In Uncertain Times, Embrace Imperfectionism

If you are an Amazon customer, chances are you have encountered household names such as Amazon Prime and Zappos. It’s less likely that you would have noted the baby steps Amazon took to expand beyond its core business into consumer financial services.

The acquisition of TextPayMe, the investment in Bill Me Later, the hiring of a team from GoPago, and the launch of a remote card payment device, Amazon Local Register were modest moves that involved little financial outlay and attracted relatively little attention at the time. They also ended in apparent failure, with TextPayMe closing in 2014, Local Register withdrawing from the market in the face of competition from Square, and Bill Me Later being acquired by rival PayPal.

Yet today Amazon is a powerhouse in consumer finance, boasting a 24% user share in the U.S. for its pan-economy Amazon Pay service, and is positioned to develop further as a global finance player. How did the company craft success at scale from a series of pint-sized and seemingly unpromising moves?

The answer is that Amazon is an imperfectionist, a concept we’ve developed over several decades of helping companies and nonprofits, and one that we believe is vital for organizations striving to prosper in today’s uniquely uncertain economic environment. As we will explain, imperfectionism is an approach in which companies grow not by following a strategic framework or plan, but through multiple and frequent experimentation in real time, incrementally building up valuable knowledge, assets, and capabilities on the way.

Why a New Approach Is Required

We all experience it: Change is accelerating, with uncertainty and threat of disruption in every industry segment. Artificial intelligence, programmable biology, robotics and other technologies are transforming existing industries and spawning new ones. External shocks cut across boundaries and impact all businesses, even those that appear to enjoy structural advantages.

In this volatile and fluid environment, established approaches to strategy development no longer work. There is no stable industry structure or new market equilibrium to return to. The pace of change and rise of global winner-take-all competition means that conventional product-market-structure approaches, as well as core competencies thinking, are difficult to implement in practice and may yield misleading answers. The player to topple a dominant company in a particular space is more likely to enter from outside with a de novo approach than to be an existing rival. Just look at the battle of Instagram, Quibi, and TikTok in the short-form video social media space — the now-winning player was unknown five years ago.  We need responsive strategy approaches to match external volatility.

One repercussion of this changed environment is that most company strategic planning processes are unsuited to an uncertain and fast changing world, leaving many management teams stuck in a nervous wait-and-see posture.  Other company leaders avoid stasis by making panicky bets, including “leap before you look” acquisitions in an attempt to “own” new spaces, typically with little success.

What should you do instead?

Embrace the Imperfect

As we describe in our book, The Imperfectionists, Amazon didn’t wait for a moment of strategic clarity, nor did it use its large balance sheet to acquire an indigestible bank. It edged out into uncertainty by taking small steps: investing in young companies with interesting technologies, hiring a team from a failed fintech, and launching internally developed services and tools. It ran multiple, parallel initiatives and learned what worked and what didn’t. It built valuable capabilities and assets with moves that were mostly low consequence and reversible. There is no notion of strategic frameworks in its approach to building a major financial services business.  The end result is what we term a staircase of initiatives, a pragmatic revealed strategy.

This step-by-step approach was no accident — Amazon followed the same modus operandi in its entry to commercial finance and in its creation and dominance of cloud computing. It now appears to be doing something similar in health care.

Under today’s conditions, we believe that real-time problem solving, not theoretical frameworks, should be the heart of strategy development. This starts with a hypothesis, often framed as an audacious question that when answered will provide strategic insight on the first steps of a solution staircase.

Strategy development then moves to seeing things through multiple lenses or vantage points, typically anchoring these outside the company in its ecosystem or beyond. Looking at your business through the eyes of your suppliers, customers, current rivals, and potential outside entrants will give you much more perspective on both threats and opportunities than you will get from staying inside your company’s prevailing mindsets and routines.  Perspective-taking workshops at all levels of the organization can help surface novel observations that can be the nucleus of a new strategic path.

Next it seeks novel data on these perspectives via experimentation, augmenting this new data by crowdsourcing external ideas and technologies to bring collective intelligence to bear. To be sure, few organizations are adept at having others toil in their garden, but one path to accessing fresh capabilities is crowdsourcing new technologies and approaches.

This leads to a pragmatic plan to step into uncertainty, gathering more information from small moves, while adding capabilities and assets, laying off risk to others where possible, and learning from mistakes and successes. It involves recognizing that strategic decisions are probabilistic wagers, and moves are made commensurate with an assessment of the odds.

Large enterprises, in particular, struggle with this. They are often reluctant to think in terms of probabilities that, in turn, gives rise to a powerful form of risk aversion, as documented in a 2020 HBR article by Dan Lovallo, Tim Koller, Robert Uhlaner, and the Nobel laureate Daniel Kahneman. They found that that managers insist on making risk-averse choices even where favorable odds are clear. A starting point to overcome this deep-set bias is to insist that all discussions of strategies include probability estimates and ranges of potential outcomes.

Our research suggests that this approach to strategy under uncertainty is being effectively employed not just by the likes of Amazon but by other savvy firms and nonprofits alike. Here are four examples, each illustrating one of the four main components of the process.

Invisalign: Seeing Through New Lenses

Until the early 1990s, orthodontic braces had hardly changed for decades.  Incremental improvements came from inside the practice of dentistry, but significant innovation was missing. Kelsey Wirth and Zia Chishti, MBA students at Stanford Business School with no qualifications in dentistry, worked to apply a fresh and different lens to orthodontics.

Zia only had the financial means to correct his teeth as an adult, shortly before business school.  When he finally got his braces off, he noticed something interesting: His retainer was almost invisible, but it still moved his teeth, which he could feel if he forgot to wear it for a few days. This small insight inspired a revolutionary hypothesis: If a simple plastic retainer could move teeth a little, perhaps the technology could be improved to move teeth a lot, but without affecting a person’s appearance as much as conventional braces.

Chishti and Wirth proposed a series of transparent, custom-manufactured removable plastic alignment retainers, developed through software-designed treatment plans. They called their solution Invisalign, emphasizing that the aligners would be inconspicuous to others. By changing to a new set every few weeks, the patient’s teeth would gradually move into the desired position without the need for metal braces.

By looking at the problem of straightening teeth through these novel customer and technology-centered lenses, Zia and Kelsey made orthodontic treatment less painful, less embarrassing, and more affordable.  Today the company has transformed orthodontics and has a market capitalization over $15 billion.

SpaceX: Relentless Experimentation

Everyone knows about software companies using A/B testing to refine products. But could this approach create strategic success in the ultimate heavy industry? In only two decades, SpaceX has massively transformed near-Earth space travel through relentless experimentation.

SpaceX engineers have learned to make small moves that add up to big improvements by repeatedly testing new hypotheses. This rapid test-and-improve approach enables them to acquire new information at a reasonable cost and to more quickly travel down the experience curve. They refer to their decision system as “fly, test, fail, fix” and accept that failure (“unplanned disassembly”) is a necessary part of testing the boundaries of what is possible. They’ve pioneered lower-cost 3D printing for rocket components, as well as making more components re-useable, and shifting away from expensive external contracting, with 80% of components produced in-house.

Moving down the cost curve required substantially increased frequency of launches. During the 30-year period of its space shuttle program, NASA averaged 4.5 flights per year. In 2021 Space X launched 31 rockets in orbit, up from 26 the prior year.  In early 2022 it was launching three to five missions per month.  More frequent launches and other innovations have dramatically reduced the cost of putting a kilogram of mass into space.  For the period from the 1970s to the early 2000s, this cost was largely unchanged, costing an average of $54,550 per kilo.  SpaceX has slashed the per-kilo cost down to $2,720, a reduction of 95%.

TNC FishFace: Tapping Collective Intelligence

The Nature Conservancy (TNC), a global conservation organization interested in protecting endangered fish stocks, was searching for effective strategies to implement conservation measures onboard fishing boats at sea. TNC wondered if digital techniques could be deployed by fishing boats to identify endangered species that they could return to the sea while still alive.

But TNC did not have the necessary AI skills to develop a solution itself, so it decided to crowdsource solutions via Kaggle, an online platform that hosts machine-learning competitions. The TNC competition attracted 2,293 entries over five months with a $150,000 prize. After selecting a winning entry, TNC now has a prototype operating with 90% to 95% accuracy.  It is called FishFace.

To solve the complex fisheries regulatory compliance problem, TNC anchored outside its own organization’s capabilities, inviting the best AI developers in the world to provide a novel strategic solution.

AI swarm platforms are a variant of this approach and are particularly useful for predicting uncertain events.  A study at Stanford medical school found that groups of doctors using Swarm AI algorithms were 22% more accurate in making diagnoses than the most advanced deep-learning algorithm that only used historical data.  The difference is that participants in swarms “think together” in real time and converge on solutions through interactions governed by algorithms.

Quibi versus TikTok: Anchoring Inside vs. Outside

In each of these examples, a substantial part of the strategic insight that led to breakthrough strategic ideas is the idea of anchoring outside the conventional boundaries of the industry.  Getting this right isn’t easy — consider the experiences of Quibi.

Quibi was a short-form streaming platform launched with great fanfare in April 2020; by December 1, the company was dissolved at a loss of more than $1 billion.  An A list of investors included Alibaba, Disney, Google, and Goldman Sachs.  An all-star management cast was led by Meg Whitman, the former CEO of eBay and Hewlett Packard, and Jeffrey Katzenberg of Dreamworks. Investors were sure that this was a team of insiders who knew what the market wanted.

The Quibi app was downloaded 4.5 million times during the second quarter of 2020, but by the third quarter of 2020 Quibi had only 710,000 paying subscribers and it was losing 90% of early users after their free trials expired. Quibi was anchored inside the old Hollywood model and failed just as short-form video competitor TikTok was explosively rising, with 315 million downloads in the first quarter of 2020.

The difference?  TikTok embraced the collective intelligence of its users with a highly sophisticated AI-curation of user-provided content.

. . .

When you look in retrospect at Amazon’s stepped entry into consumer financial services or the other examples above, they look like the work of a master strategist. But that disguises the reality: There were many small moves to build the capabilities and assets that supported those positions, including missteps and failures along the way. Strategically nimble organizations explicitly map potential pathways into new areas and update their models with their experiences. They understand the difference between modest cost, reversible strategic moves that build position and bet-the-farm decisions.