Managing innovation projects in small companies
TIPS for new players
Bob Mills, Department of Materials and Process Engineering, School of Science and Technology,The University of Waikato, Hamilton, New Zealand
Breakthrough innovation creates a legal monopoly (Schum-peter 1934) and is useful to gain competitive advantage. Incremental innovation extends this benefit (Morita 1992). Increased accessibility to alternative suppliers through the Internet is intensifying global competition and is encouraging organizations to optimize innovation processes. Rapid changes in technology and market opportunities mean that innovation processes must incorporate dynamic linkages to operational intelligence for effective control.
Innovation models typically show a high idea mortality rate in early stages (Economist 1999). Resources expended in this risky fuzzy front end are typically financed as a cost of business, (e.g., an overhead for feasibility studies), rather than as a structured investment aligned with organizational strategy. Cooper et al (1993, 2001) have developed an overview stage-gate approach for the new product development process and reports current practice for linking projects with organizational strategy through portfolio management. Management of each stage is not prescribed but can be accomplished by using traditional project management techniques specifying the gates at either end of each stage to start and finish. Disadvantages of this approach are that the process is project-centric and not synchronized to the business calendar for go/no go reviews or funding level assessment, it is not synchronized to other projects for easy comparison and it diverts the responsibility for innovation away from the people who will make it happen towards resource-hungry, self-serving systems. This process is therefore challenging for small firms and small projects.
Software development operates in a rapid change environment extensively and frequently employs approaches that fix delivery dates but truncate functionality (Mimno 1991). This has been successful with captive customers but tends not to “delight” them. Rolling-wave methods also help by allowing details to be incorporated just prior to execution (Githens 2001). However, this still presumes an overall knowledge of future activities during early planning and their sequence (dependencies).
An alternative philosophy called Time-Block Innovation Project System (TIPS) (Mills et al 2002) developed during the past six years embraces ten differences to existing approaches:
1. Projects always start now and always finish when benefit (income) is received.
2. Project achievement goals are synchronized to the business calendar (weeks, months, quarters).
3. TIPS recognizes achievement not activities, which helps create motivation and urgency.
4. Projects are monitored and controlled in relation to maximizing delivery of currently estimated future benefit and are not aligned with past plans and forecasts.
6. All past project cost and effort is considered as sunk, or in current terminology, invested in new knowledge.
7. TIPS projects are replanned from a zero base at the end of each time-block to incorporate new knowledge.
8. Each project is regularly reviewed in the context of project portfolios and assigned a unique priority rank within the organization.
Exhibit 1. Action Research/Learning Cycle
Exhibit 2. Action Research/Learning for Pinnacle Health, Connexionz, and Honey New Zealand
Exhibit 3.Traditional Project Planning Case Guidelines
9. Resource allocation is done on the basis of project priority ranking by empowered staff rather than through the dictates of a central planning function.
10. TIPS acknowledges, quantifies, and standardizes risk estimates.
11. TIPS information is formatted for convenience and practical use.
Trials using TIPS on six pilot projects completed in 2000 indicated potential usefulness but also a need to improve formatting and robustness for the implementation of TIPS in spreadsheet software. Discussion with more than sixty colleagues and peers has also indicated a much broader potential for TIPS than the intended technological innovation process. It is small organizations that are frequently involved in innovation projects and have the most difficulty in diverting resources to manage projects. Therefore this research seeks to discover issues arising from trying to introduce TIPS for innovation in small firms. The results will be used to help determine the functionality required of an innovation project management process to emulate the TIPS philosophy.
New Zealand provides good opportunity for small firm innovation project research because 96 percent of its companies employ less than twenty staff and one recent study (Frederick 2002) suggests it to be one of the world’s most innovative countries (cf. the contrary view of the Economist ). Innovation projects in New Zealand are small and typically of only between 100 and 10,000 person-hours duration (cf. 1 million person-hours for the United States military). A disadvantage of researching small firms in small countries, however, is the difficulty of using traditional Socratic methods where the observer remains independent and the observed remains uninfluenced. Action research/learning has therefore been employed to collect data.
Exhibit 4.TIPS Project Planning Case Guidelines
This paper reports on three action research/learning journeys in the form of case studies. The first case is about fourteen key staff and members from Pinnacle Health, Hamilton, a primary healthcare provider network, who wanted to learn how to manage projects to grow their organization. The second is about the problems of an experienced project manager when moving to Connexionz, Christchurch, a highly innovative firm of fourteen staff making transport management software/hardware products. And finally the case of Honey New Zealand, Te Awamutu, a company of twenty-five staff embarking on a rapid expansion program under a new CEO, employing a university student to develop novel operations management software.
Action research and its corollary, action learning, are used in education research and for introducing new management systems (Cohen and Manion 1994; Koch et al 1995). This methodology was also used to formulate the TIPS philosophy. Interaction between researcher and subjects is used to introduce new approaches to problems and obtain feedback from the subject’s attempts to implement solutions so that improvements can be made and tried. Multiple iterations of the “problem definition, proposed solution, solution testing, and reflection” cycle increase understanding of key issues and engender learning for the researcher and subject. The research/learning cycle is shown in Exhibit 1 and detailed in Exhibit 2.
Data was collected from fourteen Pinnacle Health staff and members during the project management module held for the final six months of a two-year certificate in strategic leadership run by Aiden Holliday of AH Frontiers. Staff had professional backgrounds and included general practitioners, four clinician pharmacists, a medical practice manager, a nurse, accountants, and an information technology developer. Several members of the Board of Governance attended including the Chairman. Only two staff had formal project management training prior to the module. A total of six, five-hour interactions occurred over the period from July to December 2001.
Exhibit 5. Pinnacle Project Planning Survey
The first exercise was to assess project management approach ease of use for newly trained staff. Staff members were each assigned to one of four teams by Aiden Holliday, selected on the basis of members having previously worked together in teams. After two hours of instruction at the first interaction session two teams were assigned the task of preparing a project plan for a powerboat race (McDonough 1989) using the traditional activity-based approach. The other two teams used the TIPS achievement-based approach. Guidelines for project management case study analysis were provided (Exhibits 3 and 4).
After one hour each team presented their plan to their peers. After the presentations each staff member was surveyed and asked to reflect on the ease of the planning process their own team had used and the quality of each of the resultant plans presented, including their own (Exhibit 5).
The survey used a five-point scale rating system and also invited suggested planning process improvements. This procedure was repeated during the second session with another case (Factory Fire) (Stokes and Pinches 1993) each team using the alternative planning approach to the one used in first session. Results were analyzed using t-tests, Confidence Indices, and Kruskal-Wallis tests.
Exhibit 6. First Email from Connexionz (11 September 2001)
Data was also collected at the end of the six-month training period by analyzing reflective essays on the topic: “The problems encountered and the solutions found or proposed during the implementation of project management practices in the workplace during this module with recommendations for areas requiring further development.” The key points of each essay were listed and classified under twenty-three headings. The responses broadly related to: issues for Pinnacle Health, the advantages and disadvantages of traditional approaches and TIPS, and recommendations for future project and portfolio management tools.
This case was initiated from an email enquiry from David Ford,Connexionz on 11 September 2001 (Exhibit 6).
Email exchange over the next ten days culminated in a draft project management development proposal from David Ford to his manager (21 September 2001) and a calendar-based spreadsheet, developed to track and link all the projects in a portfolio (30 October 2001).
Data was also collected from industry seminar presentation notes made by David Ford reflecting on the lessons learned during the prior six months (8 April 2002).
Honey New Zealand
The Honey New Zealand trial was initiated from a Technology New Zealand project-funding proposal written by new CEO Kerry Paul in September 2002. Data was subsequently collected at “weekly” project review meetings with Kerry Paul and a third-year computer science degree student Watkin Foote. Project status information was translated into changes in the expected achievement schedule, weekly job lists, and calculated probability of “achievement on time” using the TIPS spreadsheet (Microsoft® Excel®) developed for earlier trials (Mills 2002). The author, who acted as project overseer (rather than manager) encouraged regular weekly meetings during the period of the trial from 19 November 2001 to 8 March 2002 (sixteen weeks). No meetings were held when there was no progress possible on the project. Weekly time-blocked TIPS sheets were replaced in favor of daily time-blocked TIPS sheets from the eighth week of the trial onward.
Informal reflections from Kerry Paul, Watkin Foote, and Richard Hopkins, user of the final product process, were collected during the month subsequent to project completion.
Findings and Discussion
Results from the ease of use and quality of plans survey for the first trial suggest that the TIPS approach is significantly easier to use for project management novices (3.57 [sd = 0.53] compared to 2.33 [sd = 0.52], where 5 = very easy and 1 = very difficult) and produces significantly higher quality plans (3.77 [sd = 0.6] compared to 2.96 [sd = 0.59], where 5 = very good quality and 1 = very poor quality) (Exhibits 7 and 8). This result is tempered in the second trial a month later where the teams swapped approaches resulting in no significant difference between TIPS and traditional approaches for ease of use with ratings of 3.5 (sd = 0.55) compared to 3.83 (sd = 0.68) and no significant difference in plan quality, 3.62 (sd = 0.38) compared to 3.46 (sd = 0.5) respectively.
Exhibit 7. Pinnacle Health Ease of Use Ratings
Exhibit 8. Pinnacle Health Plan Quality Ratings
Exhibit 9. Pinnacle Health Reflective Essays Topics
In the second trial there was a tendency for those using TIPS to give it lower ratings and teams using traditional approaches provided less consistent plan quality. This may be because the initial familiarity with writing lists gives way to a preference for the visual impact of the completed Gantt chart, even though it requires more training and effort. The result might also indicate that both the teams using TIPS in the first trial were inherently able to synthesise requirements and produce better plans regardless of the approach.
The ease of use rating mean increased from 2.95 for the first trial to 3.67 for the second indicating an overall increase in competence with time and training. Likewise the quality of plans produced a small improvement from a mean of 3.37 to 3.54.
The improvements suggested for the first trial were the need for more time for planning (n = 4), more information (n = 3), and more practice (n = 2). Program Evaluation and Review Technique (PERT) was regarded as entirely unnecessary (n = 1). It was also suggested that achievement and job lists should be worked out within functional groups first before consolidating them for the entire project. The suggestions were reiterated for the second trial but with the observation that although a visual approach is preferred (n = 1), it is difficult to fit all the required information onto one page (n = 2); therefore sub-charts for micro-management may be useful (n = 1). A further observation was that planning should be worked back to the start date from the final scheduled achievement (n = 1).
The reflective essays (n = 13) at the end of six months training identified twenty-three topics for consideration (Exhibit 9).
Most cited was the mature development of Microsoft Project® software for Gantt chart generation and its potential for organization-wide implementation. The graphical nature of the output was much appreciated, most likely as a result of the requirement to produce hand-drawn charts during the initial training. Insightful comments included: “I have previously only seen Gantt charts used by building contractors and to be honest have not grasped much more than the pretty colors on different lines.” and “No one mistakes a pretty chart for careful planning.” Over half the respondents called for project management tools to have high impact visual/graphical information.
Exhibit 10. Five Report Conclusions, David Ford, Connexionz
Nine of the thirteen responses bemoaned the unclear distinction between “projects” and “core activity” at Pinnacle. Once the distinction is clarified and because multitasking is the norm, priority ranking of core activity as well as projects was thought to be useful (n = 1). The associated issue of defining when a project has started had been improved with the establishment of a Project Opportunity Assessment Form earlier in 2001. By the time the course had finished in December 2001 preliminary work had been done to align projects with the company annual plan and rank project priorities.
A significant issue mentioned by six respondents was acknowledging and quantifying project value (or benefit) in dollars to meet a fundamental requirement of TIPS. Examples of the raw benefits of projects were “the prospect of improved health” and “improved relationship with client.” These benefits needed to be translated into financial terms, a process hitherto not considered or attempted.
TIPS produced as many negative responses as positives. Immature spreadsheet software caused some frustration of a mechanistic nature but most frustration was with the constraint of enabling the selection of only one achievement per time-block. Some lesser achievements and the jobs that that enabled it were therefore at risk of being overlooked in the planning process. Furthermore, confusion was caused by listing jobs to be done within the time-block that did not necessarily relate to the desired achievement for that period. Staff still tended to look for jobs to do rather than achievements to celebrate.
One respondent said, “Lists and achievements are part of my everyday life … thus TIPS is a more natural process for me … much more user-friendly for a first-time user.” Another now uses TIPS to list project changes and identify critical assessment points and then uses the information to update his Gantt chart. One respondent concludes, “For truly innovative projects TIPS holds considerable advantage … [however] Microsoft Project seemed to be the most intuitive of the products to me … its visual representation … [and] drill-down capabilities was impressive. I am left wondering what TIPS could become if it was dressed up ….” An information technology (IT) staff member suggested that a hierarchy of information levels would be helpful: workstream (objective), achievement, and task (job).
David Ford advised that his company currently used a four-stage process to take projects from conception to implementation with fixed deliverables required at the end of each stage. He was sent details of recent relevant publications including the TIPS philosophy on 11 September 2001. After three days his thoughts had moved to the possibility of using a two-plan approach: a long-term milestone plan to satisfy business and marketing requirements and a short-term work plan schedule for the next month and updated each month starting from a zero base. He could immediately relate to points 3, 4, 6, and 8 of the TIPS philosophy but had three major problems. Firstly, human resource procurement took three or more months to arrange and his new regime would not adequately inform him of company requirements. Secondly, there was inadequate linkage between his short-term and long-term plans. Thirdly, the Critical Chain concept of measuring project health through the measurement of available buffer size was being used and provided comfort and confidence in meeting firm deadlines (Goldratt 1997).
On 17 September after further consideration David Ford concluded that assessing the probability of achievement completion on time was of more value than trying to assess the completed proportion of scheduled work. However, expecting sign-off for a plan that has only say, an 80 percent chance of being completed on time, was being unduly optimistic. He also confirmed his preference to retain “cosy buffers.”
The following day David Ford reflected on his determination to pursue change: “I am on a mission to find a better way of managing the process. It’s not that I, as a project manager, need a better way (I could just carry on making up the story as I go—I’m quite good at that!) but our company as a whole needs a better way of knowing where it is going and when it is likely to get there. As the ‘project professional,’ I get charged with finding the best way of managing our projects. If a better way means turning the company management upside down, then so be it.” The thought occurred to him too that perhaps Connexionz should have two plans, one with a 90 percent probability of completion for business and marketing and one at 60 percent to motivate the workforce! Practicality and the need for transparency of process quashed this suggestion.
To summarize, this initial journey in the search for a better way David Ford produced a draft ten-page discussion paper for his manager on 19 September 2001. The five report conclusions are presented in Exhibit 10.
By late that month, the events of 11 September 2001 in the United States (US) had rippled through to Christchurch, New Zealand and progress on project management philosophy slowed significantly. However, on 30 October 2001 David Ford reported the establishment of a time-block based portfolio tracking spreadsheet. This is used to list goals for each of Connexionz’ research and development (R&D) projects for 1, 2, 3, and 4 weeks ahead, 2, 3, 6, 9, and 12 months ahead, and 2, 3, and 4 years into the future. The spreadsheet goals are updated each week and linkage between goals across projects is indicated.
David Ford’s project portfolio spreadsheet was developed further and by April 2002 he had some success with the solution he calls Variable Time Block Planning (VTBT). In VTBT the upcoming four weeks are planned in detail; the following two months have enough detail for resource planning; the three quarterly blocks are used for milestone planning, and further annual block for visionary planning.
Honey New Zealand
The original proposal for funding the project was set out (by coincidence) using weekly goals for ten working weeks (12 weeks duration with a two-week Christmas break at the mid-point). A week time-blocked TIPS spreadsheet was drawn up for the project start on 23 November 2001 and project completion on 8 February 2002. In the event, student Watkin Foote was away sick four weeks from week 7 to week 10. The project was not actually completed until 22 March 2002 at the end of week 18.
Weekly meetings held during the first four weeks were used to report project status, discuss and agree on expectations, and make minor adjustments to the proposed plan. The calculated probability of completion on time (TIPSp) was between 0.99 and 1.00. Discussions at the end of week 4 were up-beat and even included the possibility of adding completely new elements to the project if there was enough time. Once all the raw data had been collected from the Te Awamutu site and the shape of the database model agreed the student was based at his home computer in Hamilton (25 km away).
Communication problems occurred between weeks 5 and 7, first with failure of a new email link into the student’s home and then the two-week Christmas holiday period. This was reflected in TIPSp moving to 0.93. Concerns from Honey New Zealand increased with the news of the student’s sickness at the beginning of week 8, TIPSp had now moved down to 0.61. By the following week only four weeks probability data was available on the TIPS weekly spreadsheet, which is insufficient to generate a TIPSp estimate (five values required for statistical comfort). Therefore a day time-blocked sheet was created (looking forward for up to 29 days to the end of the project schedule). The project was rescheduled in that format for the rest of the project. It was also decided at week 8 to extend the completion due date to match the duration of the sickness. TIPSp remained constant at 0.63 during the period of sickness (Exhibit 11).
Upon Watkin Foote’s return to work in week 12 after sickness it was agreed in principle to remove the requirement for some of the desirable software functionality that had been scheduled during the initial period of the project (weeks 1 to 4). However, CEO Kerry Paul was unable to attend the status meeting and the unchallenged optimism of the student rebounded the TIPSp to 0.99. Regular subsequent status reviews resulted in TIPSp values of 0.72 and 0.81 for weeks 13 and 14. During week 15 the process operator Richard Hopkins had become familiar with the capability of the software during pilot testing and requested a completely new requirement to account for inventory damage, reducing TIPSp to 0.68 for the final week scheduled.
The software was finally completed and documented to everyone’s satisfaction after two further weeks. Most of this un-scheduled two weeks was required to fix minor problems and to suit undocumented operator practices and preferences.
There was no change in the expected value or benefit of the project during the project but labour costs increased by 60 percent from $3,920 to $6,272. This reduced the benefit/cost ratio from 6.12 to 3.83 but the actual dollar advantage in completing the project was only reduced from NZ$20 080 to NZ$17 728. Had a decision been made to halt the project once originally estimated costs had been exceeded the company would have invested $3,920 in new knowledge and still have been exposed to $24,000 of risk.
At Week 8 of a scheduled 12-week duration project there was no tangible evidence that the student could complete the task satisfactorily, the student was sick with an unknown “return to work”date and over half the budget had been spent. Why was the project continued? Using project-centric data, continuation may have been considered foolish. Using business-centric data and the trust and confidence between the stakeholders developed during regular weekly meetings to develop a better plan the project continued to a successful conclusion.
Conclusions and Reflections
The TIPS philosophy appears to have some use and appeal for planning innovation projects for novices but the spreadsheet software is presently inadequate for scheduling subachievements and identifying workflows. On larger projects using traditional approaches subprojects would be used for this purpose. This suggests that achievements should be scheduled for each workflow although clearly not all workflows will be present for each time-block and the optimistic and pessimistic achievement completion times for each time-block would need to be consolidated.
Exhibit 11.The Week 9, Day Time-Blocked TIPS Spreadsheet
TIPS requires emulation in reliable software to increase flexibility and visual information presentation. Mass information (e.g., stock market data and financial ratios) appears to be satisfactory in numeric form, presumably because it is digested relatively infrequently with intense interest and has a need for accuracy. Analogue information in contrast is useful for frequent operational monitoring of essential information (cf. speedometers and pressure gauges). This style of information is not currently provided at the project level. Warning light indicators may also be helpful. Candidates for TIPS indicators at project level are:
• Individual and project time-block probabilities (e.g., red < 0.6, amber, green > 0.8)
• Benefit/cost to complete ratios (e.g., red < 2, amber, green > 4)
• Days to completion/days to date required (e.g., red < 1, amber, green > 1.2).
David Ford found it useful to create a rolling calendar-based portfolio of future project achievements to show interlinkage and dependency. Sometimes more than one achievement was important enough to be shown, supporting the findings from the Pinnacle Health trials. A disadvantage of VTBP was desynchronization from the endof-calendar-month and end-of-quarter for most of the time, thereby potentially requiring additional interim data gathering and more time spent on updating goals each week.
The case highlights a dilemma encountered by most innovation project managers. There is a desire to increase completion expectations to solicit and maintain funding and decrease it to create urgency and motivate project staff performance. The Critical Chain buffer device is particularly useful in this respect for traditional activity-based project plans. The adoption of a transparent process requires accommodation by both senior management and staff and moves the project manager into a more favored role of an honest broker.
Honey New Zealand
Using TIPS on a live project was invaluable firsthand experience for the author. Wrestling with immature software that did not provide the hoped for TIPS philosophy emulation was frustrating but rewarding. The value of a day time-block TIPS spreadsheet was readily appreciated.
The Honey New Zealand case caricatured the situation where there is a need to rely on specialists who are difficult to replace. The effective three-week communication breakdown during Christmas caused anxiety and uncertainty and tested the resolve of recently developed trust. The most important lesson learned was the need for good communication and the invaluable benefit of weekly status meetings. These early meetings, which were used to create increasingly realistic plans proved to be vital in team relationships building. On reflection, updating the TIPS spreadsheet live at the meeting would have reduced the need for follow-up communications to resolve issues generated by the TIPS updating process (this adds to the functionality required of TIPS software).
The probability of completion on time was useful as a barometer of project health and particularly helpful in making crucial decisions about scheduling. Project benefit and “cost to complete” were also valuable but were not used as frequently as probability information.
Action research/learning is a controversial methodology. The influence of a researcher in a small company within in a small country, however, is inevitable and helps justify the research approach reported here. A Pinnacle Health board member comments on this topic: “The teacher had an obvious interest in TIPS as a management modality. This is understandable and indeed it is fascinating to see firsthand the evolution of a different way of approaching project management. I had the distinct impression of being engaged as a trial in a process of ‘action research’ where the outcome was not set and the intention was to structure the tool according to the needs of the user. This is to be applauded. There was a sense of genuine research and discovery in the process with risk to both students and teacher. In a standard mode of teaching using Socratic or didactic philosophies this approach would be suspect. In a truly facilitative approach the sense of discovery is invigorating.”
Innovation projects in small companies rely on open communications and trusting relationships to prosper. Information sharing and decision-making must be done on a business-centric rather than a project-centric basis to foster the continuity of strategic intent into strategic integration.
TIPS as a philosophy appears to be credible but is of little practical value without being in a usefully flexible format using reliable media. Interpreted today, TIPS requires a common software platform capable of being shared at a distance through telecommunications networks. This may require TIPS reconfiguration from a spreadsheet to a relational database format.
The author would like to thank Aiden Holliday of AH Frontiers for allowing this research to be incorporated into his strategic leadership program for Pinnacle Health. Thanks also to the fourteen staff and members of Pinnacle Health who were willing and enthusiastic subjects for the trials and provided much appreciated feedback in their essays.
I have yet to meet David Ford from Connexionz face to face but acknowledge a kinship of spirit to search for better ways of managing R&D projects and I thank him for his willingness to share his learning experiences.
I thank Kerry Paul for his trust in using a university student to solve an important bottleneck in the business operations of Honey New Zealand, and Watkin Foote for getting well and doing a fine piece of process software development. I also thank Richard Hopkins for insisting the process software did the required job.
I would also thank Associate Professor Ray Littler who has provided very necessary statistical and biometric advice and guidance.
Finally my thanks and appreciation to the Project Management Institute (PMI®) Research team who inspired and supported me throughout.
Cohen, L., and L. Manion. (1994). Research Methods in Education, 4th ed. London: Routledge.
Cooper, R. G. (1993). Winning at New Products, 2nd ed. Reading, Massachusetts: Addison Wesley.
Cooper, R. G., S. J. Edgett, and E. J. Kleinschmidt. (2001). Portfolio management for new product development: results of an industry practices study. R&D Management 31 (4): 361–380.
Economist, The. (1999). Innovation in industry survey. 20 February: 5–24.
Frederick, H. (2002, March). Why innovation and entrepreneurship are different: implications for New Zealand knowledge economy. Institution of Professional Engineers of New Zealand Convention in Wellington, New Zealand.
Githens, G. D. (2001). Managing innovation program with rolling wave. PM Network 15 (5): 35–39.
Goldratt, E. (1997). The Critical Chain.Barrington, Massachusetts: North River Press.
Koch,N.F., R. J. McQueen, and J. L. Scott. (1995). A methodology to IS study in organizations through multiple action research cycles. Research Report Series.Hamilton: The University of Waikato.
McDonough, M. (1989, August). Staging a powerboat grand prix. The Australian Project Manager: 12–13.
Mills, R. A., A. G. Langdon, C. M. Kirk, and J. E. Swan. (2002). Managing Technological innovation projects: The quest for a universal language. In D. P. Slevin, D. J. Cleland, and J. K. Pinto (Eds.), The Frontiers of Project Management Research. Newtown Square, PA: Project Management Institute.
Mimno, P. R. (1991, January). What is RAD? American Programer: 28–37.
Morita, A. (1992). S does not equal T and T does not equal I. The First Innovation Lecture, The Royal Society, London, 6 February.
Schumpeter, J. A. (1934). The Theory of Economic Development; An Enquiry into Profits, Capital, Credit, Interest, and the Business Cycle. Boston: Harvard University Press.
Stokes, E., and G. Pinches. (1993). Factory Fire. Geelong, Victoria: Deakin University.
Proceedings of PMI Research Conference 2002