The Iterative Adjacent Possible
When we set out to do bold things, we make a critical error of judgment. We ignore a fundamental constraint because it would be a terrifying bummer if it were true. As a result we continually make brittle plans at great expense that are nearly guaranteed to fail. If instead we embrace the fundamental constraint and tweak our mindset, we can achieve results over the long term that are better than our wildest dreams at significantly lower risk. But to do so first we must embrace our own limitations, which feels, paradoxically, like giving up before we ever get going.
This analysis works for any entity that has agency — that is, that can make choices among many alternative actions over time. It might be a single individual making decisions in the context of their career progression. Perhaps it’s a team within a company making decisions about what work to prioritize. It could also be a whole company making decisions about how to navigate its competitive landscape and deliver value for its customers. The same approach works for any of these situations.
First, let’s develop a clarifying mental model for how things work.
You start out with the position you’re currently in. This might represent the amount of revenue you are extracting from a given product line today, the number of skills you’ve acquired, or what your customers think about you — anything that represents your current state before making any decisions. Your position is boxed in by a number of constraints. You now have to decide where you want to go, a cone of possibility opening up in front of you.
Arrayed in front of you are a number of different points you could jump to. Each of these points might represent a different project and its outcome. Maybe one represents building a new feature in your product, and another acquiring a new capability.
The further away each is from your current position, the more time and effort it will take to achieve it. This is because you have to move farther from where you currently are to arrive at it and movement takes applied force and time. The opportunities that are farther from your current position tend to be bigger pay offs.
Then you make a choice of which project to execute. Picking which one to do requires exercising your judgment: your brain. Once you make your choice, you put your head down and start executing. Making progress requires doing work using your strength: your brawn. Once you pick, the other options fade away, because you’ve committed to one and are executing on it to the exclusion of others.
Which project you pick will depend on which one you think will have the largest value. This value is not just in the direct value unlocked by completing it, but also the indirect value of other good spots that you are now closer to and have a better launching pad to now reach.
So far, so good. You make a choice, you execute that choice, you reap the rewards.
But of course nothing in life is certain. The likelihood of success is not constant and many projects will fail. Perhaps you sight off a given feature to bring to market, and it turns out that once you build it users don’t actually want it, or the cost to provide it is higher than the price users are willing to pay. Perhaps it would have worked but by the time you get there the market conditions have changed and it’s no longer viable. There are any number of reasons why a given project might turn out to not work.
There is some expected success rate, and it falls off the further afield you are from your current position. The future is fundamentally uncertain; the farther into the future you must peer, the more likely the fog of uncertainty obscures the details. We can visualize the likelihood of a project succeeding as being a gradient of green.
The region of success is theoretically a smooth gradient from 100% to 0%. But for simplicity we can focus on just the region that is almost guaranteed to work, above some high threshold, say 95%. This is sometimes known as your adjacent possible. It’s the set of next actions that you can take that will almost certainly work. You might say that these projects are within your grasp.
If your adjacent possible extends very far into the distance, then the world is your oyster. Pick a project, put your head down and execute it, and take the spoils. If your adjacent possible is huge, you’re a hero. We might call this approach to innovation the heroic leap.
In the tech industry in particular we believe in the foundational archetype of the heroic entrepreneur boldly making a dent in the universe. Your boss, your CEO, your investors — all of them will push you to make decisions that implicitly operate as though the adjacent possible is huge:
But here’s the bummer, the uncomfortable truth, the curse that dooms us to cycles of despair: our adjacent possibles are significantly smaller than we’d like to believe. The system of constraints operating on us at any point leave very little maneuvering room and box us in significantly. The clouds of uncertainty in the future get thicker at a compounding rate as they extend in front of us, so it’s hard to see beyond our noses, and impossible to see beyond arm’s reach.
Our adjacent possibles, unfortunately, look much more like this:
Our first reaction is that this can’t possibly be true. We celebrate the heroic archetype because there are innumerable examples of it coming true. Surely that’s not an illusion and we should all just try to be superheroes like our legends were?
If you think about it carefully you can see how those legendary heroes’ accomplishments could be real and also outside the adjacent possible. Imagine a given leap beyond the adjacent possible, with a 99.99% chance of failure. One person steps up to the plate to attempt it, a perfect picture of our heroic ideal: 99th percentile grit and intelligence. But it’s fundamentally a very low chance of success, so they almost certainly fail. Their failure is quiet and unremarkable, and they go on to do something else with no one except their closest friends and business partners even aware of their story. The next person, with every bit of grit and intelligence as the first, comes along and tries it… and again fails. If this happens a handful of times, then almost certainly everyone fails. But if it happens 10,000 times, then there’s a good chance that one person, by pure luck of the draw — right place, right time, on a day when all of the stars align — will succeed. That person who succeeds has effectively won the lottery. But luck only exists looking forward in time, so after they succeed we tell stories about their grit and intelligence that caused their success, ignoring the fact that everyone else who tried was also 99th percentile grit and intelligence. Everyone hears the story of this heroic mega-success and it’s seared into our consciousness. It’s not that these heroic leaps are impossible; it’s that they’re exceedingly rare.
Investors have an interest in everyone believing this myth because if people didn’t believe it, no one would rush to try their hand at it. 99.99% of the people who tried their hand ended up as something like suckers, duped into believing impossible odds. Those people who tried and failed aren’t total suckers: they almost certainly learned and grew from the experience, and as a consolation prize most of American society will respect their bravery for having taken the entrepreneurial plunge in the first place. The individuals almost all failed but at the level of the whole industry it still works: there really are spoils if you work insanely hard and are extremely intelligent, and most importantly get extremely lucky. Those very visible lottery winnings get people’s attention and encourage more people to play. The investor math works out so that even a single blockbuster can subsidize the funding for dozens and dozens of failures. And the self-perpetuating myth creates a never-ending supply of fresh talent. The only downside is that the system chews up and spits out the vast majority of participants. The whole system can hum along in overall equilibrium — especially in a zero interest rate environment where huge amounts of capital can be thrown at the wall to see what sticks.
If you’re in a position where you get 10,000 lottery tickets each with a 99.99% chance of failure, you should do it. If you only get a single lottery ticket then it’s probably not the right choice. Investors can purchase multiple tickets at once, but as an individual you only have one career, and typically can only have a single job at a time. A huge swarm of activity happens, most individuals and projects fail, and every so often something big and new appears.
So it is possible that our adjacent possibles are significantly smaller than we’d like to believe. This reality is such a bummer that it still seems impossible to accept. Believing this to be true feels like giving up, throwing in the towel, succumbing to nihilism. If this is true, then it feels like we surrender our agency, the very thing that makes us human, and become mere dust motes blowing along in the wind. It feels downright offensive — or perhaps even dangerous — to even contemplate. If everyone believed it, no one would bother trying anything and nothing would happen in the world!
We go to great lengths to avoid even considering that it might be true. Our desperate attempts to ignore it end up causing more pain than the underlying fact itself would.
We are not immortal superheroes. We are mortal. We are fallible. Every one of us will die. No one of us is perfect.
When you truly accept this fact, you arrive at a feeling of peace, completeness, balance, and ease. It allows you to look at the problems in front of you not from the chaotic dance floor but the balcony. You can still participate in the dance, but the dance no longer feels existential, like failure in it is death. As an individual the existential dread goes from less of a constant shout to more of a minor background buzz.
But then something truly surprising happens. You realize that by accepting this fact, not only do you feel more at peace, but you realize that it’s possible to create more business value, with significantly less risk, than you ever thought possible. Observers won’t be able to figure out what you’re doing; they won’t see any direct, obvious causal relationship to what you’re doing and what the outcomes are. They’ll conclude that you’re just extremely lucky. Others might believe you’re doing magic — either a wizard or some kind of charlatan trickster pulling the wool over people’s eyes. But the techniques are real and powerful.
Here’s what to do.
We were so focused on how constrained our adjacent possible is that we missed the most important fact: within our adjacent possible we have choice. We have the agency to pick what we will do! Our agency is the source of our power. It is what makes us alive. Fred Kofman would call this our response-ability. Far from being a nihilist stance, embracing our true agency (constrained though it may be) is an empowered stance. Our ability to choose, to apply our judgement, gives us leverage, and makes us an active participant.
It’s also not as bad as we thought that our adjacent possible is small. This is not a one-shot game, where you pick a single move and then put your head down and execute until you’re done and see how it goes. It is an iterated game. You make your choice, you execute on it, then you make another choice. Importantly, the choices that you have after one round are different from the choices you had before. You’ve moved forward in the possibility space, bringing a whole new set of possibilities into your adjacent possible. As time passes, the compounding clouds of uncertainty in front of us clear just a little bit, and a new thin slice of adjacent possible in front of us opens up, with a new set of possibilities within our reach.
The future is fundamentally covered in an impenetrable fog. Why make one single bold choice now, when you have very little information about what might work? It would be like taking a flying leap in a bank of fog on a rocky cliff-face, where a wrong step could lead to death. When you make one big bold long-range decision, you’re effectively locking in many smaller decisions before you have enough information to make a good judgment call on how to make them! Instead, slice up those decisions to start and only make the ones you need to make now, deferring the rest. You can kick the can down the road to make the later decisions once you have more clarity on how things play out. We typically think of kicking the can down the road as avoiding the problem, but it’s a great idea if the road is sloped away from you and things will get easier the farther you go. You’ll be able to make those deferred decisions at much higher quality with significantly lower risk. You’ll have the optionality to adapt the plan based on new information you discover or changes in the environment since you started.
How thin should you slice your actions? You can think about it as the size of your OODA loop. The smaller you slice the decisions, the quicker you can make them, because they become easy, very low downside, barely anything at all. Thinner slices also allow you to adapt to changing context more quickly.
In teams, everyone spends so much time debating between big, bold projects based on information that no one knows. It’s a waste of time — time that could go into actually executing. A lot of these decisions fade away into no-brainers if you slice it down small enough. One person thinks we should do Project A? One person thinks we should do Project B? What would we do differently today if we were going for Project A versus Project B? Nothing? Then stop bickering about it, just do the obvious action and wait until you have better information!
If you make each choice randomly, then you will get a random walk through the opportunity space. You’ll be making iterative, fast, low-risk judgements and getting quickly to action, but they won’t cohere into anything. After many iterations you’ll be in a random place that isn’t necessarily better overall than when you started.
Here’s the other trick: you don’t just make each decision randomly. Instead, you sight off of a north-star vision that you keep consistent. As you execute through your day, you’re constantly making zig-or-zag style decisions. If you’re flipping a coin for each one you get a random walk. If instead you have some bias towards one or the other — that is, you’re pulling in some consistent direction — then you’ll naturally drift in that direction. But you won’t just idly drift. The amount of your change in direction will compound, as it builds up over many iterations. If you do it right, you can arc to wildly different outcomes than you thought possible when you started, perhaps even outcomes that seemed outside your cone of possibility!
A north star vision helps make sure that all of the actions you take accrete into something greater than the sum of the parts. It is from this steady accretion of individual no-brainers into a coherent whole that huge amounts of value can be made.
A north star looks superficially like a big, bold, all-in-one-chunk plan, in that it is long-term and bold. But it’s unlike a plan in that it doesn’t say precisely how to get there or pretend to have certainty on each step. It’s fuzzy, lightly held, a navigation beacon more than a map.
Your north star is typically a vision: a simple narrative that can be expressed in a page or two of succinct prose describing an outcome many years out. Everyone who reads it should come away thinking that it is at least plausible (from their perspective it doesn’t require any miracles) and that the outcome would be one worth celebrating if you achieved it.
You should constantly seek disconfirming evidence: new perspectives who can help you make sure the vision really is plausible. As you find new disconfirming evidence, you tweak your north star to make sure it’s still plausible. Sometimes these tweaks lead to the north star moving across the sky significantly. But unlike rigid plans, where such a change would lead to significant thrash in short-term decisions, this movement instead just leads your individual slices of actions to arc smoothly in a different direction over time.
Your brawn is what you use to make something happen. Using your brawn feels good—you have the wind in your hair, you’re making something happen, not just sitting around navel gazing. And everyone watching will be able to see that you are Doing Things. But we tend to over-index on it precisely because it feels so good. At least some of your calories should be spent on using your brain to exert judgment to pick better options among your adjacent possible and not just jiggling randomly. Using your brain takes space and time; if you are sprinting, using your brawn to its limits, you won’t have the ability to use your brain. To use your brain you must take a step back from the action to gain a slightly wider vantage point. When you do so it will look like you are “taking it easy” or taking yourself away from what matters. It will seem almost self-indulgent. If you aren’t confident in your abilities and the value you’re offering you might be cowed into not doing it. But that’s a trap, dooming you to a downward spiral. Your brawn has a limit: you’ll never be able to run faster than a 3:40 mile, no matter how hard you train. Your brain is what steers you and gives your actions leverage. A better judgment on what to do, paired with execution, can lead you to significantly better results.
So you have a vision and you have a sense of what is in your adjacent possible. How do you actually pick your next action?
The most important thing is that the action you take will clear the bar for survival. If you die in the next round, then there will be no future rounds to play. This is a bar to clear, not to maximize.
Then, all of the rest of your energy should go into a thing that will maximize the long-term value accumulated in that round. This is not a greedy calculation about the direct value but rather a calculation of the net increase in long-term value from doing that action. That is, it should be an action unlocking real value, but also that will pull you somewhat closer to your north star. In this way you can find actions that are likely to pay for themselves (where the value extracted is greater than the cost to extract it), but also bring you somewhere more interesting.
If there is a lot of uncertainty — you don’t know which of many plausible long-term directions will turn out to be the right one — then pick the action that is at least neutral in all plausible future directions you want to go, and ideally adds at least some value in many possible futures. That is, a step that you almost certainly will not regret in the future.
A good rule of thumb is for each decision, imagine watching a video of you making this decision in 3, 5, or 10 years. You want to always make sure you aren’t actively embarrassed by it, and are ideally actively proud of that decision when you look back on it. It’s easy to get caught up in the churn and activity of the short-term; this helps make sure you take a long-term perspective.
Don’t overthink it, just do it. By slicing decisions into smaller steps in your adjacent possible, you’ve significantly reduced the downside risk of a bad decision, while still leaving yourself exposed to the upside of long-term iterative outcomes.
This approach works for any entity that has agency. But it’s significantly easier for individuals than it is for teams of people. The reason is because the coordination cost expands at a super-linear rate with the number of people who must coordinate. Coordination cost makes it harder to quickly decide which next action to do, and also makes it hard to find a vision that everyone finds compelling and worth sighting off. Even if you do find a long-term vision to sight off, over time it will tend to ossify because updating it with new information will also incur a coordination cost. All of this means that for teams, picking which of many different options to choose can take enormous amounts of coordination effort.
A typical business analysis might sight off of the long-term Total Addressable Market (TAM) of an idea. But what matters most is not the absolute size of the TAM in the far-future, but the slope of the path to achieve that TAM. If the first five years of investment are expected to have zero returns before back-loaded returns, it’s almost certain you’ll give up before you ever see any value. But if there is a no-brainer, definitely-pays-for-itself first step, there are a plausible series of inductively no-brainer, definitely-pay-for-themselves next steps afterwards that we could choose to take later, and the long-term prize is likely big enough in orders of magnitude, then just do it! You don’t need to do a big TAM analysis. You can do doorbell in the jungle style patterns of execution and get the best of both worlds: capped downside and exposure to upside. It’s an extremely powerful asymmetry.
One tactic for navigating your iterative adjacent possible is to use a scavenger mindset. Instead of creating everything you need from whole cloth, look around to find pre-existing objects that you can duct-tape together into novel combinations. If you create everything yourself, you might build a thing that isn’t viable at all — its own heroic leap. But if you use two known-to-be-viable things, the only part you’re sticking your neck out on is the duct tape, and duct tape is cheap. You’re significantly reducing the cost of trying new ideas, which makes it more likely that you find outcomes that create more value than they cost. It’s not just iterative solutions you might find in this way. In Why Greatness Cannot Be Planned, Kenneth Stanley and Joel Lehman show that most game-changing innovations come from the novel combination of two disparate pre-existing objects. Nintendo uses this technique in an approach they call “lateral thinking with weathered technology” (sometimes mis-translated as “withered”).
Another approach to navigate the iterative adjacent possible is to set expectations very low. If you fail to clear expectations, your product isn’t viable and you might fail to survive. One way to clear expectations is to make darn sure that your quality is extremely high. Another way is to ensure that the expectations are extremely low so that you can almost certainly clear them and be able to continue itearting in future rounds. Instead of having a big, loud launch to everyone at once — where the stakes are high and coming in under expectations could mean the death of the product — roll it out quietly and allow a self-selecting, highly motivated audience to find it and try it out. That self-selecting early adopter audience will be significantly more likely to be resilient to rough edges, giving you the time and space to iteratively improve your product and make it truly great.
The iterative adjacent possible mindset is a powerful technique, but it requires that you have a good grasp of what is actually in your adjacent possible. For most people in most of society, we undercount how many potentially value-creating moves we have within our grasp. In Silicon Valley — especially in contexts adgance to buckets of VC-funded capital — we tend to over-estimate. That’s because in order to make the case for lots of capital, we have to tell an absurdly optimistic story, and then it’s easiest to navigate that story if you come to earnestly believe it yourself, too.
This enlightened mindset can be extremely powerful, but it’s possible for a superficially similar but fundamentally defeatist one to take hold. For example, in a large, mature company with extremely high coordination costs, it can be easy to over-learn that “everything is so hard that attempting to make anything interesting happen is not worth the effort.” One of the reasons that people reject the enlightened version of the iterative adjacent possible is because they superficially pattern match to this dysfunctional do-nothing stance. When adopting this mindset, be on your guard to not accidentally slump into the defeatist stance. You can tell if you’re in the enlightened stance if you actively yearn to make interesting big things happen and are making steady iterative progress you are proud of.
It’s not that searching for heroic leaps is an inherently bad idea at the level of society; every so often in the cacophonous swarm of activity someone does find a winning lottery ticket, and we all benefit. But it is odd that as an industry we use the heroic leap technique as practically the only innovation approach, when there is an alternative that is more aligned with individuals’ interests and produces significant returns at low risk. It’s as though everyone is trying to be the hare and no one is being the tortoise.
So if this iterated adjacent possible technique is so powerful, why do more people not do it? It requires embracing the inherent uncertainty. Uncertainty hurts. It’s an uncomfortable feeling in the pit of our stomachs. Serious People always project certainty, so that the people who follow them don’t get stuck in the abyss of nihilism and self-doubt. Projecting certainty, even if it’s totally made up, is what Serious People do! By embracing the uncertainty you have to sit with an uncomfortable reality of not knowing. Other people might think you look weak and defeatist, that you aren’t a Serious Person. The trick is that we don’t have to have certainty in order to succeed. Another way of looking at it is how can we make sure that great things happen despite the fact we don’t know?
In teams we spend so much time bickering to choose between a set of projects all of which will almost certainly fail as described, based on information that none of us could possibly have at the moment. It’s a waste of time! Most big interesting outcomes in the world are composed of slices of uniformly small, uninteresting decisions. If you try to zoom in on the magical, bold, decision that changed everything you’ll come up empty, thinking the magic must be hiding somewhere else. But the magic emerges from the totality, not any individual decision.
As an industry we’re so obsessed with boldness that we end up doing performative heroics at great cost with poor returns. We miss ways to be better versions of ourselves and create substantially more business value at significantly lower risk. All it takes is the confidence to admit to yourself and to the world that you don’t know the answer but that you can still take actions that matter.
People are used to rigid strength. But rigid strength is brittle. This is fluid strength, able to flow through uncertainty. It’s significantly more resilient, but people who aren’t looking closely will think it’s unserious or weak. Have the fortitude to ignore the siren’s call of faux certainty and be water, my friend.