Portfolio approach to IT projects A div. Of chemical co. scrapped a SAP project –They had implemented SAP before –This time used inexperienced project manager •He misjudged amount of change management necessary –Lost millions of dollars •Major credit card co. underestimates processing requirements by >90% when outsourcing –System crashes –Service levels plunge –CIO and a side kick get fired • A mfging co. consolidates activities from 50 plants – Wait time to confirm orders averages 25 secs • More than 2 secs makes the system unusable • Two major insurance co. install same software – One sees 46% increase in sales – Other had to scrap the system at cost of $600 mill • This happened in late 1990s Whats wrong? 3 serious deficiencies in general and IT mgmt 1. Failure to assess implementation risk when the project is funded 2. Failure to consider the total implementation risk of a portfolio of projects 3. Failure to recognize that different IT projects require different managerial approaches Sources of implemenation risk Mismanagement is not included since it is not inherent part of the project. Also assume that proper info tools and methodology is used. • Project with high benefits usually have high risks • Managers produce lots of reports about benefits but very little is done about risk • Risk of time slippage, cost overruns, technical shortfalls, and outright failure seldom assessed 3 main sources of implementation risk 1. Project size • • • • • $ value of the project: bigger $ value carry more risk, relative to other projects undertaken Staffing levels Durations Number of departments affected Project Size relative to the size of the other projects undertaken by the IT department 2. Experience with technology • • • • Familiarity with hardware Familiarity with software Familiarity with operating system Familiarity with dbms etc. Notice its familiarity. A department familiar with these may face lower risk than a dept. new to it Handle risks by using systems integrator consultants with expertise in these technolo. Companies that do these include • CSC, IBM, EDS, Computer associates etc. 3. Project structure • • • Will requirements change before implementation Does project require little or high organizational change Does the project require BPR Low structure projects usually require organizational change The questions in fig 10.3 may not be appropriate for all companies but it’s a good starting point Higher assessment score mean more need for very senior mgmt approval. Show figure 10.1 which shows effect of degree of structure, technology relative to company’s experience, and project size on project implementation risk. Then show figure 10.2 which is an example One companied developed a questionnaire to assess project implementation risk. Show figure 10.3 which has some of the questions. These questions may not be appropriate for all companies but provide a good starting point. Higher score means greater need for very senior level approval. Only executive committee can approve projects with high risks. And managers should ask next slide questions before giving approval. Use the questionnaire several times during the project to show major changes in the risks. Usually risks should reduce with progress as the size of remaining tasks dwindle and people become familiar with the technology. When managers think its low risk and IT folks think its high then usually things go very bad. Using the questions in this case helps. Managers should ask themselves: 1. Are the benefits great enuf to offset risks 2. Can the affected parts of org survive if the project fails 3. Have the planners considered appropriate alternatives Administer the assessment questionnaire several times during the project. Usually risk should decline as make progress on the project. Portfolio risk Company should also assess risk with the project portfolio • A co in strategic quadrant should be concerned if portfolio has no high risk projects • A portfolio full of high risk projects is also not good – Compan operations may be at risk from disruptions • Co in support quadrant shouldn’t have lots of high risk projects in its portfolio Factors that influence implementation risk profile of portfolio Portfolio risk focus factor Low high Stability of IT development group High low Perceived quality of IT dev gr by insiders High low IT critical to delivery of current co services No Yes IT an imp decision support aid no yes Experienced IT development group Yes no Major IT fiascoes in last 2 years No Yes New IT management team no yes IT perceived as critical to delivery of future co services No Yes Co perceived as backward in IT use no yes If the stability is low then the risk level will be high Different projects should be managed differently Principal management tools for project mgmt: 1. External integration tools: used to improve coordination/acceptance/ from/with user & mgmt 2. Internal integration devices: to coordinate with project team members 3. Formal planning tools 4. Formal results control tools Formal planning tools include PM sw, etc. go to next 2 slides External integration tools/tech Internal integration tools/tech User as PM Selection of experience IT as leader User steering committee Team meetings User-managed change control Key design decision info available Distribute proj team info to key users Have users as team members Tech status review/inspections User approval process for sys specs Try to pick people who have worked together before Prototyping with users Get outside technical assistance Progress reports Have team members set goals & deadlines User involvement in other areas Team members personalities should be compatible Techniques for low HR turnover Formal planning tools Formal results control tools Project management s/w Status vs plan reports Milestone selection Change control discipline and systems System specs Milestone review meetings Project approval process Analysis of deviation from plan Postproject audit procedures Result controls are v. good for projects that have: • Clear advanced knowledge of desired results • Whoever is supposed to be effected by these mechanisms can control the desired results • Results can be measured effectively Highly structured projects using low degree of tech are very good for these. Not for low structured projects using high degree of tech. ……here use high degree of internal integration tools Which tools for which project types High structure/low tech: lowest risk, easiest to manage but rare 1. Requirements are defined clearly and do not change so don’t need change management, 2. don’t need heavy representation of users on the team, 3. don’t have to assign IT systems analysts to user departments, 4. don’t require formal user approval of design specs, 5. but still need user training 6. Can be staffed with average skill people 7. Project leader doesn’t require extra ordinary IT skills 8. Can use junior managers to gain experience 9. Pert/cpm techniques work very well 10. Likely to meet milestones and target budget 11. Result control techniques give good data to help team work hard to meet goals High structure high tech projects Example would be converting mainframe system to client-server with same functions. 1. Here interaction with the user is not crucial but is needed to ensure coordination of any changes in the input/output and to deal with system changes that are being brought in cuz the old mainframe had some short comings 2. Often find that the selected technology was inadequate so either select a new tech or some vital features are modified. 3. PM should have strong IT background and should be able to communicate with techies. 4. For large projects its important to maintain teamwork 5. Develop a record of all key decisions 6. Call subject meetings as needed 7. Will find bugs which will increase cost of implementation 8. Formal results mechanisms don’t help much but personnel controls become imp. 9. Internal integration more important than external integration Low structure low technology projects Less risky if managed very well. Fail cuz of not enuf understanding and focus on bus. requirements Get users involved in design, development, and implementation 1. Get user as PM or deputy PM 2. User steering committee to evaluate design periodically 3. Break project into several subprojects 4. Formal user review and approval on all key project specs 5. Distribute minutes of all key design meetings to users 6. Try sticking to key subproject time schedules. Keep turnover among users low, agreement with a user manager who is gone is of very low value….get agreement with the new user manager again 7. Need strict formal change control process 8. Formal planning controls good if users heavily involved 9. External integration tools very important here Low structure high tech projects • • • • • • • • • • • Requirements not clearly and completely defined Technology is complex Managers need technical expertise PM should have ability to communicate with users External integration very important Total user commitment to design specs imp Need highly experienced project leaders Need high degree of support from users If possible divide the project into smaller parts Formal planning and results do not help in the beginning Time, cost, and technical performance very diff to predict at the same time • Show table 10.3 page 592