key: cord-0061615-2qp63x09 authors: Cooper, Robert G. title: Accelerating innovation: some lessons from the pandemic date: 2021-02-10 journal: J Prod Innov Manage DOI: 10.1111/jpim.12565 sha: 43dbeff32fa748b18937cb1047a18705f00e1578 doc_id: 61615 cord_uid: 2qp63x09 The Pandemic taught us that accelerated new‐product development is more important than ever, and provided examples of firms developing breakthrough products in record time. This article outlines five approaches to accelerated development. The first two deal with adequately resourcing new‐product projects, namely the use of focused teams; and effective portfolio management to prioritize projects and reallocate resources. Newer digital tools are outlined that speed new‐products developments. Finally, two development methods are described that move development projects faster: Lean development and Agile development. Accelerated development also has hidden costs: undertaking less innovative projects, and cutting too many corners. Although important, the topic is under‐researched, and the limited research has yielded inconclusive results about acceleration’s expected benefits. The most frequently cited challenge in new-product portfolio management is too many development projects in the pipeline. 5, 6, 7, 8 Portfolio managers are "normally concerned and overwhelmed with issues like the prioritization of projects and the continuous distribution of personnel from the different projects to overcome the urgent crises". 9 The result is that people resources are spread too thinly across too many projects, so that every project, even the important ones, are under-resourced. 10 The lack of needed resources is one of the fundamental reasons why projects take so long to get to market, and is also blamed for shortcuts taken to save time and effort, that later come back and haunt the project team. 11 Benchmarking studies show that top-performing businesses are considerably more focused than others, with dedicated resources for product innovation: Half the top businesses use dedicated This article is protected by copyright. All rights reserved teams for projects, and more than half have fully dedicated product-innovation groups that only work on new products. 12 Consider this example, Project Lightspeed: On a Friday in late January 2020, Dr. Ugur Sahin, received an email with news about a deadly new coronavirus in China. 13 The following Monday, the scientist and CEO of biotech firm BioNTech SE in Mainz, Germany, summoned his board to announce that the company would start work on a Covid-19 vaccine. To accomplish this task in record speed, BioNTech staff was divided into two dedicated teams working seven-days per week (separate teams to avoid transmission). BioNTech had later partnered with Pfizer; and by early December, just 10 .5 months after the project start, the new vaccine, perhaps the most significant product development in recent times, was ready for roll-out: the first Covid vaccine approved in the developed world (on December 9 in Canada and shortly after in the USA). While a 100% dedicated team is ideal for maximum speed, this is not always practical for many manufacturers. For example, there are waiting times in such projects (waiting for equipment arrival or test results), so team members work on other projects; and team members often have other important duties. Thus, to ensure time commitment to the project, an appropriate model for most product-developers is to deploy a "focused team"not quite 100% dedicated as in the BioNTech or Agile developments, but almost; that is, to use a mostly dedicated team with some core team members devoting the majority (60-75%) of their time to the project. Some firms, such as Tetra Pak, a global packaging company, limit the number of "other projects" per team member to one or two, so that for core team members, this new-product project is really their principal job. Thus, the core team is protected from outside disruptions that divert their attention from the project. 14 The development of the Moderna vaccine at warp speed has taught us that given unlimited resources, time-to-market can be cut dramatically. (Moderna received $2.5 billion in funding from the US government). 15 But most firms do not have access to such unlimited resources, but at least they can avoid the opposite situation: under-resourcing projects to the point where they This article is protected by copyright. All rights reserved move at a crawl. Effective portfolio management is one solution to getting focused on the right development projects, and right number of projects. Here are some ways: 16 1. Prioritizing projects at regular portfolio reviews: At these reviews, senior management goes through the list of active development projects, checks that there is the right mix and balance, but most important, does a forced ranking of projects:1 to N. Resources per project are then added down the list. When the resource limit is reached, a line is drawn: Those projects below the line are killed or put on-hold, and their resources are re-allocated to the better projects above the line. The goal is to do fewer projects but better projects, resource them properly, and get them done! Example: MB Company, a manufacturer of baby-care products near Vienna, had too many active projects in their portfolio. At the portfolio review, the CEO complained that newproduct projects were taking too long, including his two "executive sponsored" projects. The portfolio manager produced a list of the 16 active projects - Table 1 which we ranked from best to worst using both the Project Attractiveness Score (based on a scorecard method) and the Productivity Index, both shown in Table 1 . Also shown is the full-time equivalent people (FTEs) needed to get the project done on time, and the current resource allocation (FTEs). Table 1 here Almost all projects were under-resourced (actual versus needed resources), with some projects having less than one-half of an FTE assigned; and the two exec-sponsored projects were consuming 66% of the resources. The total resource requirements for all 16 projects was 2.3 times the 26 FTEs available! After ranking the projects and adding up resources per project (final column), the resource limit was reached after the sixth project. Thus, 10 of the 16 projects were put on hold, and their resources re-allocated to the top six projects. As a result, besides one exec-sponsored project, the other five top priority projects saw their resources increased by a factor of 3.0, and could now be accelerated. A simple procedure, but very effective. This article is protected by copyright. All rights reserved 2. Productivity Index: Ranking projects by their NPVs or some similar profit metric is not the way to prioritize development projects. The result is a sub-optimal list of projects. Instead, the Productivity Index is an effective way to prioritize projects when there are constrained resources, which is usually the case (the method is based on the theory of constraints). 17 The Productivity Index is simply the value of the project divided by the constraining resource, usually person-days or dollars. It gauges what value is added for every additional unit of scarce resourceperson-days or dollarsthat is spent on the project: In practice, one simply takes the NPV from the project's Business Case, and divides by the person-days or dollars that must be spent in order to complete the project; and then one rankorders the projects by this index until out of resources, as in the example above. Note that sunk costs are not relevant to this prioritization model, only the "go forward costs" are counted. Ranking projects by this Productivity Index is one way to cull out the lower "bang for buck" projectsthey fall to the bottom of the ranking list. Done correctly, this method maximizes the value of the development portfolio for a given resource level. 3. Qualitative Scoring Models: Numerous factors have been shown to correlate strongly with new product success and profitability. 18, 19 As a result, research-based scoring models have been developed that predict new-product outcomes reasonably well, as high as 83 percent accurately, pre-Development. 20 Many of the models were initially developed privately within corporations, but in recent years, more are found in the public domain. 21 Seven proven criteria for such a scoring model include:  Strategic fit and importance. This article is protected by copyright. All rights reserved  Product advantageunique superior product with a compelling value proposition.  Market attractivenesslarge, growing, low competitive intensity.  Ability to leverage the firm's core competencies in the new project.  Likelihood of technical successsize of technical gap, technical complexity.  Potential for financial reward.  Risk level (negative)reward versus investment, uncertainty of key assumptions. At the gate meeting, following presentation of the project by the project team, senior management (the gatekeepers), independently of each other, score the project on the criteria, 0-10, using a scorecard. Gatekeepers' scores are displayed on the screen, and major differences are debated. A decision is then reached, and if "Go", the resources are committed. Such models are often better predictors of new-product outcomes than are financial models, which are traditionally unreliable. 22 And the transparent decision process that scoring models enable produces a fruitful discussion, and usually results in more thoughtful Go/Kill decisions. Numerous digital tools are available to accelerate the product-development process. 23 A recent survey of 200 firms reveals that they expect their investments in digital product development to increase their efficiency and performance by 19 percent over the next five years; and also expect to reduce time-to-market by 17 percent. 24 Newer technologies have made prototype development easier, faster and less expensive. Rapid prototyping based on 3D printing was the precursor of these new prototyping tools; a rapid prototype could be used not only to test technical facets of the product's design, but also to seek customer feedback and validation for the proposed new product. Rapid prototyping, now much cheaper and more ubiquitous, partially solves the number one reason for new product failure, namely the lack of understanding of customers' needs. As Steve Jobs, never a proponent of traditional market research, famously said "People do not know what they want until you show it to them." 25 Rapid prototyping makes this possible. And today, 3D printing has gone beyond just This article is protected by copyright. All rights reserved prototypes: Finished products are now being 3D printed. For example, more than a third of the components in GE's new advanced turboprop engine are made by 3D printing. 26 Other digital testing technologies are being used in a similar way. Simulations not only test new products technically, but also allow customer-testing of a product that does not yet exist. For example, Volvo Construction evaluates new truck designs before a working prototype is built by using a real-time simulator, not unlike a flight simulator. 27 And Alphabets' (Google's) selfdriving car, Waymo, logged 10 billion miles, all done by simulation. 28 In a similar vein, developers create a digital twin (or virtual protype) of the new producta virtual or digital replica. Sensors collect real-time data from the physical item, which is used to operate the digital duplicate, allowing it to be understood, analyzed, or optimized. More recently, Virtual and Augmented Realities are being used to test early versions of products, long before the real project is developed. VR and AR can be used to simulate the user's environment so that test-customers can try the product, not just in a lab, but in its intended setting. For example, in the development of FedEx's new drop-box, an early prototype was first developed from cardboard. 29 With potential users fitted with VR goggles, developers were able to allow users to see and use the prototype in a variety of environments, for example at their own homes. AI is also beginning to be used to predict the outcomes of different technical solutionsfor example moleculesthus accelerating the choice of the right technical solution. For example, medical chemists must guess which compounds might make good medicines; thus they synthesize and test countless variants, and most are failures. The same is true for material science coming up with a new molecule for a material. Much of this innovation thus involves making predictions based on data. 30 But AI and machine learning uses real-world data and analysis to predict outcomes, faster and cheaper than traditional lab-testing methods. For example, the design or "invention" of a new product can be accelerated by machine learning to select the right molecule and hence help to create new materials for solar panels or batteries. This article is protected by copyright. All rights reserved The pharmaceutical industry has recognized the power of AI: "The step-wise serial process of R&D will change to be hyper-iterative and integrated, so that real-world information coming back from development will, in real time, change the research being conducted." 31 Many firms' new-product processes have become bureaucratic and slow. Value stream analysis is a well-known Lean-Six-Sigma methodology, designed to remove waste and inefficiencies from business processes; it has seen widespread success in factory-floor settings. But the Lean method can also be applied to new-product development, specifically to remove waste and to make the idea-to-launch system more efficient. 33 Removing unneeded tasks, cutting the preparation of too-many PPT presentations, and deleting work that adds no value all save time. And re-designing the process to overlap tasks, namely parallel processingstarting one task or stage before the preceding one is 100% completecan save time too (as the Covid-19 vaccine developers did, with a "rolling approval process" in the U.S.; Figure 2 shows this accelerated five-stage process) 34 . Other ways to cut cycle time that a value stream analysis may identify are to move some of the key decision points forwardfor example, the decision to purchase production equipment (acquiring equipment is often a long lead-time item). Making this decision before the field trials are completed is risky, but a cost-benefit analysis often reveals that the payoffs of a faster launch through rapid manufacturing are worth the risk. In practice, in value stream analysis, a task force of knowledgeable product-developers maps the current idea-to-launch process for a typical project in the businessevery step, activity, procedure, and decision point. Then they walk through the process and identify all work that adds no value, and also tasks that take too long. 35 Problem-solving methods are employed to This article is protected by copyright. All rights reserved eliminate the sources of delay or to shorten lengthy tasks or steps. For example, Danfoss, the Danish controls company, undertook such a Lean exercise, and over a three-year period, and cut the cycle time for major projects from development-approval through to launch by halfa remarkable acceleration of the process at relatively little cost. 36 Perhaps not coincidentally, this Lean approach is consistent with Agile principles, which value simplicity, defined as "the art of maximizing the amount of work not done". Example: AK produces a key component used in a large piece of production equipment in the paper industry. Although the development process seemed fairly straightforward, projects were taking much longer than they should have. A small task force of project leaders undertook a value stream analysis exercise: They mapped out the entire process, idea to launch, for a typical major project on a long sheet of paper along a wall, about 10 meters in length: every activity, step, and decision point. An abbreviated version of their 10-meter map is in Figure 2 . Then the task force went through the map, step by step, and eliminated tasks that were not needed and worked to reduce the time of lengthy tasks. The longest task in Figure 2 is the field trials, 8-14 months, sometimes with several recycles, thus up to three years for field trials! The task force then used root cause analysis (a fishbone diagram) to identity the main causes of this. Over a dozen causes were identified, including that the customer had to shut down their production in order to do a field trial, at some cost; and often the field trial failed due to lack of prior technical testing, resulting in recycles. For each of the dozen causes, the task force, with other invited experts, brainstormed and arrived at numerous solutions. The result was a significant reduction in the field trial times, and the elimination of recycles. Agile Development methods were developed in the 1990s in the software world to deal with projects facing uncertain and changing information. While a number of different Agile methods This article is protected by copyright. All rights reserved were proposed, the Agile field finally came together in 2001 as a set of principles outlined in the Agile Manifesto, 37 complete with development methodologies with clear rules. 38 Agile is incremental and iterative, a series of build-test-and-revise iterations; it is adaptive and evolutionarythe product definition and project plan change as the project moves forward; it emphasizes frequent and fast delivery of results (for example, product versions), in rhythmic takt time; and it is based on self-managed project teams. Agile (the Scrum method, the most popular) works like this for software development: Agile breaks the development process into a series of short, iterative and incremental time-boxed sprints, each typically about two weeks long. At the beginning of each sprint, the development team holds a sprint planning meeting to agree on sprint goals and create a task plan for the sprint. Once the sprint is underway, the team meets every morningtheir daily stand-ups or scrumsto ensure that work is on course, share information, and solve problems. At the end of each sprint, software increments, potentially releasable, are demonstrated to stakeholders, both management and customers. Finally, the team meets in a retrospective meeting to review how they can improve the way they work and execute the next sprint better. The team then plans the next sprint based on customer and management feedback. In Agile development, the team is dedicated 100 percent to the one project and co-located physically. Agile defines new rolesthe Scrum Master and Product Owner, but no traditional project leader or project manager. The method is also very visual, with tools such as the burndown chart, Kanban chart, and sprint backlog visibly displayed in the team room. Although Agile Development has its roots in the software industry where it has become widely accepted, firms that produce physical products have benefited from Agile Development methods in recent years. 39 What leading manufacturers have done is to build Agile project management methods into the stages of their traditional gating process, replacing the more rigid project management methods such as Gantt charts, timelines, and milestones. 40 A typical gated model with Agile built in is shown in Figure 3 . By integrating Agile methods with traditional gating This article is protected by copyright. All rights reserved models, leading firms such as Honeywell, GE, and LEGO have dramatically reduced time to market as well as getting the product right. 41 percent of their projects, specifically the larger, major revenue generator projects. After seven years using the system, the firm estimates "a 20-30 percent cycle time reduction because there is much less 'redo' in projects now", as well as improvements in productivity. This new Agile-Stage-Gate approach yields three important positive results for manufacturers: 43 1. Development is faster: Sprints are timeboxeda hard stop; the team is focused, and partially or fully dedicated; and frequent scums resolve problems immediately. The result is higher productivity, closer adherence to the time schedule, and shorter development cycles. 2. Gets the product right: Product designs (features, functionality, etc.) are validated by customers (and management) as the project moves alongearly, often and cheaply. The team is self-managed, self-organizing, ideally co-located, and has decision authority. This article is protected by copyright. All rights reserved Specific gains when Agile is applied to manufacturers' projects are in Figure 4 from several large-sample studies. The greatest impacts are increased flexibility to react to changes, improved team morale, and higher customer satisfaction; shorter development times and more adherence to the time schedule are also benefits. 44, 45, 46 Figure 4results from A-SG Not all the promised benefits of accelerated development exist. The theory is clear: Faster to market means higher profits (sales and profits are realized sooner; and money has a time value). Being in the market first results in higher market share, hence more sales and profits -"first mover advantage"; and being late to market means missing the window of opportunity, or incurring other penalties (for example, annoying valued customers dependent on the product). The logic appears sound, but the research evidence is lacking. For example, there has never been a consistent set of research results to support the argument that "first in wins". Often the first to get to market does better, but not always; sometimes the "second in" learns from the mistakes of the pioneer, and does better. Indeed, a review of empirical studies reveals that research has produced inconsistent, even conflicting, results on the relationship between accelerated development and project success. 47, 48 One major problem with trying to prove that accelerated development has economic payoffs is the challenge of putting an economic value on hypothesized benefits. For example, how does one measure the economic gains from "a possible increase in market share because the product was launched sooner"; or the economic losses of "missing the window of opportunity" because the product was slow to market (opportunity costs are always tricky to estimate)? Such economic metrics are difficult to gauge, simply because an assumption must be made about "what would have been" if the product were launched sooner or perhaps later….a hypothetical situation. Some companies have compared the success of projects done before versus those done after implementing accelerated development (profits, sales etc.); 49 but that requires a large sample of projects. This article is protected by copyright. All rights reserved Some research on accelerated development looked at more easily-measured dependent variables, such as the impact of speed on development costs; these studies often yield negative results. While "development cost" is more reliably measured, it may not be a valid metric, however: One usually moves quickly not to save money, but in order to gain competitive advantage and market share. Indeed, some of the actions taken done in order to move quickly, such as those prescribed by Agilededicated teams, daily scrums meetings, multiple iterations, demos to customers and managementare resource intensive and may actually cost more. Other metrics also produce mixed results: 50 For example, the relationship between cycle time and product quality is also unclear: One study found that higher product quality is related to decreases in cycle time, 51 while others suggest the opposite effect. 52 Yet another study found shorter cycle time was correlated with an increase in the new product's sales when it is managed across all of the stages of the process. 53 Conflicting outcomes regarding the importance of accelerated development for newproduct success may also occur because the impact of cycle time reduction may pale when compared to other key success drivers; thus its effect is swamped by the noise. 54 Cycle time reduction may also yield major negatives. 55 In order to cut cycle time, a firm may avoid bolder innovations, which often involve learning and experimentation and thus take longer; the firm may instead focus on smaller, less challenging projects. There is evidence to support this fear: NPD cycle time was found to be cut significantly over a 14-year period, based on data from two PDMA best practices studies; 56 but in that same period, innovative products dropped almost in half as a percent of the total portfolio of projectsfrom 20.4 to 11.5 percent. 57 A second major negative is cutting corners and making mistakescutting short key activities such as VoC or front-end homework. But these same activities have been shown to have a strong impact on new-product success, and hence are vital. One reason for insufficient VoC and frontend homework is limited resources; but the desire to reduce time-to-market impacts negatively on resources available (people x time). 58 Time demands may also truncate the process, where important stages in a well-designed system maybe be skippedfor example, skipping the validation stage in the process (production trials or extended field trials), simply because that stage is deemed unnecessary given the tight deadline. This article is protected by copyright. All rights reserved Project teams, driven by time, may also become too committed to their project and its plan, and fail to pivot when needed. Finally, high priority projects tend to chew up a lot of resourcespulling team members from other important duties to focus on the one project. Accelerated innovation is a worthy goal in product development, and has the potential for delivering many benefits, as witnessed in developments during the pandemic. Five approaches to accelerated innovation have been outlined in the articlefrom better focus and portfolio management through to Agile and Lean development. Many unanswered questions remain however, and thus accelerated innovation is a fertile field for academics to investigate. Three closely related research questions are: 1. Does accelerated innovation really workdoes it yield benefits, specifically which benefits and how much? 2. What are the hidden costs of accelerated innovation? 3. Which proposed methods work the best, why, and under which conditions? One reason for the inconclusive results regarding the benefits of accelerated innovation is the unreliable and largely invalid metrics that try to capture the benefits of speed to market. A fourth research question focuses on the implementation of acceleration methods. Agile methods were first implemented in the software world; but developing a physical product is much different: it's difficult to do demos on-line to customers; and it's expensive to pivot and change a product's design near the end of development. Thus manufacturers have been forced to make significant modifications to the Agile approach. But more research is required here, not just for Agile, but for the other acceleration methods too. Only then will we have the knowledge, confidence and motivation to move forward with accelerated innovation. This article is protected by copyright. All rights reserved Hardly Any Value Very Large Value This article is protected by copyright. All rights reserved A stage-wise approach to exploring performance effects of cycle time reduction The pandemic pivot: the need for product, service and business model innovation Marelli's Covid-killing cabin purifier Innovation due to Covid: yes, but how? New-product portfolio management with Agile: challenges and solutions for manufacturers using Agile Development Manage pipeline bandwidth to avoid derailing new products Six myths of product development Portfolio management for product innovation Project portfolio management phases: a technique for strategy alignment Problems in managing internal development projects in multi-project environments Benchmarking best NPD practices-2: strategy, resources and portfolio management practices How Pfizer partner BioNTech became a leader in Coronavirus vaccine race Agile development for manufacturers: the emergent gating model Moderna vaccine funded by government spending, with notable private donation Making excellent R&D decisions New products -what separates the winners from the losers and what drives success The drivers of success in new-product development A Dutch test with the NewProd model Selecting winning new products: using the NewProd system Based on analyses of actual versus forecasted sales or profits of NPD projects. One study in the public domain is from Procter & Gamble; the "50% success rate" is defined as Digital transformation and its impact on new-product development for manufacturers Digital Product Development Steve Jobs: The Exclusive Biography A treat for the AvGeeks: An inside look at GE's 3D-printed aircraft engine Volvo Construction Equipment streamlines product development with a real-time, human-in-the-loop simulator Waymo has now driven 10 billion autonomous miles in simulation How Handsome used VR to test products for FedEx. Inside Design AI is reinventing the way we invent Artificial intelligence: molecule to market AI takes its best shot: what AI can -and can't -do in the race for a coronavirus vaccine Accelerated Product Development: Combining Lean and Six Sigma for Peak Performance Government Accountability Office Winning at New Products: Creating Value Through Innovation 5 th ed Agile Stage Gate at Danfoss. GEMBA Innovation Conference Manifesto for Agile Software Development Scrum Guides, 2017. The Scrum Guide The Agile-Stage-Gate hybrid model: a promising new approach and a new research opportunity Agile-Stage-Gate for manufacturers -changing the way new products are developed Accepted Article This article is protected by copyright. All rights reserved Agile Development of Physical Products: A Study on the Current State of Industrial Practice Agile Development of Physical Products: An Empirical Study About Motivations, Potentials and Applicability. University of the German Federal Armed Forces Consortium Benchmarking The impacts of speed-to-market on new product success: the moderating effects of uncertainty Accelerated product development. Product Innovation and Management, Part 5 Product development cycle time for business-to-business products The hidden costs of accelerated product development PDMA Foundation New Product Development Report Initial Findings: Summary of Response From Your NPD portfolio may be harmful to your business's health The dark side of time and time metrics in product innovation This article is protected by copyright. All rights reserved This article is protected by copyright. All rights reserved Customer Defined Iterations or sprints (about 3 weeks long) occur within the stages of the Stage-Gate process (many iterations within longer stages such as Development). Each iteration starts with a Planning Meeting, & ends with a Demo to stakeholders, followed by a Retrospect.This article is protected by copyright. All rights reserved