Abstract
Keywords
INTRODUCTION
Innovation and new product development (NPD) are key factors driving sustainable corporate growth and competitiveness (Cooper & Edgett, 2003; Cooper et al., 2004; Kavadias & Chao, 2007). As a result, significant amounts of resources are allocated to the R&D of new products on a yearly basis. According to a recent study (Brennan et al., 2020), the global R&D investment is staggering. In 2019 alone, organizations worldwide spent $2.3 trillion on R&D, the equivalent of roughly 2% of the global GDP.
Despite the sizeable investment, sales generated from new products seem far from satisfactory (Cooper & Edgett, 2003). A recent survey of more than 400 industry practitioners found that the three top executive concerns related to innovation were “
A case study of a high‐technology company
Typical reasons for this perception of underperformance are illustrated by the following real‐world case, which demonstrates ongoing innovation portfolio management challenges at a technology company (B Company) in China, which specializes in solar thin‐film technology. This technology allows the production and use of lightweight, flexible, and efficient solar cells and derivative products. In 2018, B Company launched a new product strategy, under the mission of “powering everything,” which reflected an effort to embed lightweight solar thin films into various traditional products, so that these products could self‐generate electricity and offer additional functionality. After 3 years of pursuing projects to execute this innovation strategy, results were unsatisfactory. The senior executive team concluded that its portfolio management process had stumbled over several recurrent issues, which were spread across different functional areas and levels of decision making. The main symptoms of problems were the following: Senior executives An
The example highlights the necessity and value of innovation (or new product) portfolio management with effective resource allocation and innovation strategy support (Kavadias & Chao, 2007). As we will see in the next section, the problems experienced by B Company are not unique, but rather typical for the difficult multidimensional decision challenge of allocating resources across the innovation projects of the company.
The need for a new portfolio management framework
The goal of innovation portfolio management is to support and strengthen the business's competitive position (and often it is also mentioned that it should maximize the financial returns of a company's product innovation investments). Professionals and academics alike have devoted significant attention to this problem since the 1980s (Chagas Brasil & Eggers, 2019; Cooper et al., 1997a; Kavadias & Chao, 2007; Kavadias & Hutchison‐Krupat, 2020; Loch & Kavadias, 2002; Meifort, 2016; Roussel et al., 1991; Wheelwright & Clark, 1992). Optimists maintain that business results are achievable when portfolio review is implemented properly and conducted on a regular basis (Loch & Kavadias, 2002). Moreover, a structured portfolio management process benefits a company by forming a common basis for discussion and strategy implementation (Loch & Kavadias, 2011). By putting discipline into the process and providing a consistent basis of comparison, people can compare projects and assess them on the same basis of information with the same criteria (Hutchison‐Krupat & Kavadias, 2015, 2018). Managers also expect to obtain better focus, and alignment to strategy, as well as balance across the right mix between short‐ and long‐term projects (Cooper et al., 2001).
However, despite the availability of a well‐stocked toolbox, the reality of portfolio decision making suggests that senior managers are still struggling with the multidimensionality and complexity of multiple innovation projects in an innovation portfolio, which are long‐lasting, uncertain, relate to different strategic priorities and possibly interact among one another. Portfolio management continues to be a challenge across small, medium, and large organizations (Eckert & Hüsig, 2021). In this article, we provide an overview of new product portfolio challenges and previously identified solutions with their shortcomings, and we propose a framework to address the portfolio challenges. Section 2 identifies the main challenges of managing portfolios from a literature review; Section 3, then, overviews the solution approaches that professionals and academics have proposed to address these challenges; Section 4, finally, discusses why previous approaches have been insufficient, and introduces a proposed framework to improve innovation portfolio management.
THE CHALLENGES OF MANAGING AN INNOVATION PORTFOLIO
Reflecting the fundamental role of portfolio management for successful innovation and new product development, both practice and academia have identified similar inherent difficulties in portfolio management that challenge existing prioritization and funding methods. Table 1 provides a brief overview of innovation portfolio management challenges that have been identified. Thereafter, we discuss them in more detail.
Overview of innovation portfolio challenges
The diagnoses of the innovation portfolio management challenges are not dissimilar to the problems discussed in our opening case example. The key question is, of course, how these challenges can be overcome—this has been the subject of much professional and academic effort over the last four decades. The next section offers a condensed overview of the most important solution approaches that have been proposed.
INSIGHTS FROM PAST WORK AND EXPERIENCE
Over the past few decades, there has been a wealth of academic research on the innovation portfolio problem, using a variety of perspectives. We attempt to summarize the conceptual approaches as well as the main insights and limitations of the proposed solutions. Six groups are summarized in Table 2 and then further explained. The six groups emerged through an iterative process of reviewing the papers among the three authors, applying a few distinct criteria: the methodology used to analyze the problem, the discipline “of origin” (e.g., Operations vs. Organizational Behavior or Theory), and similar terms used in previous review studies (e.g., Kavadias & Chao, 2007).
Comparison of Six Research Streams on Innovation Portfolio Management
The knapsack problem (KP)
The knapsack problem is the oldest conceptualization of the portfolio problem, as it addressed the challenge of allocating a scarce resource among claimant projects. It has been studied as early as 1896 (Mathews, 1896). It draws its name from the problem of packing the most valuable items into a knapsack with limited volume. If the items could be included fractionally, an optimal procedure would be to rank them by the index “value/volume taken up” and fill the knapsack in this order until it was full. But if items are indivisible (must be included completely or not at all), this procedure is not optimal because it may end up with left over unused capacity because the last item no longer fits—one may need to include “less valuable (per volume unit)” items to fill the capacity to the highest total value.
This ad hoc filling of capacity requires combinatorial optimization methods, and it can be shown that the problem complexity grows exponentially with the number of the items (in technical terms, it is “NP complete”) (Balas & Zemel, 1980; Dantzig, 1957; Karp, 1972; Martello et al., 1999; Martello & Toth, 1977).
The Knapsack problem has had limited impact on practice (Cooper et al., 1998; Loch & Kavadias, 2002; Martinsuo, 2013) for two reasons. First, the solution algorithms are complicated, representing a black box to users, and prone to degeneracies without careful calibration (Kavadias & Chao, 2007; Loch et al., 2001). Second, the Knapsack problem rests on assumptions that are too rigid to keep recommendations relevant—project values and resource needs are not deterministic but stochastic, project value has multiple dimensions beyond only financials, while project scopes (and therefore budgets) are not fixed but can be adjusted up and down; so projects can indeed be included “fractionally,” at least in one period.
Dynamic programming and real option valuation (DP/ROV)
Dynamic programming is a method of representing, at first, individual projects, which go through “stages” (or phases) of development in a stochastic manner (for example, projects may progress well with “high quality” or less well with “lower quality”), and the status in one stage influences a project's evolution in later stages. A key assumption is that a
Special cases of DP have become well known. First, the so‐called multiarmed bandit (MAB) problems, corresponding to the “stop or continue” arms of a slot machine can represent stop‐or‐go decisions across multiple projects that share a critical and indivisible resource (e.g., expensive lab equipment). A critical number called the
DP methods have also been able to roughly characterize what one might do when multiple projects interact. One important case consists of projects in a portfolio competing for access to scarce resources and causing waiting (queueing) delays (Kavadias & Loch, 2003). A robust simple policy prescribes giving priority to the project with the highest delay cost divided by the expected processing time (Harrison, 1975; Smith, 1956); this requires linear waiting costs to be optimal but represents a reasonable approximation even for convex waiting costs (Ha, 1997; Van Mieghem, 2000; Wein, 1992). Instead of prioritizing on delay costs, one may also prioritize on “value at stake” (Loch & Kavadias, 2002; Zschocke et al., 2014). In this case, more available capacity lowers the threshold for project continuation (Kettunen & Salo, 2017; Lewis et al., 1999; Stidham, 1985), while the continuation threshold increases when capacity is scarce immediately (Kleywegt & Papastavrou, 1998) or in subsequent stages. Similar priority schemes may be useable when considering the capital budgeting of competing projects and accounting for their real option values (Childs et al., 1998). 2
DP applications also (like Knapsack applications) become complex quickly when multiple projects are in a portfolio, and thus they become “black boxes” that managers can no longer connect to the accuracy or distortions of the underlying data. Also, a DP/ROV approach faces limitations in the presence of unknown unknowns (ambiguity) as discussed in Pich et al. (2002), Adner and Levinthal (2004), and Klingebiel and Adner (2015). Therefore, “optimization” applications of DP have not been widely used (with some exceptions where the evolution of projects is very simple, and data are easily available). However, DP has been very powerful in building a conceptualization and an intuition of how to think about dynamic reviews over the course of project evolution (from one stage to the next), what to look for (e.g., updated value, delays, quality), and based on what criteria to change actions (e.g., kill, continue, run a test to reduce uncertainty, adjust resources, modify). Projects are not “fixed items” as in a portfolio knapsack that, once included, are committed until failure, but they evolve and their treatment is updated. Thus, DP has been influential by creating a conceptual language of how to think about project changes and respond to them.
Decision analysis (DA)
The first two streams discussed in Sections 3.1 and 3.2 aim (at least in principle) at optimization, that is, they seek to determine “what” the portfolio composition should be. In contrast, the third stream, Decision Analysis (DA), provides tools and frameworks from a decision theory perspective to assist decision makers: it tries to address the “how to” of project portfolio selection processes. The key contribution of DA to new product portfolio management is the introduction of Multiple‐Criteria Decision Making (Henriksen & Traynor, 1999) 3 : KP and DP start with the assumption that there is one agreed‐upon value maximization criterion (e.g., “value” or “total return”), where in reality, a portfolio has multiple purposes (financial return, as well as segment coverage, as well as robustness, etc.). Moreover, multiple stakeholders in the organization maybe looking at different relevant criteria. Multi‐criteria valuation builds upon the classical, and simple, “pros and cons” table, listing the strengths and weaknesses of each project on each relevant criterion. It may also assign importance weights to the criteria. Through these (importance) weights, the multiple criteria values can be aggregated into a composite “overarching” project value. More sophisticated aggregation methods than linear combination with importance weights also exist, such as the analytic hierarchy process (Hammond et al., 1998; Saaty, 1994). DA may be seen as the “prerequisite” to applying DP (or other methods of making portfolio inclusion and prioritization decisions), as DA provides the agreed upon value dimension to base its project decisions on.
Multicriteria decision tables are not sufficient to make portfolio decisions because they fundamentally consider project values in isolation: the aggregated values (obtained from weighted averages of all relevant criteria) are obtained from individual project data alone, as if each project existed in isolation. Therefore, they can at most achieve combinatorial comparisons across individual project values. However, by their very nature, portfolios are collections of projects that must “act” together to support and achieve the strategic priorities of the organization. Metaphorically, even if one likes oak trees the best, a forest that consists only of oak trees will likely not fulfill the objectives of the forest master who tries to achieve various goals (for animal populations, soil support etc.).
Other proponents of DA emphasize that it must support information exchange and consensus regarding selection criteria (Brenner, 1994); DA tools should facilitate communication, the interpretation of different individual visions, and collective problem structuring (Lawson et al., 2006; Montagna, 2011). This element of DA already points to the next stream of methods, organizational, and stakeholder interactions.
Organization and team (OT)
The fourth stream (OT) highlights the insight that innovation portfolio management is not only an optimization and decision‐making challenge, but in its core a multilevel organizational problem. Therefore, effective portfolio management needs to achieve consensus toward a common goal, which requires organizational processes (rather than ad hoc decisions): (1) strategy (firm's strategic orientation and aspirations); (2) organization (organizational characteristics, cultural aspects, organizational learning); (3) leadership (leadership style, leadership involvement, competence of different management levels); (4) process (formalized project management and project reviews at specified intervals); (5) risk management; (6) incentives and coordination (incentive structure, coordination, and control roles). Successful portfolio management requires frequent project reviews, assigned responsibility, sufficient slack capacity, and stable core teams on projects (McDonough III & Spital, 2003). Innovation portfolio decision‐making effectiveness (portfolio mindset, focus, and agility) is positively related to innovation portfolio success (Kester et al., 2014).
OT elements allow constructively influencing the relationships between the actors in the organization and portfolio outcomes. Since the early work of (Bower, 1972), researchers have considered different relationship dimensions to understand portfolio outcomes. The diversity of perspectives among the stakeholders that may approve portfolio projects (Criscuolo et al., 2017; Oraiopoulos & Kavadias, 2020), the delegation of decision rights (Hutchison‐Krupat & Kavadias, 2015, 2018), the establishment of shared incentive schemes beyond individual project bonuses (Crama et al., 2019; Schlapp et al., 2015), as well as the specificity and breadth of the communicated strategic objectives (Chao & Kavadias, 2008; Hutchison‐Krupat, 2018; Klingebiel & Rammer, 2014) shape which projects are included in the portfolio. Different configurations of OT can support innovation portfolio management, and which “rules” the organization applies should be contingent on high level portfolio strategic objectives (e.g., how many radical innovations are desired, or how many per market segment). Much more research is necessary to be able to characterize the “normative” OT support of innovation portfolio management.
Psychology and behavior (PB)
Most formal models of operational and decision analysis assume that the people involved in operating the systems are perfectly rational, or at least can be induced to behave rationally. Indeed, the very purpose of decision systems is to support the rationality of decisions, meaning the inclusion of considerations that the organization recognizes, values, and feels it can defend against internal or external questioning. However, uncertain and complex decisions such as those shaping innovation portfolios are
Intuition is powerful and important because it allows people to come to conclusions quickly when required, and to implicitly aggregate information that they have access to, formally or privately, in their own heads. However, intuition is also well known to be prone to biases and distortions. Some biases result from the heuristic shortcuts people use when faced with complex decisions (Benartzi & Thaler, 2007; Chao & Kavadias, 2008; Gigerenzer & Gaissmaier, 2011; Gino & Pisano, 2008; Kester et al., 2011; Keyes, 1972; Loch, 2017). Examples are salience, such as the position of a project in a list, where early placement increases the selection probability (Muthulingam et al., 2013; Schiffels et al., 2018) or the way individuals fill the available resource budget (Pape et al., 2020). Other relevant biases are short termism (paying less attention to consequences far in the future), overconfidence (the underestimation of ignorance about outcomes), or loss aversion (e.g., spending more effort to avoid a small failure than to ensure a large upside). These are individual decision biases. They can be sufficiently important to warrant changes in discussion protocols, for example, presenting projects in varying order over the day, or explicitly discussing long‐term effects to make sure they are not implicitly (and unspokenly) discounted (Criscuolo et al., 2021).
Beyond individual biases, there are also social preference biases, or unconscious social needs in groups. In particular,
Qualitative portfolio analysis (QPA)
The last stream (QPA) is motivated by the observation that many formal methods have not had a large impact on practice because they are too complicated and require data input that is either hard to obtain or manipulable (such as underlying financial analyses, scoring models or DPs). This stream proposes the use of qualitative portfolio diagrams (“bubble charts” or other visual representations), instead of optimization models, to represent the key strategic trade‐offs that the organization needs to resolve (Cooper et al., 1997b; Roussel et al., 1991; Wheelwright & Clark, 1992). Rather than the optimization of some type of composite value index, the explicit purpose of QPA is to achieve a
In addition to these tools, Cooper et al. (1997a) propose a “strategic bucket” model, which in its simplest form states that “we should have 90% incremental projects that support the current business, and 10% of radical projects that create new business.” The 10% here would reflect, of course, the high risk of radical projects, but what the right bucket categories are, and what the right target percentages, is hard to “prove” (and therefore prone to manipulation). Chao and Kavadias (2008) examine the business environment conditions under which the strategic bucket approach can be productively applied (environmental complexity and instability affect the applicability of strategic buckets), whereas Hutchison‐Krupat and Kavadias (2015) show that strategic buckets can mitigate information asymmetry challenges in hierarchical organizations and ensure that project leaders align team efforts with the overall strategic objectives.
WHAT SHOULD INNOVATION PORTFOLIO MANAGEMENT LOOK LIKE?
Our discussion in Section 3 suggests a peculiar situation. Amidst a general awareness of the importance of the innovation portfolio, a significant number of conceptual frameworks and organizational procedures have been put forth over the last few decades, each capturing pieces of portfolio management and addressing parts of the problem (e.g., how to address biases with decision procedures). Yet, management methods have not managed to arrive at an overarching “professional” procedure that would give us hope to overcome the challenges identified in Sections 1 and 2. In other words, well‐justified methods are available, but they appear as fragmented answers to the overall challenge, and this way are insufficient and have failed to win wide acceptance by practicing managers.
Recent discussions of “Integrated” portfolio management
There is some evidence that we need to, in some way, combine different approaches to get good results: Cooper et al. (1998) found that leading companies used a
Another recent proposal builds on “agile development” (Thomke & Reinertsen, 1998) concepts. Cooper and Sommer (2016, 2018, 2020) propose an agile‐stage‐gate hybrid approach that enables decision makers to get better information earlier through repeated iterations and validation demos. These proposals come at a highly aggregate level, and they do not specify whether they suggest more early testing, so more solid information is available for planning the stage gate process (which would be merely an extension of the stage gate philosophy similar to Cooper, 1994), or whether they actually suggest that more testing and information updating should be incorporated in projects (which would be consistent not only with DP thinking but also with recent approaches to the management of innovative projects including pivoting and parallel trials, see Lenfle, 2022). Certainly, intensified reviews of individual projects would enable projects to “pivot” and allow the portfolio itself to possibly change its composition more frequently rather than sticking to large projects over many years. This would be consistent with a general trend in project management over the last decade toward more flexibility and adjustment (Davies et al., 2022; Loch et al., 2006). However, the call for frequent reviews and evolving the innovation portfolio (even if this is not “programmed” by a DP planning) is not by itself a fundamental deviation from previously discussed portfolio management principles.
A different interpretation of the observed innovation portfolio challenges
The discussion so far suggests the possibility that the professional and scientific innovation management community perhaps have actually produced sufficient methods—solving the portfolio management challenges may not require new additional methods, but it may be sufficient to apply the existing methods in a different order and spirit. Figure 1 summarizes the known innovation portfolio methodologies with a characterization of their emphasis.
A “process” of innovation portfolio management
The innovation portfolio needs to first and foremost (1) support the organization's business strategy (embodying its market‐offering‐related change over time). In order to accomplish this, (2) a collection of
Within our discussed collection of methods, knapsack, decision analysis, and qualitative portfolio methods can be connected to Step 3i, DP and agile methods can be connected to Steps 3ii and 4, and OT and PB methods can to Steps 5 and 6. While we may see Step 1 as outside the scope of portfolio management (one cannot at every turn question the strategy), we cannot but notice that Step 2 has simply not been addressed. Implicitly, it has been taken as given that a reasonable candidate list exists from which one can select—the projects have been seen as “falling from the sky in sufficient number and fully formed” (perhaps being produced by some other organizational process that we do not need to concern ourselves with). Therefore financial or multicriteria decision making have been taken for granted as the right methods to select the “best ones.” Moreover, we seem to be assuming that the other mysterious process has produced an already “good” list, which can be relied upon to already address our strategy, which allows us the luxury to use
As a result, the usual method discussions miss an important qualitative difference in the phases of the portfolio management “process”: the creation of the portfolio (based on strategic priorities) remains at its core a
In contrast, the methods that professional and academic experts have proposed are
A proposal to create rather than merely select portfolios
Our discussion above suggests the conclusion that the core portfolio challenges diagnosed in Sections 1 and 2 of this paper have been treated with the wrong “recipe,” with practicing encouraged by the advice of academics and professional experts. This recipe has
Thus, we propose that instead of trying to invent more (and newer) methods to address the widely lamented innovation portfolio problem, we may benefit from correcting the logical order of steps (see the extension of Figure 1 in Figure 2): in a strategic discussion, the deciding body agrees on the strategic innovation goals. Then, the decision makers need to
Then analytic methods come in: do the projects fulfil minimum financial requirements? Are there excessive risks involved? Are there important negative (or positive!) interactions among some projects? Are the resource requirements too far outside of the acceptable capacity (in budgets or person‐years)? The formal methods support the design process by ensuring its soundness, by pointing out holes and weaknesses. Typically, the first draft will not work, and one or two more rounds are required, where profitability is improved or budgets cut, or a project added. Sometimes, feasibility (e.g., resource constraints) can mandate that some target fulfilment must be renegotiated. Thus, Figure 2 has an arrow going back from the portfolio analysis to the strategic goals: although we stated earlier that strategy should not be questioned every time around, it may happen that the portfolio design uncovers a constraint that must change the strategic goals (and sometimes it does indeed happen that the portfolio uncovers an opportunity that may even improve the strategic goals!).
After a portfolio has been created as well as evaluated (steps 2–3), discipline can then be built into the further
Proponents of the qualitative portfolio approaches (discussed in Section 3) have long emphasized that portfolio diagrams need to stimulate an active strategic discussion by the decision makers. However, we have pointed out earlier that the way this has tended to be applied has been via standardized diagrams (e.g., “return over risk,” or “process change over product change,” or “technical novelty over market novelty”). The use of standardized portfolio diagrams has prevented a true discussion of the strategic goals of
Indeed, review articles have repeatedly called for extending innovation portfolio management from isolated decisions of the R&D funnel to a holistic process (Kavadias, 2014; Kavadias & Chao, 2007), for viewing innovation portfolios as a multilevel organizational problem (Meifort, 2016), and for integrating across levels of analysis with a macro lens and a micro lens (Chagas Brasil & Eggers, 2019).
Honest strategic debates about the innovation portfolio may represent a challenge for some management teams, whose very senior members have too much status and political capital at stake to risk ever being contradicted or even outvoted. On the other hand, a shared view of the innovation initiatives and their priorities, which stems from an honest discussion in the interest of the organization, makes the portfolio more robust as well as makes it easier for the management team to explain it consistently to the organization and stakeholders. Table 3 lists the arguments why our proposal addresses the challenges from Section 2.
Why our proposal may improve the shortcomings of innovation portfolio management
The arguments so far are conceptual, based on an assessment of results from previous practices and studies. The next question is then what the application of our proposal, distinguishing innovation portfolio design versus evaluation and management, may look like in practice. We now present an illustrative case study, and then discuss that our proposal needs to be tested with additional research.
AN ILLUSTRATIVE CASE EXAMPLE
Table 4 compares three companies in the solar energy market (the names and numbers are disguised for confidentiality reasons). Company A is the market leader with a revenue of $8.1B and the highest profit. Company B is the oldest and smallest of the three, but with the highest net profit margin. Company C is the youngest and least profitable (but profits are growing fast).
Three solar energy companies
The companies have distinct strategies: A grew out of physics and has strategy of being the technology leader with the highest quality and solar energy conversion efficiency. B comes out of mechanical engineering and wants to lead with thin‐film applications across markets (three of which have already been developed, construction, transport, and solar paper backpacks). C's founder came from financial services, and the company is the cost leader.
Not surprisingly, the three strategies lead to different strategic new product and R&D goals (Figure 3). Company A emphasizes growth and technology leadership; B emphasizes growth and presence in target markets, and C emphasizes growth, lean management, manufacturing excellence and cost reduction. These different goals require different kinds of projects in their R&D (product innovation) portfolios (Figure 3 lists not individual projects, but aggregated programs of multiple projects with common themes).
Strategic goals and program portfolios of the three companies
Company A has programs on new technologies, and programs for enhancements of functionality, reliability, and energy conversion efficiency. Company B has two types of programs: solar cell and panel technologies that support their base products (which still account for slightly over half of sales), plus exploration of new markets and product adaptations for the various verticals. Company C focuses on programs that are entirely focused on cost efficiency: production equipment, factory layout and systems, supply chain efficiency, and manufacturing process development and improvements. Although these three companies compete with one another, their R&D and product portfolios are entirely different—as they should be, as the companies have very different logics of value proposition and market position.
The three very different innovation portfolios cannot be explained with financial evaluation methods or standard bubble diagrams. Each portfolio must have been produced (even if implicitly) by a translation of the company's specific innovation goals into a project collection: A's portfolio tries to achieve functional superiority, B's portfolio a balance across (core and new) markets (so, here the standard bubble diagram of market versus technology newness would actually apply), and C pursues supply chain and process excellence. This is most simply shown with the arrows in Figure 3; a more sophisticated representation could put qualifiers on the arrows of strength, or value, or (ir)replaceability. The “balance” of the projects’ strategic contributions is entirely specific to the strategy of
This is consistent with our proposal: only a design step can capture these portfolio priorities, no standard set of criteria, no matter how elaborate, can achieve this. Analytical methods are important in helping to make sure that the new technology projects included in the portfolio of company are sound, but analytical methods fundamentally cannot identify that new technology projects are needed.
This interpretation of the portfolios at the three companies is supported by a failure that company B diagnosed within itself: as selection criteria to the projects, even the new‐market projects, it used to a large degree technical quality and conversion efficiency. This was adequate for the cell and panel programs, but it violated its own strategic logic in its new markets and market introduction programs: customers (both consumers and industry clients) did not value the subtleties of technical performance (and were indeed not bothered by slight
Thus, the project inclusion criteria did not match the stated objectives in the design process (winning customers in new markets). This mismatch of selection criteria (at the detailed project level) with the strategy caused the products to be less attractive, as well as too expensive, which compromised company B's market share, growth, and profitability. This kind of failure cannot be avoided by increasing the sophistication of analytical methods; it requires a clear consciousness of the strategic priorities, shared down to the project level, so people at the operating level can propose and select consistently and make consistent design decisions.
CONCLUSION
This article has diagnosed a widespread impression that innovation (or new product, or R&D) project portfolio management in companies does not seem to be working to general satisfaction. Significant dysfunctionalities seem to persist, from politics and data manipulation over just wrong and short‐sighted decisions to a failure to follow‐up and evolve portfolios over time, which we illustrated in an example and from surveys. We have then overviewed available methods that have been proposed in academia and professional circles over the last 60 years. These methods look (collectively) competent and relevant, but have apparently not gained sufficient application across industries, or they do not address the problem correctly.
We have then argued that the failure perhaps stems not from a nonapplication of the methods, but from an application in the wrong order and with the wrong logic: the creation of a collection of initiatives that actively support and further the business strategy goals is just a
As a second step,
Analytical methods can support the creative activity of portfolio construction (by insisting on the appropriateness of the projects included) but cannot drive portfolio construction. But portfolio management in practice has often tried to construct portfolios with formal criteria (rather than as a collective design act, which then becomes a “how to explain this” afterthought).
Although already Roussel et al. (1991) proposed that senior management teams must agree on the portfolios as strategic instruments, our proposal turns widely used practice on its head, placing the demand on senior manages to regularly engage in a design (or re‐design) process that requires subtle and difficult strategic negotiations. But who else than the senior team that owns the innovation portfolio can undertake these negotiations, which must happen somewhere in the organization?
We demonstrate with a case example, comparing three solar energy companies, that the “innovation portfolio design” as separate from portfolio evaluation and management approach may offer better outcomes. We are the first to concede that our proposal is, so far, based on a new view of previous work rather than on empirical evidence. Empirical research is needed in order to (a) show what the portfolio design process might look like (although design processes for managerial actions are well known in related areas, a transfer will have to be accomplished), and (b) to build evidence that this approach indeed benefits the organizations who apply it. We see this as important work for the near future.
Footnotes
1
).2
) is not discussed here because it has focused on the capacity need of a portfolios (and the necessity to have capacity buffers),rather than on the choice of which projects should be prioritized to join the portfolio.3
4
) say that the company needs to decide what it wants,but a key part of what they propose is their “innovation ambition matrix” that has as its two dimensions (axes) market novelty and product novelty,a well‐known standard categorization of innovations. This may be the right way to distinguish the innovation emphasis for some organizations but not for others. It does not address the need of companies to clearly articulate what it is they want to achieve.