PDF Summary:Loonshots, by Safi Bahcall

The ideas that change the world—ideas like radar, insulin, and computer animation—all have something in common: They begin as flawed, incredibly fragile ideas that almost don’t make it. That’s because they’re what Safi Bahcall calls “loonshots,” ideas so original and ambitious that they and their creators are initially dismissed as crazy (loonshot = loony + moonshot).

In Loonshots, Bahcall argues that organizations can deliberately cultivate these kinds of ideas. He presents four rules for nurturing loonshots and explains how organizations can keep innovating no matter how big they grow. Drawing on his background as a physics PhD, biotech CEO, and business consultant, Bahcall explores topics such as why loonshots die three deaths, why organizations are like traffic jams, and why you don’t want to be a leader like Moses.

In this guide, we’ll build on and challenge Bahcall’s ideas by connecting them to similar books on business and creativity. We’ll also suggest a bonus Shortform fifth rule that complements Bahcall’s original four.

(continued)...

On the other hand, Bush also had to bring feedback from the military to his innovation team. For example, when pilots tried the first versions of in-aircraft radar, they found them too complicated to use during combat—ease of use during a dogfight hadn’t occurred to the engineers in the labs. With the pilots’ feedback, OSRD’s innovators developed more user-friendly displays and controls which the pilots then adopted. (Shortform note: In Cagan’s terms, even after OSRD’s engineers found a feasible way to build radar small enough to fit in an airplane, in-aircraft radar was not a viable product until they made it more understandable to its intended users.)

Rule 2 for Small Teams

Rule 2, as Bahcall explains it, assumes an organization that is big enough and established enough to have completely separate innovators and implementers. But what about startups, which might be small enough that everyone is involved in all aspects of the business at first? In these cases, it’s still true that innovative ideas need to have some kind of practical application, but the idea of mutual exchange between two specialized groups doesn’t make sense.

In The Lean Startup, Eric Ries tackles this and similar problems. Ries suggests starting out with a build-measure-learn approach. To use this approach, start with a hypothesis about what your customers want or need from your business. Then, build the minimal product or service needed to test your hypothesis. The point is to expose your intended customer to your intended product or service and collect data. Finally, you analyze that data to refine your product idea and your strategy.

In The Innovator’s Dilemma, Christensen suggests that a similar process can help businesses figure out how to respond to innovations that might interrupt their market. In the absence of market research, he advises that businesses focus on learning rather than execution.

Project Champions

One of the takeaways from Bahcall’s first two rules is that loonshots need champions. A loonshot champion is someone who passionately believes in the idea and promotes it to outsiders; this person needs to be able to liaise between the idea’s creator(s) and the people who have a practical stake in implementing the idea. In the case of the OSRD, Bush served as a project champion.

Bahcall notes that it’s rare that the person who originates an idea is also the right person to champion it. He suggests that organizations should specifically appoint and train project champions.

How to Be a Loonshot Champion

Bahcall doesn’t go into a lot of detail about who loonshot champions should be or how they should approach their jobs. To understand what such a champion might look like in a business setting, we can look to Marty Cagan's description of a product manager.

Cagan explains that a product manager needs to understand the intersection between technological innovation, the market for that innovation, and how the innovation and its market fit into the company’s business plan. He says that a product manager should be smart, curious, creative, persistent, and collaborative. The product manager needs to be willing to advocate for the creative team’s ideas, and in order to do so, he or she needs a basic understanding both of the technology in question and of how the business works.

In the OSRD, Bush fit this definition of a product manager well. He had a solid grasp of the technologies his team was developing as well as a clear idea about the military’s practical needs. Moreover, he had a good rapport with scientists and generals alike, which meant he was in a position both to advocate for his team’s ideas and to provide practical feedback about what would and wouldn’t be useful on the ground.

Rule 3: Focus on Process, Not Outcome

As we’ve seen, loonshots are fragile, and you should expect them to fail often. Because early failures are inevitable, Bahcall says it’s crucial to focus on your processes rather than your outcomes.

In particular, Bahcall stresses the importance of analyzing your decision-making. If you try something and it fails, Bahcall says that you should ask yourself how you decided on that action in the first place. He suggests doing the same when you succeed. This practice helps you determine whether you’re making good decisions, and it also helps you sort out whether your results (good or bad) are due to the quality of your ideas or due to luck. (Shortform note: Similarly, Pixar’s Catmull suggests that you follow up every project with a postmortem. Doing so lets you reflect on what you’ve learned and plan your next steps, and it also helps clear the lines of communication between teams, which will help with future implementation of Bahcall’s Rule 2.)

How to Learn From Failures

Bahcall says that when a loonshot fails, you should probe into why it failed and determine whether you can learn anything to improve the idea and try again. He also says that it’s important not to be defensive or dejected, but instead to ask as many questions as you can about your rejections and failures. For example, Endo persisted by looking into why his rat trial failed instead of assuming his drug didn’t work. Likewise, when pilots didn’t use the first in-aircraft radar systems, Bush asked why instead of giving up on the idea (or insisting that pilots adopt a technology that didn’t work for them).

More Ideas for Learning From Failure

In The Magic of Thinking Big, David J. Schwartz offers some further advice for learning from every setback. Schwartz advises that you don’t assign blame—to others, to yourself, or to luck—and that you be open to experimenting with new approaches to a problem.

In an organizational setting, Catmull suggests openly embracing failure as a key part of the creative process. In Creativity, Inc., he writes that leaders can destigmatize failure by discussing their own failures and explaining how they’ve learned from them.

For a specific method for analyzing failure, Ries suggests the Five Whys in The Lean Startup. He explains that when something goes wrong, you should ask yourself why five times in a row. He says that doing so will lead you to the root causes of a problem, which will let you make adjustments that can prevent a wide range of future problems.

Rule 4: Balance Stake and Rank

Whereas the first three rules all have to do with nurturing innovation, Rule 4 is about how to maintain innovation as a company grows. As noted earlier, once loonshots are successful, they tend to evolve into franchises, replacing all-new innovation with incremental improvement. But Bahcall argues that if you pay close attention to the relative weight that stake and rank have in your organization, you can balance growth with continued innovation. We’ll explain what stake and rank are and how to manage them, but first, we’ll look briefly at the theory behind Bahcall’s ideas.

(Shortform note: Maintaining innovation is only one concern for organizations as they grow. In Start With Why, Simon Sinek points out that growth can cause an organization to forget about the purpose that brought it success in the first place. But even if innovation isn’t your central purpose, the principles Bahcall outlines might help you keep your eyes on whatever it is about your organization that really matters to you.)

Phases of an Organization

One of the central ideas in Loonshots is the notion that organizations operate in different phases, a concept Bahcall borrows from his background in physics. Matter can exist in one of several different phases—the basic ones are solid, liquid, gas, and plasma. Bahcall says that groups also exist in phases that are tied partly to group size.

He makes an analogy to traffic jams: When a road reaches a critical mass of traffic traveling at a certain speed, spontaneous traffic jams become a certainty. The system snaps from free-flowing to jammed. Bahcall argues that the principle of group phase dynamics explains why organizations tend to leave innovation behind as they grow: Once an organization reaches a certain size, it changes from the loonshot phase to the franchise phase.

(Shortform note: Bahcall relates these ideas to what he calls the “magic number.” He uses several anecdotes and a mathematical formula to suggest that an organization changes phases when it hits a specific number of members (150). His explanation of this number is somewhat confusing and the evidence for it (by his own admission) is questionable. More importantly, his only point in introducing the magic number is to argue that you can raise that number, which is really just a more complicated way of saying that you can maintain innovation as your organization grows. Therefore, we’ve left out the details of Bahcall’s magic number theory and focused instead on his suggestions for maintaining innovation.)

The good news, Bahcall says, is that system phase changes are dictated by control parameters—and size is only one of these. In the case of traffic, the main parameters are the number of cars and their average speed. That means you can prevent jams by reducing the number of cars or by lowering the speed limit and introducing stoplights. By extension, Bahcall argues that by manipulating the parameters that control organizational behavior, you can maintain innovation even as your organization grows.

Limitations of Bahcall’s Phase Theory

It’s worth noting that Bahcall’s theory of phase management is only relevant to organizations that combine loonshots and franchises under one roof. Vannevar Bush didn’t manage the OSRD’s phases. The OSRD was a self-contained loonshot think tank. It never intended to develop into a franchise; it existed to provide loonshots to the already franchised military establishment.

Bahcall himself points out that some industries consistently operate with this level of separation, as with the previously mentioned symbiosis between biotech startups and big pharmaceutical companies. In these cases, phases—and the principles that govern them—don’t seem particularly important.

Besides, sometimes your goal might be to grow your loonshot into a franchise. If that’s the case, your concern isn’t managing the phase shift so much as getting to it in the first place. Instead of worrying about the magic number, you’d probably be better served by thinking about growth mechanisms and about how to get customers to adopt your idea as soon as possible.

Stake and Rank Control Organizational Behavior

Bahcall argues that besides group size, the two main factors that control organizational behavior are stake and rank. He says that these are the two motivating forces in any organization and that they are in constant competition with each other. According to Bahcall, when the importance of rank outweighs the importance of stake, the organization snaps into the franchise phase.

Bahcall notes that in a small, new, innovative business, everyone has a high stake in the organization’s success, and their rank relative to each other is not that important as compared to their stake in the business’s success or failure. Once the business is established and heading into the franchise phase, rank (which comes with promotions, job titles, better salary and benefits, and so on) becomes more important.

(Shortform note: Throughout this section, we’ll also be looking at organizational politics. Bahcall sees organizational politics as a byproduct of the franchise phase and seems to regard them as the generally undesirable opposite of project work, by which he means both innovative development and a general focus on the quality of your work.)

Therefore, Bahcall says, the key to keeping an organization innovative as it grows is managing the relationship between stake and rank. He identifies four key factors that go into this balance:

1) Salary step-up: How much more money you earn for a promotion. The more money conferred by higher rank, the more incentive you have to engage in politics rather than innovation.

2) Span of control: How many direct reports each manager has (the lower this number, the more managers there are relative to the size of the organization). The more managerial positions exist (the smaller the span), the greater your incentive to work your way up the ladder rather than focusing on innovation or quality.

3) Equity fraction: How much your pay is tied to the quality of your work, how much the quality of your work affects the overall success of the company, and how much that overall success is reflected in your compensation. The less equity you have in the company, the less incentive you have to innovate and the more incentive you have to politic your way into better compensation.

4) Organizational fitness: A parameter that measures how well the company matches workers’ skills to projects and how important politicking is at the company. High fitness means that the company assigns workers to projects they are skilled at and places little importance on politics. Low fitness means companies give workers projects they are ill-equipped to handle and allow politics to dominate the workplace. In a low-fitness organization, you have little incentive to do excellent work or to focus on innovation.

Stake and Rank Control Job Satisfaction

While Bahcall is interested in the stake-rank balance because of its role in maintaining innovation, this same balance may also help control how satisfying a job is.

In How Will You Measure Your Life?, Clayton Christensen argues that rank-related factors like salary, benefits, stock options, and other perks are all incentives. People will work for incentives, but as Christensen points out, many people (like unpaid volunteers) are willing to work hard in exchange for small incentives or no incentives. And as Bahcall points out, incentives can also cause people to do the minimum required to earn their incentives, or even to game the system to earn better ones.

On the other hand, stake-related factors like appropriate challenge, responsibility, the chance to learn and grow as a person, and the opportunity to make a difference in the world are motivators. Christensen says that whereas incentives need to be in place to make sure that employees don’t get dissatisfied, only motivators can make people satisfied with their jobs.

So balancing stake and rank actually has multiple benefits. Not only will a good balance help maintain innovation, as Bahcall suggests, but it’s also likely to keep your workforce happier and more fulfilled.

Balancing Stake and Rank

Now that we understand the variables that affect the magic number equation, we can look at ways to manipulate those variables to sustain innovation as an organization grows. Bahcall offers the following suggestions for adjusting the balance between stake and rank:

Use Rank Carefully

Bahcall points out that you can directly control the influence of rank if you choose the correct spans (number of direct reports per manager) for each part of your organization. Bahcall explains that loonshot groups do best with large spans (fewer managers). That’s because large spans encourage creativity rather than control. Bahcall also notes that the fewer managerial positions there are, the less anyone cares about rank and the more peers are likely to help each other with projects rather than competing with each other for promotions. Plus, he says, creative people will be surrounded mostly by their peers, and that’s whose opinions and input they respect most.

(Shortform note: Also, the fewer managers there are, the less micromanaged the creatives in your organization will feel.)

On the other hand, franchise groups do better with smaller spans (more managers). As Bahcall explains, smaller spans encourage control and consistency, which are more desirable than creativity in the franchise phase. You don’t want military officers or soldiers experimenting with new tactics or new technologies in the middle of a battle, and you don’t want an established company completely changing its products or its business strategy on a whim.

(Shortform note: Along with span of control, it’s also worth considering how centralized or decentralized control is in the organization. In Turn the Ship Around, L. David Marquet promotes a leader-leader model in place of a traditional leader-follower model. Even in a rigid military rank structure (Marquet’s book is based on his tenure as a naval submarine captain), Marquet shows that it’s possible to decentralize control in order to promote better and more flexible decision-making at all levels. Techniques for doing so include managers letting supervisees make their own decisions rather than jumping in and asking employees to take responsibility for their own work rather than monitoring their every move from the top down.)

Maximize Project-Skill Fit

In addition to thinking about how you manage your teams, you should also pay attention to how you assign team members to projects. Bahcall points out that employees who are underskilled for a project stand little chance of success. They are likely to be frustrated or unhappy and to get fired when their project fails. On the other hand, employees who are overskilled will do a good job quickly, then get bored and have little left to do but play politics.

Bahcall recommends aiming for an ideal balance in which employees are challenged but have a high chance of success. He says management should intervene proactively if the fit is bad. He believes it’s worth spending money to get project-skill fit right, because your projects will do better, your employees will be happier, and talented candidates will want to work for you.

How to Recognize and Manage Fit

In Radical Candor, Kim Scott discusses the issue of employee fit in detail. She identifies five types of employee fit and gives suggestions for managing each type:

• Superstars—people with high performance and rapid growth; Scott recommends keeping these people challenged, staying out of their way, and not necessarily trying to promote them to management.

• Rock stars—people with high performance and gradual growth; Scott recommends supporting rock stars by giving them fair reviews (by which she means judging them based on their work and not just on their managerial potential) and by consistently recognizing their achievements.

• People with low performance and no growth; if there isn’t a better fit for such a person elsewhere in the organization, Scott says you probably have no choice but to fire this person.

• People with low performance but expected rapid growth; Scott says this person may be suffering from a difficult transition into his or her new role, or may be affected by outside factors (such as personal trouble). However, Scott also acknowledges that if you had high performance expectations that an employee isn’t meeting, you may have created the kind of poor fit that Bahcall describes above.

• Mediocre performance; Scott suggests that sometimes mediocre performers can be pushed toward growth, and other times their stagnation may be a sign that they are not a good fit for the organization.

Use Smart Incentives

A final way to raise the magic number is to craft incentives that promote the sorts of behaviors you want. Bahcall warns against large step-ups in salary or in hard equity (stock options and the like). He argues that large step-ups encourage people to focus more on promotions (and any related politicking and backstabbing) than on their projects. Bahcall says this problem is especially pervasive at middle management, where it is hardest to evaluate performance and where, with the wrong incentives, the best way to improve your position is to lobby for promotion rather than doing your best project work.

Instead, Bahcall suggests that it’s better to use soft equity: Put people in positions where they are motivated by earning respect and recognition from their peers, not just by earning more money. Bahcall says this goes hand in hand with giving people autonomy and visibility, so that the pressure is on to do a good job rather than to network and self-promote.

(Shortform note: In The Five Dysfunctions of a Team, Patrick Lencioni offers a few more suggestions for creating soft equity, such as creating public goals and standards and rewarding teams rather than individuals for meeting these goals. Doing so, he says, creates public accountability and team spirit and discourages individualistic behavior.)

Likewise, Bahcall points out the need to recognize and avoid perverse incentives—incentives that sound good but actually encourage unwanted behavior. He recommends hiring a specialist to be in charge of incentives (a chief incentives officer). His reasoning is that good incentives are a strategic challenge and require considerable thought and innovation in their own right. (Shortform note: Literature gives us one example of perverse incentives: 19th-century novelists like Alexandre Dumas were often paid by the word or by the line, which encouraged extremely long books (to maximize word count) and long dialogue exchanges of one or two words at a time (to maximize line count).)

Traps That Kill Companies

If you’re trying to build an organization around loonshot ideas, Bahcall’s rules can help—but only if you follow them carefully. Otherwise, you can fall into one of several traps.

Blindspots

Bahcall warns that sometimes, proponents of product loonshots focus on only the newest, biggest, best products or technological innovations and thereby become blind to the innovative strategies that allow seemingly lesser competitors to outdo them.

For a dramatic example of blindspots in action, we can look at the US’s fortunes in the wars that followed World War II. World War II, as Bahcall notes, was in part a war of technological innovation, a race for the newest, best airplanes, tanks, and armaments. The US got good at this game, thanks to the OSRD, and have continued to develop powerful military technologies ever since. But in Vietnam, they encountered for the first time a new strategic innovation in the form of guerrilla warfare against a nebulous opponent rather than open combat against a regular standing army. They were unprepared for this new strategic paradigm, and no amount of technological superiority could carry the day.

(Shortform note: Blindspots can afflict strategic innovators as well. Blockbuster built their business not around a product, but around the service of video rental. Eventually, they fell prey first to a strategic innovation (Netflix’s mail-order service) and then to a product innovation (streaming video, which, ironically, they helped develop).)

The Moses Trap

Bahcall argues that a company can run into trouble when it is led by a strong visionary who dictates the exact course of innovation. Bahcall likens such leaders to the biblical figure Moses since they lead their organizations through sheer force of will and vision. In cases like this, the company achieves the first rule (separate innovators and implementers) but fails at the second (foster dynamic exchange between the two groups). In effect, the company becomes so enamored of its technology that it doesn’t think enough about how that technology will be implemented.

Bahcall gives the example of Edwin Land at Polaroid. Bahcall says that Land loved film technology and kept Polaroid focused exclusively on developing new film cameras even when magnetic tape (VCRs and camcorders) and digital cameras entered the market. Bahcall points out that Land did this even though he knew that digital technology was viable, having previously developed and championed digital imaging for US intelligence satellites. Unfortunately for Polaroid, the newer technologies were cheaper and more convenient for users, and other companies soon undercut Polaroid’s market.

(Shortform note: In Originals, Adam Grant points out a different kind of Moses trap that happens when people with previous success in one domain get overconfident in their abilities to predict success in other domains, like when Steve Jobs, Jeff Bezos, and John Doerr all got behind the Segway, which ultimately flopped.)

Bahcall says that Polaroid’s missteps also highlight the fact that leaders should manage the exchange between innovators and implementers rather than managing any specific technology. Land got too personally invested in film technology and didn’t stop to ask whether it was still the best technology in the marketplace.

(Shortform note: In Start With Why, Simon Sinek suggests that you can help avoid falling into this kind of trap by staying focused on why your organization does what it does. In Bahcall’s terms, asking why forces you to place innovation in dialogue with implementation. Asking why might also help you avoid blindspots. For example, if Land had focused on Polaroid’s why (providing people with affordable and convenient imaging technology) instead of the what or how (film-based cameras), he might have capitalized on his background in digital imaging to get into the digital camera market early.)

The PARC Trap

The other thing that can happen if an organization achieves separation without exchange is that loonshots can be kept so separate that the innovative ideas never make it to implementation and instead die in the lab. Bahcall calls this the PARC trap, after Xerox’s former research division. PARC developed a whole host of innovations that Xerox never brought to market, but which went on to shape modern computing.

(Shortform note: Xerox’s short-sighted neglect of PARC is well-known and shows that sometimes companies simply fail to capitalize on good ideas. But as Christensen points out in The Innovator’s Dilemma, big businesses also face some legitimate obstacles when it comes to adopting new innovations. For example, a business might have a hard time marketing new products to different customers, or new innovations might not fit into a company’s business plan—not to mention its manufacturing setup.)

Shortform Bonus Rule 5: Build a Critical Mass

Separate from his four rules, Bahcall introduces the idea of critical mass as a final component needed for sustainable loonshot development. To explain the idea of critical mass—and how it benefits both innovators and implementers—let’s return to our example of the drug industry as a symbiosis between biotech startups and pharmaceutical giants. As this industry demonstrates, critical mass has two aspects:

One, you need enough loonshots to have a reasonable chance of success. A small percentage of loonshots work out. Those investing in them need to invest in a large number of loonshots to hedge their bets against failure. This is one way that a phase-separated company—or industry—benefits both sides. The ongoing success and stability of the franchise side funds the experimentation of the loonshot side.

(Shortform note: In Antifragile, Nassim Nicholas Taleb makes a similar point when he recommends investing in good people rather than in good ideas. Because loonshots are so unpredictable, it’s hard to know in advance what a good idea even is. Plus, as Taleb argues, trying to dictate the course of a loonshot only limits its potential.)

Two, for people working in the loonshot side of things, you want enough loonshot factories that if something goes wrong, there are other places around where you can continue your work. If your biotech startup folds, or if it rejects your drug idea, the idea can survive if it has somewhere else to go. Bahcall gives the example of scientific innovation in imperial China versus in early modern Europe. When an idea or project was shut down by the emperor, it had nowhere to go and died for good. When an idea was shut down by one royal patron in Europe, the inventor or scientist could seek patronage elsewhere and continue the work.

(Shortform note: This kind of safety net can also encourage innovators to build up a critical mass of new ideas, since they won’t have to worry about what will happen if any one idea fails. If your loonshots have only one outlet (whether that’s the Chinese emperor, a rigid parent company like Xerox, or a Moses-type boss like Edwin Land), you’ll either have to restrict your innovation to ideas you know will be approved or else risk wasting time and energy and maybe losing your job. But if you work for a company with a good loonshot factory (like Pixar) or in the loonshot side of a divided industry (like pharmaceuticals), you’re freer to explore even the looniest of ideas knowing that some missteps here and there are expected and even welcomed as a byproduct of innovation.)