The Innovation
Quotient

A Note From Josh & Kaiser

The Innovation Quotient is a quarterly research publication from Platypus Labs. Each issue draws on peer-reviewed research, industry data, and organizational case studies to surface findings relevant to leaders navigating disruption, competing in crowded markets, and building organizations capable of sustained innovation.

This quarter’s theme is Optimizing the Conditions for Innovation. Our last two issues looked at the human skills that matter most in the AI era and at the speed at which innovation now moves. This one turns to the part of innovation that leaders control most directly: the conditions inside an organization that decide how many ideas surface, how quickly they get tested, and whether people feel free to pursue them. Plenty of organizations hire brilliant people and then ask them to invent the future on a calendar packed like a rush-hour subway car.

The pages that follow start with the evidence that conditions shape what talented people produce. Then they look closely at three areas where conditions matter most: experimentation, collaboration, and play. Section 05 also offers a first look at Co-Flow, Josh’s forthcoming book on how groups reach a shared creative state.

We hope you take a few of these ideas and try them with your team this quarter.

Josh Linkner & Kaiser YangCo-founders, Platypus Labs

About Platypus Labs

Platypus Labs is an innovation consulting, training, and research firm helping ambitious organizations build the creative capacity their strategies actually require. We work with Fortune 500s, high-growth startups, and the teams in between: on stage, in offsites, and embedded inside the work.

Josh LinknerCo-founder
Platypus Labs

Kaiser YangCo-founder
Platypus Labs

What’s Inside

• 01Executive SummaryThe report’s key findings and their implication for leaders.
• 02Statistics That Surprised UsThis issue’s most consequential numbers, at a glance.
• 03Why Conditions MatterThe evidence that conditions shape what talented people produce.
• 04The Conditions for Rapid ExperimentationHow fast a test gives an answer, who it reaches, and what happens next.
• 05The Conditions for Creative CollaborationHow groups do their best creative work, with a first look at Co-Flow.
• 06The Business Case for PlayWhat play and unstructured time produce, and where the evidence stops.
• 07Hard LessonsSonos’s app redesign, and what it teaches.
• 08Innovators We AdmireFour leaders who built the conditions for other people’s ideas.
• 09To-Test ListFive small experiments to run this quarter.
• 10Resource GuideWhere to go deeper.
• 11Partner with PlatypusHow we can help.

In Brief

Innovation depends heavily on the conditions people work in, and leaders can change those conditions faster than almost anything else in their organization.

Most organizations already employ capable, creative people. What varies is how much those people are able to produce, and the research in this issue shows how closely that output tracks the conditions around them: how fast an idea can be tested, whether a result changes anything, whether people feel safe offering a half-formed thought, and whether anyone has time to explore. Leaders shape every one of these conditions, mostly by habit. They build up like barnacles on a ship’s hull: a meeting habit here, an approval step there, a raised eyebrow at a new idea. This issue makes the case for shaping them on purpose. It draws on peer-reviewed research, industry surveys covering more than 55,000 workers and leaders, and company case studies.

Four Findings

What We Mean by Good Conditions

Throughout this issue, good conditions are the ones that make it easier for people to come up with ideas, test them, and say them out loud. Bad conditions make each of those steps slow, risky, or pointless. Good conditions can still be demanding. In the research ahead, the projects people called their most creative involved challenging work, scientists with more freedom were still reviewed (just less often), and the playful programs that worked had a set time and a clear purpose. What hurt was a different kind of difficulty: extreme deadline pressure, approvals that added delay without catching problems, and ideas that disappeared into silence.

Where to Start

If we had to pick three places to start, they would be these:

• Open time. Room to explore that nobody has already filled, protected on the shared calendar so no one has to defend it alone.
• A short loop. The trip from an idea to a real customer’s reaction to a decision about what to do next, with fewer waits along the way and a named owner for the decision.
• How ideas are received. Quiet time to think before the group talks, a calm reaction to every idea, and an answer to each one, including a clear no.

To find out which one needs attention first, try the one-meeting exercise in Section 03. Most teams already know where their conditions fall short. Few have been asked.

By the Numbers

An Accidental Experiment

Most conversations about innovation start with people: who to hire, who to promote, who has the spark. The strongest evidence we found says the conditions around those people matter as much, and that most leaders underrate how much of that is in their hands.

The cleanest evidence came from an accident. In 2024, two researchers published a study of manufacturing plants where a supply-chain shortage had unexpectedly stopped production, leaving workers idle with no other task to switch to. In the three weeks that followed, employees who had been through a stoppage produced 58% more ideas than colleagues whose work had carried on. Not every break worked this way. Interruptions that forced people onto other work hurt their creative performance, and planned breaks showed no gain. (Schweisfurth and Greul, Organization Science, 2024.)

The workers showed up with the same skills and the same lukewarm break-room coffee they had the week before. What changed was the room around their work, and their output changed with it. The ideas came from open time that nobody had filled, and most calendars leave none. A leader who wants more ideas may need to protect some open time on purpose, and then resist the urge to fill it.

Two Projects, One Company

One accident could be a fluke, so we looked for a broader pattern. In a study published in 1996, Harvard Business School professor Teresa Amabile and her colleagues asked managers at a large electronics company to think of two projects they had worked on: the most creative one and the least creative one. The managers then rated the work environment around each. The environments around the two kinds of projects differed sharply. Around the most creative projects, people described challenging work, encouragement from the organization and from supervisors, support from their own work group, and freedom in how to do the work. Around the least creative projects, they described more obstacles inside the organization. (Academy of Management Journal, 1996.)

Each manager described both projects, and every comparison stayed inside one company. The teams on the two projects were not necessarily the same people, and the researchers were careful to say their data could not prove cause. Still, every item on that list is something a manager shapes every week, and the managers could tell the two environments apart just by looking back. Most teams hold the same knowledge. A leader can collect it in one meeting using the study’s method: ask each person to describe their most creative project and their least creative one, then listen for what was different about the conditions around each.

What Happens Day to Day

Amabile and her colleague Steven Kramer then went inside the work itself. They asked 238 people on 26 project teams at seven companies to fill out a short diary at the end of every workday, which produced about 12,000 entries. On the days people reported their best moods, they had made progress in their work 76% of the time. That matters because, in the same diaries, people were more creative on days when they felt better, so progress connects directly to creative output. (Harvard Business Review, 2011.)

Time pressure showed up in the diaries too. On days people rated time pressure at the top of the scale, they were 45% less likely to think creatively than on lower-pressure days. (Harvard Business School working paper; Harvard Business Review, 2002.)

Most managers miss the power of progress. When Amabile and Kramer asked 669 managers to rank what motivates employees most, only 35, about one in 20, put progress first. Taken together with the time-pressure finding, the research gives managers two jobs: help the work move forward, and keep deadlines reasonable. Progress is cheap to supply: clear an obstacle, shorten a wait, make a small win visible. Pressure is where good intentions can backfire. A manager who adds tighter deadlines and extra check-ins to speed things up may push the team toward the top of the pressure scale, where creative thinking dropped.

When Conditions Change, Output Changes

If conditions matter this much, people working under different conditions should produce different work, and people who change conditions should see their output change. Two studies, one in science and one on Wall Street, show both.

The Howard Hughes Medical Institute funds scientists for long periods and tolerates early failure. Economists Pierre Azoulay, Joshua Graff Zivin, and Gustavo Manso compared its investigators with similarly accomplished scientists funded under shorter, more frequently reviewed grants. The Hughes scientists produced 39% more publications, roughly twice as many papers in the top 1% by citations, and 35% more papers that drew very few citations. (NBER working paper; RAND Journal of Economics, 2011.) More hits and more misses is the signature of bold work, and it is the pattern the authors predicted: funding that tolerates early failure encourages scientists to take bigger swings. Venture investors know this pattern well. A fund expects many of its investments to disappoint and a few to produce most of the returns, so a portfolio with no failures can mean the fund played it too safe. Many corporate review processes try to eliminate failed projects instead. One quick check for leaders: if none of your new projects has failed in the past year, your review process may be screening out bold ideas before they start.

Wall Street shows the same thing from the other direction. In a study of more than 1,000 star stock analysts at U.S. investment banks from 1988 through 1996, stars who switched firms saw an immediate decline in performance that persisted for at least five years. Stars who moved together with their teammates showed no significant decline. (Harvard Business Review, 2004; Management Science, 2008.) Much of what looked like individual talent belonged to the setting the analysts left behind, like a lead singer who goes solo and discovers how much the drummer was carrying. For leaders, this cuts both ways. A star hired away from a team brings less than the record suggests, and the people and systems around your own stars are part of what makes them stars.

What Leaders Control

To be clear, talent still matters a great deal. A 2021 study of 37 years of U.S. patent data found that differences between individual inventors explain far more of the variation in patent output than differences between the companies that employ them. (Management Science, 2021.)

Talent explains more of the total, and talent can grow. But the conditions a team works in are often the fastest thing a leader can change. They are set by everyday decisions, such as the meeting calendar, the approval chain, and how a leader reacts to a strange idea, and those decisions can change next Monday.

The Conditions Today

Amabile’s diaries date from the 1990s, and recent surveys describe a workday that works against everything above.

In Microsoft’s 2025 survey of 31,000 knowledge workers, 48% of employees and 52% of leaders said their work feels chaotic and fragmented, and 80% said they lack enough time or energy to do their work. Respondents to Deloitte’s 2025 Global Human Capital Trends survey estimated that 41% of their time every day goes to work that does not contribute to the value their organization creates. That figure is easier to grasp on a calendar. If respondents’ estimates are roughly right, 41% of a five-day week is about two days, and few leaders would approve a budget line that cost two days of every employee’s week. Miro’s 2025 survey of 6,148 knowledge workers found that for every hour of creative or strategic work, people spend three hours on maintenance work such as meetings, email, and paperwork, the office equivalent of folding laundry.

Each of these surveys comes from a company that sells a remedy for the problem it measures, but they point the same way as the academic research: the open time, steady progress, and freedom that creative work needs are what a crowded workday squeezes out. The next three sections take these conditions one at a time: testing, collaboration, and play.

What Makes an Experiment Pay Off

Most organizations already run pilots, surveys, and trials, and many still change very little. The conditions around a test decide whether it teaches anything, and the research points to three: how quickly a team finds out whether an idea worked, whether the test reaches real customers, and whether anyone acts on the result. Together they form a loop from idea to test to customer reaction to decision. The shorter that loop, the more a team learns in a year.

How Fast the Answer Comes Back

Speed matters because every answer is a chance to adjust. A team that hears back in a day can try, fix, and try again dozens of times in the months a slower team spends waiting on one result.

The most detailed data on feedback speed comes from software, where Google Cloud’s DevOps Research and Assessment program (DORA), the largest ongoing study of software teams, surveys development teams every year. In 2024 the fastest group, 19% of respondents, could get a change in front of users in less than a day. The slowest group, 25% of respondents, needed between one and six months. An earlier DORA survey found that teams with external approval requirements for significant changes, such as sign-off from an outside board or a senior manager, were 2.6 times more likely to rank among the low performers, the teams that score worst on DORA’s measures of speed and stability, and their changes did not fail measurably less often. (DORA, 2024 and 2019.) The data shows an association rather than proof of cause, but the pattern looks like a speed bump.

The same pattern holds far from software. Stanford researchers gave students the same amount of time to build a device that would protect an egg in a fall. Students who could test their designs along the way built devices that survived an average drop of 6.1 feet. Students who were prevented from testing before the final drop averaged 3.3 feet. (Dow, Heddleston and Klemmer, 2009.)

Approval steps tend to pile up because each one feels safe to the person who adds it. For leaders, a useful exercise is to count the waiting. List every approval, review, and queue between an idea and the first real customer reaction. For each one, ask two questions: when was it added, and what problems has it caught since? Keep the steps that catch real problems or guard against serious risks, such as safety, legal, or security exposure. The ones that only add time, or repeat a check made elsewhere, are the ones to cut, starting with the slowest.

Getting the Test in Front of Real Customers

A test that never leaves the building can only tell you what the building thinks. The evidence favors putting ideas in front of real customers early and asking them often.

A randomized study in Rwanda shows what that habit can be worth. Researchers worked with 274 small businesses and helped half of them spend 12 months systematically asking customers what to improve. Compared with the other half, those firms earned 62% more monthly revenue and 54.5% more monthly profit. Customers who were never asked for feedback bought more too, which suggests the firms changed what they offered everyone. (Kaul, Anderson, Chintagunta and Vilcassim, Marketing Science, 2024.) The detail about customers who were never asked is the telling one: the feedback did its work by changing what the firms sold, which makes a feedback habit a product-development tool as much as a customer-service courtesy.

Large organizations can build the same habit into their structure. Major League Baseball tests rule changes in real games before the major leagues adopt them, using the independent Atlantic League and its own minor leagues as a test bed. By the time the pitch clock (a timer that limits the time between pitches) and related rules reached the majors in 2023, Commissioner Rob Manfred said, they had been tested in approximately 8,000 minor league games. The average nine-inning game in 2023 ran 24 minutes shorter than in 2022, when it lasted 3 hours and 4 minutes. Not every change survived the test bed: a longer pitching distance was tried and dropped. What baseball built is a permanent place where changes meet real conditions before they reach everyone. It is the kind of test bed Sonos lacked, as Section 07 shows.

What Happens After the Test

A fast test that reaches real customers still pays off only if someone acts on the result, and that step fails more often than most leaders assume.

The newest evidence comes from city governments. From 2015 to 2019, 30 U.S. cities worked with a national behavioral science team, the Behavioral Insights Team, to run 73 randomized trials. Most tested a new or revised letter, email, or other message to residents. Economists Stefano DellaVigna, Woojin Kim, and Elizabeth Linos then checked which of the tested ideas each city kept using. The cities adopted 27% of them. Ideas with statistically significant positive results were adopted 30% of the time, barely more than ideas that showed no effect or a negative one (25%). The strongest predictor of adoption was whether the test changed something the city already sent. Changes to an existing letter or form were adopted 67% of the time, and brand-new messages 12%. (Journal of Political Economy, 2024.)

So the strength of the evidence mattered less than the effort it took to act on it. Rewording a letter the city already mails takes an afternoon. A new message needs someone to own it, fund it, and keep it running. The authors note that no comparable data exists for companies, but most leaders will recognize the pattern: the pilot wins, everyone applauds, and the old process stays exactly where it was. The practical reading is to plan the adoption path along with the test: name who will act on the result before the test starts, and for anything new, line up the budget too.

McDonald’s shows what a deliberate “after” looks like. It opened CosMc’s, a small chain of drink-focused stores, partly to learn whether customized drinks would slow down a regular kitchen. Its chief executive later said 80% of CosMc’s orders needed no customization, a sign that the drinks could work in a regular McDonald’s kitchen. The company closed the stores in 2025, began testing CosMc’s-inspired drinks in more than 500 McDonald’s restaurants that September, and said in April 2026 it would launch six new drinks at participating U.S. restaurants on May 6. The test had a question, the answer came back, and the company acted on it within a year.

The simplest version of this discipline is a debrief: a short meeting after a project, test, or exercise where the team reviews what happened and decides what to change. A meta-analysis of 46 samples found that debriefs made teams and individuals about 25% more effective than going without one, and the average debrief lasted about 18 minutes. (Tannenbaum and Cerasoli, Human Factors, 2013.) That is less time than most status meetings, spent on the one question a status meeting rarely asks: what will we do differently?

Groups Have a Flow State Too

Most people know the feeling of flow: the state of full absorption that psychologist Mihaly Csikszentmihalyi described, when time seems to bend and the work carries itself. Anyone who has been in a meeting where ideas built on one another faster than anyone could write them down has felt the group version. Josh’s forthcoming book calls it Co-Flow, a collective flow state in which a group thinks as one without losing the individual voices inside it. The book argues that Co-Flow works more like a dial than a switch, and that the conditions around a group decide how far it turns.

The science of group flow is young. A small 2021 brain-imaging study found that team flow looks different in the brain from solo flow, with more synchrony between partners’ brains. (Shehata and colleagues, eNeuro, 2021.) With only 15 participants, it is an early signal. The research on how teams create is much deeper, and it matters to leaders for a practical reason: much of the creative work in organizations happens in groups, and the conditions of a group’s working session are set by whoever runs it.

Co-Flow is the state to aim for. The book names five forces that cultivate the conditions for it: Shared Intent, Adaptive Agility, Inventive Reach, Risk Permission, and Receptive Presence. This section previews three conditions the forces build on, each with the research behind it: a shared aim, the rhythm of working alone and together, and the way a group receives ideas.

A Shared Aim

In a meta-analysis of 104 studies of what makes teams innovative, a shared vision had the strongest link to innovation of any team-level factor the researchers examined. (Hülsheger, Anderson and Salgado, Journal of Applied Psychology, 2009.)

The crew of Netflix’s 2025 series Adolescence filmed each hour-long episode as a single continuous take, after a week of actor rehearsals and another week of full-crew rehearsals. With no edits to hide behind, every camera operator, sound technician, and actor was working toward the same unbroken hour, take after take. The series won eight Emmy Awards, including Outstanding Limited or Anthology Series.

Most teams will never shoot a one-take episode, but every team can start a session the same way that crew did, with everyone clear on what they are making. Of all the conditions in this section, it is the cheapest to supply: one sentence, said before anyone pitches an idea.

Time Alone and Time Together

The best-performing groups do not stay in constant contact. In a laboratory experiment, three-person teams worked on a hard route-planning puzzle. Teams that checked in only every few rounds got the best of both: they found the best possible answer about as often as teams that worked in complete isolation, and their average answers were as strong as those of teams in constant contact. Teams in constant contact found the best answer less often. (Bernstein, Shore and Lazer, PNAS, 2018.) Constant contact pulled everyone toward the same answer early. The intermittent teams explored on their own, adjusted to what their teammates had found, and went back to exploring.

The same rhythm works in a meeting. In a 2025 study of 157 students in small online groups, groups that brainstormed in writing (each person writing ideas on their own before reading anyone else’s) produced more ideas, more original ideas, and higher-quality ideas than groups that brainstormed out loud over video. (Baruah, Jimenez and Paulus, Journal of Creative Behavior, 2025.) Many brainstorms run like a radio call-in show, where whoever gets through first holds the airtime.

For a leader running a session, the lesson is to schedule the quiet. Let people think alone before the group talks, combine what they bring, and then separate again to improve the best ideas.

How a Group Receives Ideas

A group can only build on ideas that people are willing to say out loud. A meta-analysis of 136 samples covering more than 22,000 people found that psychological safety (the sense that you will not be punished for speaking up) was most strongly linked to asking questions, admitting mistakes, and sharing information. Its link to individual creativity was relatively weak, and its link to team creativity was moderate. (Frazier and colleagues, Personnel Psychology, 2017.) Psychological safety has become a popular answer to many team problems. This meta-analysis suggests its main effect is narrower: it makes people more willing to speak up and share what they know. Turning those contributions into creative work depends on what the group does with them.

How a leader answers matters as much as whether they say yes. In two experiments, people whose idea was approved by an angry-looking manager reported feeling less safe and less willing to speak up again than people approved by a manager who looked happy or neutral. In the second study, an angry approval left people feeling as unsafe as an outright rejection. (Guzman, Tempesta Fernández and Hamstra, Scientific Reports, 2025.) People read the face before they hear the answer. This means a leader’s bad mood can do real damage. A manager who approves an idea while frustrated about something unrelated may believe they were supportive. The team may conclude it is safer not to suggest anything again. When your mood has nothing to do with the idea, say so. It costs nothing.

Sometimes safety looks like a curtain. At a large mobile phone factory in southern China, Harvard researcher Ethan Bernstein found that workers hid better ways of working whenever managers were watching. “We have all of these ideas . . . but how do we feel safe to try them?” one worker asked. In a field experiment, four of the factory’s 32 production teams worked behind a curtain like the ones around hospital beds, out of managers’ constant sight, while their output was still measured. Over the next five months those lines were 10 to 15 percent more productive than the others. (Bernstein, Administrative Science Quarterly, 2012.) Close watching can hide the very improvements a manager hopes to find.

Many idea systems answer with the silence of a voicemail box nobody checks. In research on thousands of organizations that crowdsource ideas from the public, 88% gave the people who submitted ideas no feedback of any kind, positive or negative, and newcomers who got an explicit rejection were more willing to submit another idea than those who heard nothing. (Dahlander and Piezunka, 2020; Piezunka and Dahlander, Academy of Management Journal, 2019.) The research covers outside contributors, but there is little reason to expect employees to react differently. A clear no keeps people offering ideas. Silence teaches them to stop. A short no with a reason takes less time to write than it took to read this paragraph.

Once ideas are in the room, the best teams build on them. Across 150 studies of 9,702 teams, the quality of a team’s communication predicted its performance far better than how often members communicated. The strongest single factor was building on what teammates shared. (Marlow and colleagues, Organizational Behavior and Human Decision Processes, 2018.) Building on an idea also tells the person who offered it that the group heard them. So the place to start improving collaboration is the meetings already on the calendar, and how people respond to each other’s ideas in them. None of it requires a budget.

These three conditions are a starting point. In Co-Flow’s terms, a shared aim is the root of Shared Intent, the rhythm of working alone and together is part of Adaptive Agility, and the way a group receives ideas runs through Risk Permission and Receptive Presence. The fifth force, Inventive Reach, pushes a group past its obvious ideas. Working together, the five forces are how a team turns the dial from a good meeting toward the state where the group thinks as one. The book shows how teams get there.

What We Mean by Play

Play can sound like a perk, the kind of thing that shows up as a foosball table in a recruiting video. The research treats it as a working condition, one that affects how many ideas people produce and how willing they are to try them. It also shows where the popular claims run ahead of the evidence, which matters to any leader deciding whether to invest.

The word covers several different things in the research: positive mood, improvisation, protected time to explore, and playful approaches to everyday tasks, such as turning a routine job into a game. They are not interchangeable, and the evidence differs for each, so this section keeps them apart.

What People Want

Start with demand. Asked what they hope the future of work looks like, 70% of full-time office workers in Gensler’s 2026 Global Workplace Survey said they hope their work becomes somewhat or much more creative. Gensler surveyed 16,459 full-time office workers in 16 countries. In a 2026 SurveyMonkey survey of U.S. workers, 95% described themselves as at least somewhat curious, but only 30% said their workplace strongly rewards curiosity, and more than half (53%) said more unstructured time would help them be more curious at work. Most of the curiosity a leader could want is already on the payroll.

Protected Time

The obvious fix is to give people time, and the obvious fix is harder than it looks. The software company Atlassian once gave its developers 20% of their time for projects of their own choosing. After a year it counted the results: 34 people had put in 248 days of effort across 48 projects, which the company calculated as “1.1% Time.” It allowed that the true figure might be two or three times higher, and still called that very low. The most common complaint was finding the time.

A few Atlassian teams then tried a different arrangement: instead of scattered personal time, they set aside a full week of innovation time for the whole team at the start of each release. In 2012 the company reported that those teams had shipped 12 features and improvements from their first two Innovation Weeks. We admire Atlassian for publishing such an unflattering number about its own program. The 20% existed in the policy, but the calendar never made space for it. Innovation Weeks seemed to work better because they changed the calendar itself.

What the Research Shows

The research on play at work is smaller than the topic’s popularity suggests. Most studies find gains, some small and some sizable, and the newest ones show that those gains depend on how play is introduced. On mood, a meta-analysis of 102 effect sizes found that positive moods produce more creativity than neutral moods, though the effect was small. (Baas, De Dreu and Nijstad, Psychological Bulletin, 2008.) On improvisation, an MIT doctoral study found that improvisational comedians came up with 20% more product ideas, and 25% more creative ones, than professional product designers. (Kudrowitz, 2010.) On playing together, a randomized experiment at Brigham Young University found that newly formed teams that played video games together for 45 minutes showed a 20% productivity improvement on later tasks over teams that did a traditional team-building exercise. (Keith and colleagues, 2018.)

The newest studies look at playful approaches to everyday work, and they point to the conditions that decide whether those approaches take hold. In a 2025 study of 256 employees, people who felt their organization supported playful approaches to their work used them more often. (Ștefan, Pap and Vîrgă, Journal of Occupational and Organizational Psychology, 2025.) In two randomized studies published in September 2026, simply being assigned exercises for making work more playful did not change how people worked. The change showed up only among people who chose to do the exercises, and in one study only for the competitive kind of play (turning tasks into challenges). (Scharp, Van Woerkom and Junker, Career Development International, 2026.) A play program that nobody uses does about as much good as a ping-pong table locked in a storage closet.

Two older studies fill in the rest. In a 30-day study of 77 naval cadets, colleagues rated the cadets’ work higher on days they made it more playful, and the effect was strongest when work pressure was low. (Bakker and colleagues, Journal of Vocational Behavior, 2020.) And in a field experiment, games rolled out to every salesperson raised positive feelings at work only among those who had bought into the game. Among those who had not, positive feelings fell, and performance did not improve in either group. (Mollick and Rothbard, Wharton working paper.)

Put together, the research gives leaders a short list. Playful approaches help most when people choose them, when the organization visibly backs them, and when the pressure is low enough to leave room for them. Mandated fun fails on the first count. Play seems to work like jazz improvisation. A bandleader can choose the song and set the tempo, but nobody can order a musician to play with real feeling.

Play in Practice

The company stories that follow are illustrations. They show what those conditions can look like inside real organizations, though no single story can show that play caused the result.

When Nintendo began work on Super Mario Bros. Wonder, the team set aside its usual fixed schedule at the start. Everyone on the team, including programmers and sound designers, wrote gameplay ideas on sticky notes, more than 2,000 of them, and the team built quick prototypes of promising ideas on the spot. By Nintendo’s estimate, the game sold 4.3 million copies in its first two weeks, which Nintendo says made it the fastest-selling Super Mario game since it began collecting this kind of sales data in the Wii and DS era. Look at the conditions Nintendo set: ideas could come from anyone on the team, trying one cost very little, and most were expected to be thrown away.

The New York Times Games team holds an annual internal game jam where anyone can pitch a game they would like to see, and Connections, the daily word puzzle, grew out of the 2021 jam. At the small studio Aggro Crab, burnt out after a project that had taken more than three years, three developers spent a month in Seoul in early 2025 with four from another studio, Landfall, and built the core of a cooperative climbing game called Peak in about four weeks. It sold 2 million copies in nine days, and Aggro Crab now gives its staff every other Friday to work on whatever they want. In both cases the play had a container: a set time, a small group, and permission to make something nobody had asked for.

Play also works at the scale of one person on an early shift. In October 2022, Paris baker Stéphane Louvard started stuffing chocolate-chip cookie dough into croissants and baking them again, mostly for fun during his long shifts. The crookie stayed a small item for regulars for more than a year. After TikTok videos took off in early 2024, he sold 2,300 in a single day. Plenty of good ideas start as something a person does for fun on a slow shift. A leader’s job is to notice them and give them a place on the menu.

Prototyping in Cardboard

Play’s low stakes are useful for serious work too. When Children’s Healthcare of Atlanta was building a new hospital, its team judged virtual reality too limited for teams from different specialties to work through together. It rented a 100,000-square-foot warehouse and built cardboard versions of 15 clinical areas instead, essentially a life-size dollhouse of the new hospital. The project ran for a year and included 40 days of hands-on testing, in which more than 800 clinicians acted out 146 care scenarios in the cardboard rooms. It found 722 design issues before construction, and the hospital system says the exercise helped it avoid more than $90 million in late change orders on a $1.5 billion project. It is not play in the usual sense, but it shares play’s key feature: mistakes cost almost nothing. The cheapest moment to change a plan is before anything is built, and cardboard keeps that moment open longer.

Where to Start

We would not start a play program by buying anything. The research points to a few low-cost moves, and each one can run as a small experiment with a single team.

• Find the play that already exists. Ask your team which parts of their work they have already turned into a game, a contest, or a side project, and protect those before adding anything new.
• Put it on the shared calendar. If you offer unstructured time, give it to the whole team at once, as Atlassian’s Innovation Weeks did, so nobody has to defend it alone.
• Make it optional. Invite people in and let them decline. The salespeople who were drafted into games ended up feeling worse.
• Pick a quieter stretch. Try it in the week after a deadline. The cadets’ playfulness paid off most when pressure was low.
• Hold a one-day jam. Borrow the Times Games format: anyone can pitch, small groups build rough versions, and the best ideas get a second look.

Before you start, pick one thing to count, such as the number of new ideas the team brings forward in a month, and count it again afterward. Then you will know whether play is working in your organization, which matters more than whether it worked at Nintendo.

Sonos’s App Redesign

Sonos is this issue’s Hard Lesson because its trouble came from the conditions around a single launch.

Sonos built its business on making home audio simple: buy a speaker, open the app, play music in any room. The app is how customers control everything they own. A Sonos speaker that cannot connect to it can start to feel like a very expensive bookend. A few years before 2024, the company decided to rebuild it. “A few years ago, we decided to embark on a complete ground-up rewrite of our app,” CEO Patrick Spence later told investors.

The company had the resources to do it well. It reported $1.66 billion in revenue in fiscal 2023, the year before the new app, and ended June 2024 with $277 million in net cash.

The new app reached customers on May 7, 2024. Features they relied on were missing, including sleep timers and the ability to edit playlists and the queue of upcoming songs. As the software reached more and more users, Spence told investors three months later, “it became evident that there were stubborn bugs we had not discovered in our testing.” He had already apologized on the company’s blog (“I want to begin by personally apologizing for disappointing you”), and on the call he tied the failure to his own priority: “Since I took over as CEO, one of my particular points of emphasis has been the imperative for Sonos to move faster. ... With the app, however, my push for speed backfired.”

The cost kept growing. Sonos delayed two major new product releases planned for its fiscal fourth quarter and said it expected to spend $20 million to $30 million in the short term to fix the app. It announced a reduction in force of about 6% of employees in August 2024, and a reorganization with a reduction of about 12% in February 2025. Revenue fell 8.3% in fiscal 2024 and 4.9% in fiscal 2025. Sonos attributed the fiscal 2025 decline to challenges from the app rollout and to softer demand. In January 2025, Spence stepped down, and the company eliminated the chief product officer role.

The Repair, and the Lesson

The Repair. On October 1, 2024, Sonos published seven commitments that came out of an internal review of the launch. Read together, they describe the conditions the release had lacked. Beta testing would “include more types of customers and more diverse setups for a longer testing period.” Releases would change shape: “In contrast to the all-at-once automated app release we issued in May, any major change to the Sonos app will be released gradually, allowing customers to adjust and provide feedback before it becomes the default.” A new Quality Ombudsperson would “ensure our employees have a clear path to escalate any concerns in terms of quality and customer experience.” The company also created a customer advisory board and made executive bonuses for fiscal 2025 conditional on improving the app and rebuilding customer trust. The board later judged that condition met.

Tom Conrad, a board member who became interim CEO in January 2025, went after structure next. “We’ve become mired in too many layers that have made collaboration and decision-making harder than it needs to be,” he wrote to employees, announcing “flatter, smaller, and more focused teams.” The board named him CEO in July 2025. When Sonos began planning its next round of app changes, Conrad said he had described them on Reddit and asked customers for feedback, and that the changes would go to the company’s beta channel first, where users could opt in. The redesigned navigation that shipped in September 2026 was, the company says, “shaped and tested with the Sonos community.” Revenue in the quarter that ended June 27, 2026, was 9% higher than a year earlier.

The Lesson. Sonos had the engineers, the money, and the hardware. What failed was the setup around the launch: testing that missed bugs customers then found at home, a release that reached everyone at once, and, by its own later account, a structure with too many layers between a problem and a decision. In retrospect, Sonos sped up the release and left out the loop that would have told it the software was not ready. “We just changed too much too fast,” Conrad said in 2026, “and made a bunch of tactical errors along the way of rolling it out to our customers, and then paid a very dear price. As did our customers.”

For leaders under pressure to move faster, and most are, Sonos is a useful warning. A fast release and fast feedback are different things. The fix Sonos chose was to test changes with real customers before they reached everyone.

Five Experiments to Try This Quarter

Each one changes a single condition, and each comes from the research in this issue.

Each of these changes one condition and can be measured before the quarter ends. Pick one, run it with a single team, and compare the results with how that team worked before.

On the Shelf This Quarter

• The Progress Principle
  Teresa Amabile and Steven Kramer · 2011
  The diary study behind Section 03, and the clearest account of how everyday conditions shape creative work.
• The Friction Project
  Robert Sutton and Huggy Rao · 2024
  Two Stanford professors on how leaders can make the right things easier and the wrong things harder, starting with the approvals, meetings, and other drag that slow good work.
• Tiny Experiments
  Anne-Laure Le Cunff · 2025
  A neuroscientist’s case for treating work as a series of small, curious tests.
• Big Little Breakthroughs
  Josh Linkner · 2021
  Our co-founder’s case for everyday micro-innovations over swing-for-the-fences bets.
• Crack the Code
  Kaiser Yang · 2021
  Our co-founder’s keys to unlocking innovation in everyday work.
• The Find a Way Field Guide
  Josh Linkner and Kaiser Yang · Amplify Publishing, 2026
  Our new workbook of hands-on tools for unlocking creative capacity. The ebook is free now, and a printed edition is coming soon.

Other Useful Pieces of Work

• “The Power of Small Wins”
  Teresa Amabile and Steven Kramer · Harvard Business Review, 2011
  The short version of the diary research in Section 03.
• “Getting Rid of Stupid Stuff”
  Melinda Ashton · New England Journal of Medicine, 2018
  How one health system asked its staff what to remove, and what happened next.
• “Bottlenecks for Evidence Adoption”
  Stefano DellaVigna, Woojin Kim, and Elizabeth Linos · Journal of Political Economy, 2024
  The study behind the 27% figure in Section 04.
• “Ask the Developer Vol. 11, Super Mario Bros. Wonder”
  Nintendo, 2023
  The team’s own account of the sticky-note sessions described in Section 06.

Make Innovation an Everyday Business

For more than a decade, Platypus Labs has helped the world’s most ambitious organizations build the creative muscle that this issue has been arguing for. We are an innovation consulting, training, and research firm. Our work sits at the point where strategy, culture, and craft meet, and the deliverable, every time, is an organization with more creative horsepower than it had when we walked in.