designed Used! Delivered business goal? • Developers focus on technical deliverables, as usual • But who keeps them on the right track? • Required, designed, not delivered (cuts through deadline pressure?) • Not required, designed, delivered (an analyst's bright idea?) • Required, not designed, delivered (the user bribed a programmer?) Assurance Slide 11 with Intelligence Slide 11 Four-eyed plans Increasingly gy Inaccurate IT- focused Initial Plans The plan is a model of the project. The real project consists of people, organisation, goals and risks. Assurance Slide 12 with Intelligence © 2003-7 Gerrard Consulting Limited Version 1.0 Slide 12 Page 6 Assurance with Intelligence Testing provides the “intelligence” Our deliverables are intermediate project goals Testing can answer these questions Are they available? Are they acceptable? Do they meet requirements?? No tests completed? The deliverable isn’t available Test successes/failures indicate quality and whether deliverable meets requirements We manage product risks and goals, not activities: - We analyse W l the h risks i k associated i d with i h this hi d deliverable li bl We plan tests to address uncertainty We monitor tests planned, passed and failed. Slide 13 Assurance with Intelligence Poor contracts result in poor plans and poor project management Inputs to project plan for delivery The Contract The Requirements (business or technical level) The requirements may be good or bad But if the contract is poor or inflexible - Things get forgotten, underestimated Testing, risk management, reporting all suffer Project Management attention on timescales and costs (inputs) not on deliverables, quality, benefits (outputs) Acceptance process and criteria are loose, so plans are woolly when precision is most important. Slide 14 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 7 Assurance with Intelligence Introduction to Project Intelligence Slide 15 Assurance with Intelligence Testing throughout the life cycle: the W model Write Requirements Test the Requirements Logical Design Install Test the Design Physical Design Build System Test the Specs Code Slide 16 © 2003-7 Gerrard Consulting Limited Version 1.0 Build Software Acceptance Test System Test Integration Test Unit Test Assurance with Intelligence Page 8 Assurance with Intelligence Testing through the lifecycle Almost all project activities generate deliverables Project goals depend on deliverables Every deliverable has an associated test activity Testing provides the evidence of achievement Static tests: reviews, inspections, static analysis etc. Dynamic tests: component, link, system, acceptance tests etc. Of intermediate goals throughout a project Of final project goal (This is standard PRINCE approach, but is appropriate for all methodologies, by the way). Slide 17 Assurance with Intelligence Three types of risk Internal/Process Risk (Logistics, Planning and Control) External/Project Risk Management Domain Few, easy to understand, manage variances in planning and estimation, shortfalls in staffing, failure to track progress, lack of quality assurance and configuration management resource constraints, external interfaces, supplier relationships, contract restrictions Not usually under management direct control Usually under management control Product Risk Deliverables Practitioner Domain Many, hard to understand, manage The “problem of quality” Testers are mainly concerned with Product Risk Slide 18 © 2003-7 Gerrard Consulting Limited Version 1.0 complexity, design quality, coding quality, non-functional issues, test specifications. Assurance with Intelligence Page 9 Assurance with Intelligence What is a test? Deliverable Test Testing doesn’t change the product in any way. W do We d it i to gather h ‘intelligence’. Slide 19 Deliverable (unchanged) intelligence g Assurance with Intelligence Testing generates “project intelligence” If you don’t know where you are, a map won’t help If you don don’tt know the status of the deliverables deliverables, the best plan won’t help you manage your project A project is like driving off-road in the dark Testing provides the headlights for the journey Testing provides data on the status of deliverables and generates “PROJECT INTELLIGENCE” Stakeholders and management need intelligence to make decisions, command and control the project. Slide 20 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 10 Assurance with Intelligence Slide 21 Military Intelligence? Assurance with Intelligence Project Intelligence starts with goals What are our project goals? Our O goals l d depend d on the h deliverables d li bl How could our deliverables fail? Which are the key deliverables? What are the dependencies between them? Evidence required to be sure they won won’tt fail? How can I be sure our goals are met? - Evidence required to demonstrate they are met? Slide 22 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 11 Assurance with Intelligence Key sources of project intelligence The value of project intelligence depends on how well we achieve and communicate: 1. 2. 3. 4. Coverage – planned and achieved Product Risks – and the status of those risks Business Goals – and the status of those goal Test Status – passes, failures, incidents Time to deliver intelligence depends on the test process Willingness of the organisation to accept and act upon project intelligence is key. Slide 23 Assurance with Intelligence Based on established disciplines The PI approach uses a combination of established techniques (you might know one or two already!) - Results-Based Management Results Chain Analysis (modelling) Benefits Realisation Performance Measurements Software and System y Testingg Best Practices. Slide 24 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 12 Assurance with Intelligence PI Techniques Results Chain Analysis PI Strategy S Test Strategy Completion (Assessment and Testing) PI Reporting Slide 25 Assurance with Intelligence PI and the project lifecycle Results Chain Analysis PI Strategy Test Strategy Acceptance Goal Assessment Project Intelligence Planning Goal Assessment Project Planning Goal Assessment Project Initiation PI Management Stakeholder Involvement/Project Assurance/Governance Key: Project planning and initiation Project Intelligence Activities Development activities Review and Test activities Slide 26 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 13 Assurance with Intelligence Results Chain Analysis Results Chain Analysis Results Chain Analysis PI Strategy Test Strategy Acceptance Goal Assessment Project Intelligence Planning Goal Assessment Project Planning Goal Assessment Project Initiation RCA defines the relationship between ggoals,, activities and risks PI Management Stakeholder Involvement/Project Assurance/Governance Key: Project planning and initiation Project Intelligence Activities Development activities Review and Test activities Slide 28 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 14 Assurance with Intelligence Goal, Benefits and deliverables Technology Features Change in Work Practices Electronic communication with suppliers Web Closer working with suppliers Electronic entry System Integration Automatic reordering Portal System accessible to all requisitioners Project D li Deliverables bl Bar Coding Etc. etc Expedition and progressing facilities Electronic information service Supplier use and performance evaluation Improved supplier understanding More timely procurement More accurate deliveries More timely deliveries To reduce production downtime Project G l Goals Improved coordination between functions Increased time for buyers to add value Optimised selection of supplier to meet given need To reduce costs Improved negotiating position Etc. etc. Slide 29 Business Goals Benefits Assurance with Intelligence The Results Chain Programme/Project Management Inputs Activities Outputs Outcomes Impact Administrative By-products Business Results Spending the budget, budget meeting deadlines, filling timesheets, compiling RAG reports, ticking boxes, endless meetings Achievement of business goals as a consequence of the investment in the programme/project. Waste Products? Slide 30 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 15 Assurance with Intelligence Results Chain Analysis (RCA) Takes place as part of the overall project planning process Intended to document the relationship between: Activities (initiatives) Deliverables (outcomes) Dependencies (contributions) Risks and assumptions (assumptions) Intermediate and final project goals (outcomes again) Used to identify risks early and tie activities to goals and risks. Slide 31 Assurance with Intelligence RCA Diagram elements Slide 32 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 16 Assurance with Intelligence Simple Results Chain example Slide 33 Assurance with Intelligence Using RCA RCA models the benefits realisation process Stakeholders and management g can see the road mapp leading to the business goal Derived from interviews and workshops Promotes discussion, consensus and commitment A shared understanding of the linkages between business and IT initiatives Makes implicit thinking explicit, explicit brings hidden assumptions to the surface and enables better communication and decision making. Slide 34 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 17 Assurance with Intelligence PI Strategy PI Strategy PI Strategy generates the assessment process Results Chain Analysis PI Strategy Test Strategy Acceptance Goal Assessment Project Intelligence Planning Goal Assessment Project Planning Goal Assessment Project Initiation PI Management Stakeholder Involvement/Project Assurance/Governance Key: Project planning and initiation Project Intelligence Activities Development activities Review and Test activities Slide 36 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 18 Assurance with Intelligence Use goals and risks to define the process for intelligence gathering Test process supports the intelligence gathering p process Cannot rely on fixed, inflexible test processes (but we keep the best of ‘best practice’) Test process design driven by the need to report progress against goals and risk NB: coverage goals are included as usual NB2: we will need new forms of assessment beyond traditional static/dynamic testing. Slide 37 Assurance with Intelligence Goals and risks as PI Drivers Strategy for goal-based progress reporting: Intermediate and final business goals Deliverables: documents, software, infrastructure, training, resources etc. Coverage goals: functional/non-functional requirements, scenarios, regulatory requirements etc. “To demonstrate that these goals have been met or identify the problems that block the goal” Strategy for risk risk-based based progress reporting - Product risks: “to demonstrate that risks have materialised, or have been addressed”. Slide 38 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 19 Assurance with Intelligence Project Intelligence Strategy Project Phase PI Drivers Ass. Obj. Reqs Design Build Integ Systest UAT Trial Prod. Business goals Requirements Objectives for each test phase are easily identified Risks Assurance Slide 39 with Intelligence Slide 39 Test Strategy Completion © 2003-7 Gerrard Consulting Limited Version 1.0 Page 20 Assurance with Intelligence Test Strategy Completion Test Strategy Completion provides the test process Results Chain Analysis PI Strategy Test Strategy Acceptance Goal Assessment Project Intelligence Planning Goal Assessment Project Planning Goal Assessment Project Initiation PI Management Stakeholder Involvement/Project Assurance/Governance Key: Project planning and initiation Project Intelligence Activities Development activities Review and Test activities Slide 41 Assurance with Intelligence Test Strategy Completion E.g. System test… - Objectives (from PI strategy) Environment (technology, mainly) Responsibility (plan, spec, execution, approval) Deliverables (plan, spec, incidents, reports) Techniques (test design, intelligence reporting) Tools (p (plan,, spec, p , execution,, data pprep, p, reporting p g etc.)) Entry/exit criteria (goals met, risks addressed) Etc. etc. Slide 42 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 21 Assurance with Intelligence Test Strategy Completion 2 Coverage policy Timescales/estimates Ti l / i (or ( the h policy li ffor estimation) Regression testing policy Incident management process Reporting R ti approach h Usual stuff…etc. etc. Slide 43 Assurance with Intelligence PI Process – designed to handle change Project Risk Management PI Reporting 3 Project or Process Risks 5 Product Risks Out of Scope Risk will not be tested (no test objective) Risk is not product-related, or is programme-critical 7 Goals Outstanding g Risks Outstanding Coverage Outstanding 12 Goals Achieved Risks Addressed Coverage Achieved All new goals/risks are “outstanding” initially Test Process Management Goals and Risks Test Phase 4 Evaluation & Analysis 2 Classification Test objective defined 6 Allocation to Test Phase 8 Define/ Design Tests 9 Run Test Failed tests must be repeated when fixes received Tests have revealed a new risk or changes to a risk New risk identified Coverage Goals 1 Requirements Analysis/ coverage 10 Identify Regression Test Change affects requirements (and tests) Change 13 Impact Analysis Slide 44 Slide 44 © 2003-7 Gerrard Consulting Limited Version 1.0 Change requires regression tests 11 Run Regression Test Assurance with Intelligence Page 22 Assurance with Intelligence G l Ri Goal, Riskk and dR Requirements i B Based d Testing Slide 45 Assurance with Intelligence Requirements Coverage – Typical System Test Tests Requirements q 1 2 3 Manage Customers X X X Manage Products 4 5 6 X X X Create Requisition Authorise Requisition Raise Sales Order Allocate Stock Fulfil Order and Distribute Raise Invoice Match Payment 7 8 9 10 X X X X Large number of tests, each focusing on a limited scope And so on… Slide 46 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 23 Assurance with Intelligence Requirements Coverage – Typical User/End to End Test Tests 3 4 Manage Customers X X Manage Products X X Requirements q 1 Create Requisition Authorise Requisition 2 X X X X Raise Sales Order X X X X Allocate Stock X X X X Fulfil Order and Distribute X X X X Raise Invoice X X Match Payment X X X 5 6 7 8 9 10 Smaller number of tests, each focusing on a business process path And so on… Slide 47 Assurance with Intelligence Product risks as potential modes of failure Product risks are speculative failures and incidents The risk impact scale is the SAME as incident severity Some modes of failure are unacceptable so we select tests to demonstrate selected failure modes are unlikely NB: these tests COULD be in any project phase Means we focus on the tests that MUST pass We know defects that cause these failures MUST be corrected If we can’t correct failures of risk-based tests we can say these risks have materialised Critical risks block acceptance - Force a change of plan: re-schedule, redesign, re/de-scoping or abandonment! Slide 48 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 24 Assurance with Intelligence Risk-Based Testing complements requirements-based Project Phase Product Risks Test Obj. Reqs Design Build Integ Systest UAT Trial Prod. High Objectives for each test phase are easily identified Medium Low Assurance with Intelligence Slide 49 Risk-based reporting start today Planned end all risks ‘open’ at the start residual risks of releasingg TODAY Progress through the test plan Assurance Slide 50 with Intelligence © 2003-7 Gerrard Consulting Limited Version 1.0 Slide 50 Page 25 Assurance with Intelligence Goal-Based Testing completes the picture – Project Intelligence All projects have underpinning goals/benefits Stakeholders S k h ld want to know: k The status of these goals The status of the risks that block those goals The confidence with which they may make a decision ((test coverage) g ) These are the three areas where test evidence is required: Project Intelligence. Slide 51 Assurance with Intelligence Project Intelligence and Coverage Any test may be designed to address the needs of one, two or all three drivers. The tests in scope cover requirements, risks and goals. Assurance Slide 52 with Intelligence © 2003-7 Gerrard Consulting Limited Version 1.0 Goal Coverage Corollary: a test that doesn’t address any driver is a pointless test. Slide 52 Page 26 Assurance with Intelligence Project Intelligence Strategy Project Phase PI Drivers Ass. Obj. Reqs Design Build Integ Systest UAT Trial Prod. Business goals Requirements Objectives for each test phase are easily identified Risks Assurance Slide 53 with Intelligence Slide 53 PI reporting Project Phase PI Drivers Ass. Obj. Reqs Design Build Integ Systest UAT Trial Prod. Goals Requirements Risks Progress of achievement is clearly seen. Outstanding objectives are highly visible. Assurance Slide 54 with Intelligence © 2003-7 Gerrard Consulting Limited Version 1.0 Slide 54 Page 27 Assurance with Intelligence Objective Objective Objective Benefit Objective Benefit Beenefit Goal G Goal G Goal G Goal G Goal-based test reporting Open Risks R Closed Open Closed Closed Open Open Closed Slide 55 Goal achieved, Benefits available Assurance with Intelligence Close Slide 56 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 28 Assurance with Intelligence PI Summary Historically, coverage (requirements, code etc.) was the only game in town Risk-based introduced a second dimension Goal-based is the third Testing is transformed from just defect-detection to intelligence gathering and decision-support Project Intelligence Extends from business case through to post deployment assessment Involves document, code, system, user, service testing Not necessarily more tests, but more focused tests, much better control and stakeholder reporting. Slide 57 Assurance with Intelligence Finally Best PM practice: manage goals, not activities PI ggives management g visibilityy of risks, goals g and progress p g towards them Project managers and stakeholders are PI customers The test process is the nervous system of your project, designed to provide useful PI in a timely manner What’s new? - Goal and risk-based testing - New styles of test reporting throughout the lifecycle - Appreciation of testing as a key source of Project Intelligence PI is usable in ALL methodologies. Slide 58 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 29 Assurance with Intelligence Managing Software Projects With Intelligence Thank-You! gerrardconsulting.com d lti uktmf.com Slide 59 © 2003-7 Gerrard Consulting Limited Version 1.0 Assurance with Intelligence Page 30