The Innovation
Quotient
Your quarterly briefing on innovation.
Download the ReportA 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 The Compression of Innovation Timelines. Products that once took years to build now ship in months, and new technologies reach the whole world in weeks. That speed has put leaders under enormous pressure, and the data in this issue shows just how much. The pages that follow trace where the speed came from, what it costs the people at the top, and what the fastest-learning organizations do differently. Their answer, as you will see, has less to do with budgets than with trust.
We hope it is useful to you and your team.
— Josh Linkner & Kaiser YangCo-founders, Platypus Labs
About Platypus Labs
Platypus Labs is an innovation consulting, training, and research firm with a fifteen-plus year track record 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.
Platypus Labs
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.
- 03Innovation Timelines Are Getting ShorterThe evidence that everything is speeding up, long before AI.
- 04The Pressure on LeadersWhat the speedup is doing to the people responsible for keeping up.
- 05It’s Not a Money ProblemWhy spending more is the most natural response, and the least effective.
- 06Core, Adjacent, TransformationalThe innovation portfolio: the map, the math, and the collapsing clocks.
- 07How Standout Companies Handle ExperimentsHow they start experiments, and how they decide what to end.
- 08Hard LessonsThe Duolingo memo, and what it teaches.
- 09Innovators We AdmireFour leaders worth watching.
- 10To-Test ListFour experiments to try this quarter.
- 11Resource GuideWhere to go deeper.
- 12Partner with PlatypusHow we can help.
The report’s key findings and their implication for leaders.
In Brief
The ability to learn fast is a competitive advantage, and learning speed is heavily influenced by how a company runs its experiments.
Innovation timelines are collapsing, and the collapse began long before AI. Leaders feel that speed as pressure, and the most common response is deploying capital: new platforms, new tools, sweeping initiatives. The evidence in this issue says deploying more resources is most likely not the optimal response. The organizations pulling ahead have replaced a few large, executive-chosen initiatives with many small experiments run close to the work, and they are as disciplined about ending experiments as they are about starting them.
This report draws on peer-reviewed research, industry surveys covering more than 17,000 executives and leaders, and company results from 2024 through mid-2026. The pages that follow summarize four key findings, each examined in detail in the body of this report.
Innovation is getting faster, and it has been for two centuries.
The pressure on leaders is rising.
More spending does not mean more innovation.
The standouts are great experimenters with “freedom to fail” cultures.
“Your board will ask what you are doing about AI. The better question is how many experiments your people can run, and how fast you can tell the winners from the losers.”
Josh’s Implication for LeadersThat Surprised Us
The most consequential numbers in this issue, sized to travel. Each appears with full context in the sections noted.
By the Numbers
Are Getting Shorter
The evidence that everything is speeding up, and why it started long before AI.
How Fast New Technology Spreads
Every generation feels like the world is moving faster than the one before it. Our grandparents said it about the automobile, our parents said it about the personal computer, and we say it about AI. It may sound like nostalgia, but the numbers back it up.
Consider how fast new technology reaches people.
Read those numbers as a business leader rather than a consumer: two months is how long it now takes for something new to reach an audience the size of your entire market.
The rise of AI gets the credit for today’s breakneck pace, yet innovation timelines were compressing long before AI came onto the scene. Two economists, Diego Comin and Bart Hobijn, studied fifteen major technologies as they spread across 166 countries over the past two hundred years. They found the same pattern every time: each new technology reaches people faster than the one before it. The world has been speeding up since before the light bulb. AI is just the latest and steepest stretch of a very old trend, which means waiting for the pace to return to normal is waiting for something that has never once happened.
The Cost of Experimenting Collapsed
A big reason timelines keep shrinking is that trying an idea keeps getting cheaper. When Amazon’s cloud launched in 2006, economists got a rare chance to measure what happens when the price of testing a business idea falls sharply. A study in the Journal of Financial Economics, titled “Cost of Experimentation and the Evolution of Venture Capital,” found that startups in the affected sectors needed about 20 percent less money to get off the ground, and that investors responded by making far more bets, each one smaller.
Investors watched the price of launching fall. Upfront Ventures put the cost of launching a tech startup at $5 million in 1995 and $5,000 by 2014. By 2026, business school students were launching working companies with AI and no-code tools for a few hundred dollars, sometimes less than the cost of a textbook. These are practitioners’ own estimates rather than audited figures, but they agree on the direction.
The newest experiments are cheaper still, because many of them now run as simulations: ideas tested against AI models before any customer sees them. Querying an AI model at a fixed level of performance cost $20 per million tokens in November 2022. By October 2024 it cost seven cents. (Stanford AI Index, 2025.)
a tech startup
per million tokens
Business Timelines Are Shrinking Across the Board
This is not just a story about famous tech companies. These numbers describe the companies you compete with, and the ones you have not heard of yet:
From building cars to closing sales, almost every clock in business is running faster than it did five years ago, and your competitors are moving at the new speed whether you are or not. An advantage that once lasted a decade now has to be defended every quarter.
No one voted on this or controls it. It is simply happening, in every industry at once. The rest of this issue looks at what that speed is doing to the people who run companies, and what the organizations that handle it best are doing differently.
on Leaders
What the speedup is doing to the people responsible for keeping up.
The Pressure Is Real
Faster innovation sounds like good news until you are the person accountable for a company’s future. If the last section left you a little uneasy, you are reading it correctly, and you are in good company.
Start at the top. PwC asked 4,701 chief executives around the world a blunt question: will your company still be economically viable in ten years if it stays on its current path? Forty-two percent said no. The people with the best information about their own companies, nearly half of them, believe the clock is running out on the business as it stands today.
That fear tracks reality. In an analysis of 35,000 US-listed companies, BCG found that a public company now has about a one-in-three chance of disappearing from the market within five years, whether through failure or takeover. Fifty years ago the chance was about one in twenty.
CEO Turnover and Stress Are Rising
The pressure shows up in who holds the job and how it feels to hold it. Boards named more new CEOs at large companies in 2025 than in any year since 2010, and Spencer Stuart, the firm that tracks these transitions, attributes part of the churn to boards losing patience with leaders they see as slow to transform for an AI-centric future. The message to every sitting executive is hard to miss: move, or be replaced by someone who will.
The Gap Between Priority and Readiness
Meanwhile, the demand to innovate keeps growing. Eighty-three percent of companies now rank innovation among their top three priorities. Three percent say they are ready to deliver on it. (BCG) That gap, between what boards demand and what organizations can actually do, is where the pressure concentrates.
Pressure like that pushes people toward action, any action, because doing something feels safer than waiting. IBM surveyed 2,000 CEOs and found that 64 percent admit they invest in some technologies before they have a clear understanding of the value those technologies bring, driven by the fear of falling behind. By their own account, nearly two out of three of the world’s most senior decision-makers are spending first and understanding later.
That instinct is the single most natural response to everything in this section. The next section is about why it does not work.
“The pressure on leaders is real and the data proves it. But there is a difference between moving fast and being in a hurry.”
Kaiser’s Implication for LeadersMoney Problem
Why spending more is the most natural response to pressure, and the least effective.
The Instinct to Spend
When the pressure rises, spending is the easiest way to respond. A big software contract, an expanded R&D budget. Each feels like motion, and each produces something to show the board.
Whether that motion pays has been tested at scale. Booz & Company, the consultancy now known as Strategy&, ran a study called the Global Innovation 1000: an analysis of the world’s 1,000 biggest corporate R&D spenders across the industrial, technology, and consumer-goods sectors. The researchers looked for a link between how much a company spent on innovation and how it performed, measured by sales growth, profitability, market value, and shareholder return. In the study’s first year they ran 495 separate analyses hunting for that link. They found none.
Only two of the ten companies executives ranked as the world’s most innovative were also among the ten biggest R&D spenders.
AI Spending Shows the Same Pattern
The same pattern is playing out right now, at record scale. McKinsey’s State of AI research finds that 88 percent of companies use AI in at least one business function. Only 39 percent can point to any effect on earnings, and most of those describe the effect as small. Set that beside the finding from the last section, that 64 percent of CEOs admit to buying technology before they understand its value. Companies are spending heavily on AI before they understand it, and most have nothing measurable to show for it yet.
Money Cannot Buy Foresight
Part of the reason spending underperforms is that no budget can see the future. At Google and Bing, where testing is constant and cheap, only 10 to 20 percent of experiments produce positive results (Kohavi and Thomke, Harvard Business Review). The people running those companies are among the best-informed decision-makers in business, and they still guess wrong most of the time. Hollywood offers the same lesson from a different industry: a statistical model that scored screenplays picked a slate of films that outperformed a portfolio built to mirror the industry’s usual green-light choices (Management Science). Experience, conviction, and a big budget did not beat a spreadsheet.
The Cost of Moving Slowly
Meanwhile, the cheapest gains available, faster projects and faster decisions, go mostly unclaimed. A study in the Strategic Management Journal examined 459 large oil-and-gas investment projects and found that 88 percent could have moved faster than they did. Finishing a single month sooner was worth about $6.3 million per project. McKinsey estimates that slow and ineffective decision-making costs a typical Fortune 500 company roughly $250 million a year in wasted time. Companies are paying for speed with one hand and throwing it away with the other.
The line separating companies that out-innovate from companies that fall behind is drawn somewhere other than the budget. The next section goes looking for where it actually runs, inside the companies that handle experiments better than everyone else.
“Innovation is constrained less by resources than by the willingness to learn.”
Josh’s Implication for LeadersTransformational
The innovation portfolio: the map, the math, and what compression is doing to the clocks.
The Three Tiers of Innovation
Strategy researchers commonly sort innovation into three tiers. The framework entered the mainstream through the 1999 book The Alchemy of Growth, is often called the Three Horizons model, and many companies know it as 70-20-10, after the resource split associated with it.
Improving what you already do, for the customers you already have. Most innovation spending lives here.
Carrying what you do into nearby markets, products, or segments.
Creating something new to the world.
The Math
In a study published in Harvard Business Review, Bansi Nagji and Geoff Tuff found that companies allocating roughly 70 percent of innovation resources to core work, 20 percent to adjacent, and 10 percent to transformational were associated with better performance. The returns ran in the opposite direction: transformational work, the smallest slice of spending, produced about 70 percent of the long-term returns.
Core · adjacent · transformational: allocation associated with outperformance vs. long-term returns. (Nagji and Tuff, HBR)
Few companies actually hold that line. Benchmarking surveys have found real-world allocations sitting closer to an even split, and drifting further toward the core whenever the economy tightens (InnoLead and KPMG). Read that against Section 04: under pressure, the money retreats to the safest tier, which is precisely the tier with the smallest long-term payoff.
The Toolbox
The framework’s practical point is that each tier has to be run differently, and companies that ignore this pay for it. Nagji and Tuff’s study is specific. Core and adjacent work belongs inside the business unit, funded by its annual budget and judged by ordinary financial metrics. Transformational work needs the opposite on every count: funding protected from the annual budget fight, ideally by the CEO; people comfortable with ambiguity, designers and scenario planners rather than analysts; and different scorekeeping, because, in the authors’ words, using traditional financial metrics too early “can kill potentially great ideas.” The most successful practitioners judged early transformational bets on whether the company was likely to learn from them, and let economic measures come later.
IBM took the framework seriously enough to build a management system on it, running its emerging businesses under separate rules from its mature ones; the businesses managed that way contributed more than $15 billion in growth within six years (California Management Review). Intuit tuned the ratio to its own situation, running roughly 60-30-10 under then-CEO Brad Smith.
The Timelines
The tiers were never supposed to share a clock, and the timelines were never fixed. Mehrdad Baghai, a co-author of the original framework, has noted that when the model was built, forest-products companies put their third horizon 10 to 20 years out while semiconductor companies put theirs at 18 months. Compression has been dragging every industry toward the semiconductor end ever since.
The practical conclusion: the three tiers still need different money, different people, and different scorecards. What they no longer offer is different amounts of time. Whichever tier your next bet lives in, the disciplines in the next section now apply to it.
Handle Experiments
Two patterns surfaced.
What We Found in Outliers
We went hunting for companies with distinctive and successful approaches to experiments, across every tier of innovation. Two habits showed up again and again: how these companies start experiments, and how they decide which ones to end. Underneath both sat the same organizational choice, repeated in company after company: decisions pushed down to the teams doing the work, instead of waiting on upper management to green-light or kill an idea.
One more reason this discipline is urgent now. The old assumption was that transformational threats announced themselves a decade early, leaving plenty of time to respond. The market has stopped honoring that assumption.
That urgency now applies to every kind of innovation.
How Outliers Start Experiments
They make starting cheap and permission-free. Adobe hands employees a literal red box with a $1,000 prepaid card inside and an instruction to go test an idea, with no approval and no expense report. More than 1,000 boxes have gone out, and the ideas that work earn bigger funding. The model has spread far beyond Adobe. Swisscom, the Swiss telecom, adapted the program and built it into a 4,000-person experimenting community that has produced more than 500 ideas and CHF 4.8 million in realized value, and versions of the kit have been adopted at companies including Caterpillar, Mastercard, and P&G. At Siemens, a program called Quickstarter lets employees crowdfund each other’s ideas with real budget, and management has no veto. The idea is older than any of them: 3M has given employees free rein over 15 percent of their time since 1948, which is where the Post-it note came from. And it works far from software: FAVI, a French metal foundry, scrapped its middle managers and let shop-floor workers run their own projects and order what they needed to try things, and it became the European market leader in its niche while cheaper competitors chased it.
They make failure safe to admit. Enel, the Italian utility with roughly 65,000 employees, runs a program called My Best Failure: employees post the story of something they got wrong, colleagues read and vote, and winners earn a career-development prize, such as time shadowing a senior leader. The logic is backed by Google’s own research: across 180 of its teams, the company found the number one predictor of success was whether people felt safe to take a risk.
How They Decide What to End
They let results decide. Microsoft runs controlled experiments on the changes it makes to its products, and the results are humbling: only about 1 in 3 of those changes improves the outcome it was designed to improve, even though every one was built by a team convinced it would work. Conviction is a poor predictor at every level of a company, including the top. So the strongest companies let the test make the call, rather than the most senior opinion in the room.
Some go further and ask the whole company. When Google let employees bet on business outcomes in an internal prediction market, the crowd’s forecasts beat the company’s own expert forecasters; research on markets like these, at Google, Ford, and a third firm, found they cut forecasting error by as much as a quarter. Smaller companies have proven the same point. Rite-Solutions, a Rhode Island defense-software firm, gave every employee $10,000 in play money to invest in colleagues’ ideas, and more than 15 of the winners became real products that now account for about a fifth of revenue, by the company’s own count (Management Innovation eXchange). Best Buy’s internal market once forecast that a new service would sell far below management’s official projection, and the crowd was right.
They reward teams for ending their own projects. Google X, the company’s moonshot lab, pays a bonus to teams that end their own project, including a team of 30 that had worked on theirs for two years. The lab killed more than 100 ideas in a single year. One that survived became Waymo, the self-driving car company. This is transformational-tier discipline: the bigger the bet, the more valuable the early, honest ending.
They fund like venture capitalists. Flagship Pioneering, the firm that created Moderna, launches about 15 startups a year and deliberately tries to kill each one. Only the few that survive become real companies. In the founder’s words, each one “is meant to be beaten up and killed, and if it fails to die, it becomes a company.”
What These Companies Share
In every company above, the power to try something and the power to judge it sits with the employees closest to the work. People start experiments without permission slips, and teams end their own projects and get thanked for it. What got them there was letting more people take more swings, and making the verdicts fast and honest. The standing sections that follow show what happens when a company gets this wrong, and introduce four leaders getting it right.
“The companies that win give their people two powers: the power to try without asking, and the power to end without apology.”
Kaiser’s Implication for LeadersEvery issue, we pick one recent failure and examine what it teaches. This quarter: a top-down AI declaration, and the fastest backlash of the year.
“I did not expect the amount of blowback.”
— Luis von AhnCEO, Duolingo · to the Financial Times, 2025
The Duolingo Memo
In April 2025, Duolingo’s CEO Luis von Ahn posted a memo declaring that the company was going “AI-first.” The memo said Duolingo would gradually stop using contractors for work that AI could handle. The reasoning was the same pressure described earlier in this issue: AI and its competitors were moving fast, and a company that waited risked being left behind.
Duolingo was better placed than most companies to make that bet. The company built its growth on relentless product testing, and its use of AI in course creation was real and effective. The problem was the shape of the decision. A change affecting the company’s contractors, employees, and customers was announced from the top as a finished conclusion.
The Backlash. The reaction came from a direction the company did not expect: its own users. Duolingo’s social channels filled with variations of one sentiment, that “AI first means people last.” Longtime users posted screenshots of themselves canceling subscriptions and deleting the app, giving up years-long streaks to make the point. The story spread as a symbol of companies discarding people in the rush toward AI, and Duolingo, a brand famous for its warmth, became the example.
The Walk-Back, and the Lesson
The Walk-Back. A few weeks later, von Ahn began walking the message back. He said he did not see AI replacing his employees and that the company was continuing to hire. He kept defending the broader AI-first direction for months, but the original claim about replacing contractor work had done its damage. Talking to the Financial Times, he was direct about the miscalculation: “I did not expect the amount of blowback.”
The Lesson. Set this story against the companies in Section 07 and the difference is who made the call. The standout companies push choices about what to try down to the people doing the work. The Duolingo memo ran the other way: a sweeping call, made at the top, delivered as an announcement. Nothing about it was tested small first, even though testing small is the company’s core competence. An experiment would have surfaced the backlash cheaply, before the whole world was watching.
The pressure von Ahn felt is the same pressure sitting on every leader in Section 04 of this issue, and his instinct, to move visibly and all at once, is the instinct the data warns about. The market corrected the message within weeks. The deeper lesson has little to do with AI. Under pressure, announcing a decision feels faster than testing one. The Duolingo memo shows how expensive that kind of speed can be.
We Admire
Leaders gracefully surfing the compression wave.
Lisa Su
Su spent a decade rebuilding AMD from near-irrelevance into one of the world’s most important chipmakers, and in 2024 she made the era’s defining commitment to speed: a new generation of AI accelerator every single year, replacing the industry’s traditional multi-year cycle.
The cadence has held, with each generation shipping on schedule through 2026. Committing an organization of tens of thousands of people to a permanently faster clock, in public, with no room to slip, is compression as a leadership choice. Her company now competes at the front of the most demanding hardware market on earth.
Jay Chaudhry
Chaudhry built and sold four companies before founding Zscaler in 2008, and his fifth bet was the boldest: delivering security entirely from the cloud at a time when the whole industry sold hardware appliances. He walked away from the prevailing model years before the market forced anyone to, and the head start compounded.
Zscaler crossed $3 billion in annual recurring revenue in 2025, with its founder still running it. The first company he started was funded by his and his wife’s life savings. Zscaler was the fifth time he bet early on a shift the rest of the industry had not accepted yet.
Tim Latimer
Latimer spent his early career as a drilling engineer in oil and gas, where wells got faster and cheaper with every repetition. When he co-founded Fervo to produce geothermal energy, he brought that repetition discipline with him. Fervo treats each well as an experiment that teaches the next one, and the results have been dramatic: drilling time fell from about 70 days to 21 in roughly a year of iteration.
Geothermal has been the perpetual energy source of the future for half a century. Fervo’s bet is that the barrier all along was the learning rate rather than the physics.
Garry Ridge
Our evergreen entry.
The culture that makes fast experimentation possible was built at WD-40 decades before the tools existed, and it produced fantastic results: in Ridge's 25 years as CEO, the company's market value grew from about $250 million to more than $2 billion, and employee engagement stayed above 90 percent in a world where Gallup puts the norm near 20.
His best-known invention was a phrase, the "learning moment," which made it safe to report what went wrong while it could still be fixed.
List
Four small experiments to run this quarter. Open moves, not a checklist.
Four Experiments to Try This Quarter
Each is an open move, not a checklist item: a small, cheap way to test how your team starts experiments, and how it ends them.
The details of each experiment are on the next few pages. Try one. Reach out and let us know how it went. We’d love to hear from you.
Build Your Own Red Box
Make experiments cheap, low-stakes, and permission-free. Hand a team a small pot of money to test an idea with no sign-off required, set aside dabble time, or run a contest.
If you do not hold the purse strings, pitch your manager on carving out a small no-approval fund and offer to be its first test case.
Give Away a Decision
Push one call about what to try, or what to stop, to the people doing the work. Even better, ask your team which decision you should give to them.
Anyone can speak up to greenlight a small idea, or make the case for ending one that has stopped earning its keep.
Celebrate a Failure Out Loud
Show people that a smart swing that misses is safe. Give a small award for the best failure of the quarter, or simply name it in a meeting: they tried this, it did not work, and we are glad they took the shot.
Map Your Real 70-20-10
Take your last twelve months of innovation effort and sort it into the three tiers from Section 06: core, adjacent, transformational. Many teams discover nearly everything they call innovation lives in the core tier.
You do not need to hit any particular ratio. You do need to know what your ratio is.
Where to go deeper on the ideas in this issue.
On the Shelf This Quarter
-
Experimentation Works
The definitive book on building a testing culture, from the researcher behind our Booking.com material.
-
Trustworthy Online Controlled Experiments
The practitioners’ handbook from the people who ran experimentation at Microsoft, Google, and LinkedIn.
-
The Experimentation Machine
Finding product-market fit in the age of AI, from the HBS lecturer and Flybridge co-founder.
-
The Geek Way
MIT’s McAfee on the norms of fast-learning companies: science, ownership, speed, and openness.
-
The Alchemy of Growth
The book that introduced the Three Horizons model behind Section 06.
-
Big Little Breakthroughs
Our co-founder’s case for everyday micro-innovations over swing-for-the-fences bets.
-
Crack the Code
Our co-founder’s keys to unlocking innovation in everyday work.
-
Find A Way Field Guide
Our free workbook of hands-on tools for unlocking creative capacity.
Other Useful Pieces of Work
-
“Managing Your Innovation Portfolio”
The original 70-20-10 study behind Section 06.
-
“McKinsey’s Three Horizons Model Defined Innovation for Years. Here’s Why It No Longer Applies.”
The argument that transformational ideas now move at core-innovation speed.
-
“The Unexpected Benefit of Celebrating Failure”
The head of Google X on paying teams to kill their own projects.
-
“An Exploration of Technology Diffusion”
The two-century study behind Section 03, for readers who want the primary source.
Platypus
Make innovation an everyday business.
Make Innovation an Everyday Business
For fifteen years, 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.
Keynotes and Workshops
Josh or Kaiser on your stage, in your offsite, or at your all-hands. Sessions range from ninety-minute sparks to multi-day workshops.
Assessment and Diagnostics
Custom quantitative and qualitative work to map where your creative capacity actually lives and where it is leaking.
Innovation Amplification
A structured engagement that materially increases the rate and quality of ideas an existing team ships, using the SPARK methodology.
Growth Accelerator
A go-to-market partnership for teams with an innovation agenda and a commercial deadline.
Innovation Consulting
Senior-level strategic work on the problems that keep your leadership team up at night.
Innovator in Residence
Platypus embedded inside your organization for a defined term, working alongside your people on the things that matter most.
If any of this sounds like the kind of partner your team is looking for this year, we’d love to hear from you.
Get in Touch →Photos: Lisa Su by Fuzheado (CC BY 4.0, Wikimedia Commons) · Jay Chaudhry (CC0) · Tim Latimer courtesy Fervo Energy · Garry Ridge courtesy The Learning Moment · Luis von Ahn courtesy Duolingo.