Innovation Starts with Defining the Right Constraints

In 2012 MIT Professor Amos Winter was asked to develop a lighter, cheaper prosthetic leg for the huge Indian market. And not just a bit cheaper: the new limbs needed to be 90% cheaper than those sold in western markets to meet the needs of the over half a million amputees unable to afford prosthetics that often cost tens of thousands of dollars and lasted only 2-3 years. Under these dramatic constraints, Winter’s team went back to fundamentals and reframed the problem: what could the science of movement teach us about how to design and deliver a radically different prosthetic? Rather than taking a traditional approach, which sought to mimic a human foot, the team focused on a tunable but passive foot design that would instead mimic lower leg movements. By 2019, Winter’s team had unveiled their new, low-cost solution — one that could cheaply and easily be tailored to a patient’s weight and height. It was fundamentally different from existing products in terms of cost, design, and material. This achievement was only possible because the initial constraints imposed on the challenge forced a complete re-thinking of the problem.

This story reminds us of a consistent lesson from the research on innovation: while unshackled creativity might intuitively seem to be the best route to novelty, actually some of the most innovative outcomes are produced when innovation is constrained.

But the type and quality of constraints matter. Leaders we talk with (especially in the public sector) typically impose two common constraints on their innovation teams: budget and risk. They tell teams their innovation must cost no more than a certain amount — a figure based on assumptions about what kind solution the team will deliver. Relatedly, the constraints on risk communicate to the team: “don’t do anything too risky, especially something that might cannibalize existing business” or “don’t fail because that’s not what gets you promoted.”

Unfortunately, neither constraint is particularly helpful in practice. In fact, they often produce unintended negative consequences by narrowing innovation to tried and tested solutions and precluding a radical reframing of the problem. As a result, a team’s potential to reimagine solutions is fundamentally hampered from the start.

But during the Covid-19 crisis, the urgency of the problem prompted traditional budget/risk constraints to be replaced by two alternatives — one that we believe support genuine innovation:

1. constraining outcomes, e.g. “the new antibody test must have this level of sensitivity” or “the new ventilator must have this functionality”;

and then, within the context of this new target outcome:

2. constraining time, e.g. “we need a reliable test by Q2” or “the new hospital must be ready in a week.”

Constraining outcomes and time has precedent in a number of existing strategy concepts. One of us (Johnson) has shown how, in the content of strategic change, leaders need to specify clear, long-dated outcomes if managers are to be able to respond with the best ideas about how to deliver a new strategy or change. Another similar concept is that of the Commander’s Intent. Here, military leaders give their soldiers a clear message about what needs to be achieved and by when, e.g., “we need to capture that territory by the tomorrow,” but leaves it to those on the ground to decide how to do this. Notice that the Commander’s statement doesn’t say: “you can’t spend more than this much” or “we can’t lose any men.” Instead, it states outcome and time constraints so soldiers own how this gets done and can respond as on-the-ground conditions change and enemy tactics emerge.

So, why do these alternative constraints — outcome and time — work better, and how can we most effectively use them?

The Outcome Constraint

The key feature of the outcome constraint is, unsurprisingly, its focus on the end result. It defines what does a good solution does for users, payers, or investors rather than the process or rules by which it’s produced. We have identified three approaches by which leaders impose effective outcome constraints.

First, the organization can choose to set a single, big new constraint — one that forces people to think about the problem in a fundamentally different way. Make a new prosthetic leg that’s 90% cheaper than existing products is an example of this.

Connect the world’s remotest areas to the internet is another. This was the outcome-based challenge that Google, through its ‘X’ unit (formerly Google X), set the ‘Project Loon’ team. By focusing on the outcome rather than making assumptions about how it would be delivered, this team developed the idea to connect remote areas using giant stratospheric balloons rather than by laying cable in the ground. what might seem like a crazy idea is now being realized: Project Loon is now partnering with telecoms companies to provide internet services in remote areas of Africa and Latin America.

Of course, such 10x rather than 10% projects won’t always succeed. But even if the ideas they produce don’t all fly, they are often a source of genuinely new ideas. And that’s because having big, audacious goal creates an environment where the ideas need to be so big and new in order to deliver the outcome, that big, new ideas are exactly what you get.

A second approach is to set conflicting constraints. This is where two or more seemingly conflicting outcomes — ones that appear impossible to deliver together — force a fundamental re-evaluation of the solution space. When designing what became the Lexus line, Ichiro Suzuki, Toyota’s chief engineer, stipulated that the new car needed to be faster, lighter, and more fuel efficient than existing luxury sedans. The order was full of contradictions. Making a car faster usually meant having a bigger, heavier engine; making it lighter without compromising power meant stripping out luxury that was essential for this segment. So, the Lexus team returned to fundamentals and re-evaluated their most basic assumptions about how to build a car. Alongside tens of smaller new ideas, they designed and built a first-of-its-kind aluminium engine that made the car 120 pounds lighter, improving weight and fuel efficiency, thereby delivering on a seemingly impossible demand

A third approach is to specify what’s not allowed. This is particularly helpful when there are many possible solutions available. For example, when MIT was working out how to bring students back to campus during the pandemic in Fall 2020, the university initially operated as if all options were on the table. As a result, progress was hard and slow. MIT then decided that some outcomes would not be acceptable — for example, ones that would prevent students scheduled to graduate in 2021 from doing so. This immediately delineated who were the priority students. And if those people were on campus, this told them how many other people couldn’t be. By specifying what was not acceptable, MIT reduced the scope of the problem and created a more focused, and therefore faster, process.

Having set a clear outcome constraint, leaders can now turn their attention to the second alternative constraint.

The Timeframe Constraint

The second useful constraint is a deadline. The relationship between time pressure and performance is well established: the Yerkes-Dodson Law has been around since 1908. People benefit from being under some pressure because it makes them focus and work faster, so long as people believe that not meeting the deadline has real consequences — we all know the difference between a real deadline and a fake one — and provided the pressure is not so high that it triggers stress and other performance-killing reactions.

But why is a time constraint more effective for innovators than a budget constraint? It’s largely because people experience the passage of time much more viscerally than they do the running down of a budget, where the information on a project’s cash position is often invisible to front-line managers and accountability for the over-spend isn’t always clear.

The key is to set timeframes for delivering innovation that are appropriate for the outcome you’ve established. This contrasts with the all-too-common habit of linking timeframes to the internal cadence of your business, e.g., the annual planning process, the budgeting cycle etc. In contrast, a timeframe that is “appropriate” means one that is long enough to enable true innovation, which often requires an investment J-curve, but short enough to mean people feel a sense of urgency to deliver and to move through multiple learning loops as quickly as possible.

So how do you set the right timeframe? Sometimes events or competitors take care of this for you, but often it falls to leaders to set internal deadlines. We’ve found three approaches work particularly well.

The first option is to set the overall timeframe and then subdivide it, structuring the overall deliverable into a series of workstreams. For example, one timeframe could be imposed on the idea generation stage, where teams explore the problem. Having called time on brainstorming, and chosen the two best ideas, leaders would then choose a second timeframe for the next stage of the work, where these ideas are developed into operational (i.e. testable) propositions.

More appropriate for organizations than projects, another option is for leaders to set big, multi-year strategic objectives — for example, a certain market share or cost ratio — broken up into quarterly milestones to gauge progress. The combination of long-term targets and shorter-term milestones provides the best of both worlds: a large enough target long enough away to enable fundamental change but with regular milestones to make the size and scale of the task feel both more manageable and more immediate. This simultaneously gives the work a greater sense of urgency (thanks to the short-term milestones) but also the opportunity for managers to undertake bigger, riskier projects with multiple iterations, because the long-term target gives them the runway to do so

A third approach, is simply to ask, “what can we accomplish in a week/month/year?” For example, agile approaches often focus on how many new features and elements can be incorporated in a given period, rather than spending time and effort on setting more micro-objectives before beginning work. Most salient in settings where urgency trumps all other priorities, this is another way for leaders to reduce the degrees of freedom in which teams operate, thereby emphasizing the speed with which they need to produce workable solutions. For this option to work, leaders will need to broadly agree on the outcomes that need to be delivered but by defining the scope purely in terms of the time available, effort is focused on what can be done quickly.

Combining Outcome and Timeframes

While both outcome and time constraints are, we argue, individually better than budget or risk constraints, they are at their most effective when combined.

We’ve seen this in the development of Covid-19 vaccines. There were clear outcomes: it protects you from the virus without doing you more harm than good. The timeframe was also clear: as soon as possible. Regulation and testing protocols provided guidance around the outcome but urgency forced people to think differently about how testing and evaluation could be done.

The sequence in which these two constraints are set also matters. Timeframes typically work best once an outcome has been agreed. DARPA’s development of autonomous vehicles is a good example of combining outcome and time constraints. The target outcome was to build an autonomous vehicle that could complete a three-mile journey, and DARPA specified that the prototype had to be delivered by on a certain date. The addition of the deadline (which, like all DARPA deadlines, was short) meant that the exploration work — which could have continued indefinitely — had to be completed fast in order to leave time to develop the prototypes. This combination of ambitious outcome and time constraints meant entrants quickly produced prototypes that were genuinely innovative.

So, how can leaders make this approach work for them and their teams? They should start looking at their own working style. We’ve identified a few essential skillsets and mindsets that leaders should cultivate.

• Making decisions around outcomes/time rather than budget/risk requires leaders to shift to how they use their own To become more future-focused, leaders need to spend more time with technical, operational, and future-focused people both inside and outside of the organization. To do that, they need to reallocate time away from controlling risks or budgets within the business. That is likely to require a fundamental redesign of their planning and budgeting cycle.
• To set effective outcomes, leaders must develop a perspective about how the future may unfold. Leaders we’ve worked with who develop this often start their strategic narratives with, “Because we believe [fill in the blank] about the future, we will be targeting this particular technology, market, or customer segment.” Within this context, they then develop clear hypotheses — each mapped to these target outcomes — with which to test and refine their initial point of view. For example, Noubar Afeyan, one of the early backers of Moderna (and a former colleague at MIT Sloan), has described how he first envisions several possible futures where science or technology can address an unmet need. Then he’d choose the one he wanted to pursue first by trading off the likelihood of reaching it against the likely impact of doing so. Having chosen the priority to pursue, the paths to reach it are then developed and tried out.
• To shape better outcome-oriented decisions, leaders need to develop and reward scenario-building skills — recognizing that these are different from “planning” skills. Too often, future-focused scenarios are insufficiently different from the current state of a market or operation, largely because the people shaping these scenarios are “planners” rather than “futurists.” This may mean you need to hire in specialist consultants for this part of the process, but it’s worth it if it generates some truly different scenarios.
• And because the trend data will always be contradictory, with little obvious distinction between what is signal and what is noise, teams must be able to thrash out their points of view without fear or embarrassment. That requires teams to have built sufficient psychological safety so they can have productive fights about what the conflicting data tells them and which of the myriad options they should pursue.
• To impose effective deadlines requires bravery. In particular, leaders must be prepared to potentially waste investment on the development of multiple ideas because too much time spent developing or polishing ideas at the start reduces the time you have to get your idea to market. Time is finite, so trade-offs will have to be managed.

We believe that by adopting two different constraints — outcomes and time — organizations can become more innovative. They can then afford to be less constraining of both budget and risk — the traditional constraints — largely because the clear imposition of outcome and time-bound constraints helps deliver more valuable innovation (the objective of a budget constraint, after all) and at lower risk. As the people working in your innovation teams will tell you, that’s very good news for genuine innovation.