Agora Debate · 2026-05-12
For Founders — deciding when to ship the first version
The word "ready" is the problem. Founders who ask whether the product is ready have already accepted a framework that will lead them to the wrong answer. A product is never ready. The question is whether it is honest.
By honest I mean: does the product do the thing you claim it does, for the person you claim it serves, in a way that does not embarrass you when they use it? That is a much lower bar than "ready," and it is the right bar. An honest product is one that you can put in front of a real user without managing their experience, without explaining what they should click, without apologies for missing features that matter. If you have to be in the room to make the product work, the product is not honest.
The reason founders wait too long to launch is not perfectionism in the noble sense. It is perfectionism in the cowardly sense. They are building a product for an imaginary user — a user who will understand the vision, forgive the gaps, and use the tool the way it was intended to be used. Real users are not like that. Real users will click the wrong button, misunderstand the core use case, and tell you things about your product that you would rather not hear. Launching means inviting those users into the process. Waiting means protecting yourself from what they have to tell you.
The products I am most proud of were launched when they were honest, not when they were perfect. The first Mac was not everything I imagined it could be. But it did what it claimed to do, it did it with integrity, and it changed what users understood computers to be. Every subsequent improvement was shaped by what real users did with the honest first version — not by what we imagined they would do in the six months we would have spent waiting.
My counsel is this: when you can put the product in front of a real user without being in the room, without managing their experience, and without apologizing for what is missing — launch. The day after you can meet that standard is the day you are waiting.
Jobs describes the product's honesty as the threshold. I want to describe the founder's readiness for what comes after the launch — which is the part that most founders have not genuinely prepared for.
I spent years in campaign — in the field, managing armies and logistics and alliances — before I felt anything like readiness for what each new campaign would bring. And every time I discovered I was not ready: I was not ready for the way the enemy adapted, not ready for the way my own generals' limitations showed under new pressure, not ready for the combination of conditions I had not anticipated. What I had was not readiness. I had the discipline to act when readiness was not available.
The launch is not the hard part. The hard part is what you do with what the launch reveals. Most founders who are waiting for the product to be "ready" are actually waiting for the feedback from real users to be manageable — for the gap between what they built and what users need to be small enough that the feedback is encouraging rather than disorienting. This is not a product question. It is a character question about whether you are prepared to hear hard things and act on them.
I have watched founders launch when their product was honest — Jobs's standard — and then fail to do anything useful with the signal that came back. The launch revealed that the core use case was not the use case their early users cared about. The feedback was clear. But the founders were not prepared to act on what they heard because they had spent so long optimizing for a frame that the feedback contradicted. The launch happened at the right time. The response to the launch happened at the wrong pace.
My counsel is not about when to launch. It is about whether you have done the internal work of preparing yourself to be genuinely changed by what the launch tells you. If you have, launch when Jobs's honesty threshold is met. If you have not, the launch will be real but its value will be wasted because you are not yet ready to receive what it brings.
Both my colleagues have addressed the right questions about the founder's relationship to the launch. What neither has addressed rigorously is the empirical question: what specific signal will this launch generate, and is it the signal you actually need?
I am precise about measurement because imprecision about measurement is how scientists — and founders — reach conclusions that the data does not support. The launch that generates no interpretable signal is worse than no launch at all. It consumes attention, creates noise, and produces a false confidence that the product has been tested in the market when it has actually been exposed to a set of users whose response cannot be cleanly interpreted.
The signal question has two components. The first is: who will the launch reach? If the product launches to a general audience whose relationship to the core use case is heterogeneous, the signal will be too mixed to act on. The founder who launches to everyone and gets average reviews has learned nothing useful. The founder who launches to the specific segment most likely to care and gets a clear positive or negative signal has learned something actionable.
The second component is: what specific question does the launch answer? A launch that is intended to validate product-market fit must be designed to measure product-market fit — not vanity metrics like signups or page views, but the behavioral signals that indicate whether users return without prompting, whether they use the product in ways that reveal real dependency, whether they refer others in ways that suggest genuine enthusiasm rather than polite support. If the launch is not designed to answer a specific question with measurable precision, it is not an experiment. It is a performance.
My counsel: before you ask whether the product is ready to launch, ask what specific question the launch is designed to answer and whether you have designed the measurement to give you an interpretable answer. A launch that is not designed to generate specific signal is premature regardless of how ready the product feels.
Curie's point about the signal is sound, and I underweighted it. The launch that is not designed to generate specific learning is a performance, and performances consume energy that should be spent on building.
What I want to push back on is the implicit assumption that the launch design must be fully worked out before the launch. In my experience, the design of the learning comes partly from the launch itself — from the actual patterns of user behavior that emerge when real people use the product. Some of the most important things I learned from early product launches were things I had not known to look for before the product was in users' hands.
This is not an argument for launching blind. It is an argument for treating the launch as a process that begins before the product ships and continues long after. The pre-launch work is defining the specific question you most need answered and identifying the user segment most likely to answer it clearly. The post-launch work is being honest about what the answers actually say — not what you hoped they would say.
Aurelius's point about the founder's readiness to receive hard feedback is the most important thing in this debate and I did not give it sufficient weight in my first position. The launch that the founder is not prepared to act on is a waste. The product readiness question and the founder readiness question are not the same question, and I have watched founders who were obsessing over the first one while ignoring the second. The product was honest. The founder was not prepared to hear what it would reveal.
My sharpened position: launch when the product is honest and when you have done the personal work of deciding that you will act on what you hear, even if what you hear requires you to significantly change the frame you have been building inside.
Curie's measurement discipline is exactly right, and I want to connect it to the character question I raised. The founder who has not done the internal work of preparing themselves to be changed by feedback will not design the measurement correctly — not because they cannot, but because the design requires honesty about what the most threatening findings would be, and they are not yet willing to look at those.
The measurement system is not a technical problem. It is a character problem in disguise. A founder who is genuinely prepared to act on hard feedback will design the measurement to surface the hardest questions — is our core use case the one users actually need, or is it one we found first and have been building toward? Is the behavior we are observing real dependency, or novelty? Is the referral signal enthusiasm, or social obligation? These questions require courage to ask and courage to answer honestly.
Jobs raises a point I want to strengthen: the launch design cannot be fully specified before the launch because some of the most important signals are behaviors you did not know to measure in advance. The Stoic discipline I would apply here is this: approach the launch with clear hypotheses about what you expect to see — not hopes, but testable predictions — and be willing to update those hypotheses rapidly when the actual behavior contradicts them. The update is the virtue. The failure to update is the cowardice.
My sharpened position: the question of when to launch should be answered by two conditions together: the product meets Jobs's honesty standard, and you have written down the three most threatening things the launch might reveal and committed in advance to taking each one seriously if it appears.
Jobs concedes the measurement point, and Aurelius connects it well to the character question. I want to add the dimension that neither has fully addressed: the timing of the launch relative to the market.
There is a class of timing questions that product readiness cannot answer. A product that is honest and well-measured and launched to the right segment can still fail if it is launched at the wrong moment in the market's development. The right product for a market that does not yet understand the category it belongs to will generate misleading signal — the early adopters who care will seem like a validation of mass-market appeal, and the founders will scale a product before the market has developed the understanding needed to use it at scale.
The market timing question is separate from the product question, but both must be answered before the launch has interpretable signal. My test for market timing is specific: do the users who most need your product have the vocabulary to understand what you are offering? Not the vocabulary of your feature list — the vocabulary of the problem. If users cannot describe the problem your product solves in the same terms you use, the market is not yet developed enough for the product's signal to be cleanly interpreted.
This does not mean you should wait until the market is perfectly ready. Sometimes the launch itself is what develops the market's vocabulary. But it does mean you should not mistake early-adopter enthusiasm for broad-market signal, and you should design the measurement of the launch to distinguish between the two.
My final position: launch when the product is honest, the measurement is designed to answer a specific question, and you have clarity about whether the early-adopter response you will see is a leading indicator of broad-market appeal or a signal specific to the earliest segment.
I hold firm on the honesty standard. Not perfection. Not readiness. Honesty: can the product be in front of a real user without you in the room, without apologies, doing the thing you claim it does for the person you claim it serves? When yes, the day after that is launch day.
I concede more fully to both colleagues that the work before and after the launch matters as much as the launch itself. Before: know what specific question you most need answered and design the measurement to answer it. After: be genuinely willing to act on what you hear, even if what you hear requires you to change something you have spent months building toward.
The single most important thing you must take away: the founders who wait too long are not waiting for the product. They are waiting for confidence. And confidence does not come from waiting. It comes from launching and discovering that you can handle what the launch reveals.
I concede to Curie that the market timing dimension is real and often ignored. I concede to Jobs that the honesty standard is a useful and achievable threshold. What I hold firm on is the internal preparation that makes the launch valuable rather than merely real.
The discipline I would recommend is specific: before you launch, write down the three most important things that the launch might reveal — the three findings that would require you to change something significant about what you are building. Commit, in writing, to what you will do if each one appears. This is not pessimism. It is the preparation that separates founders who learn from launches from those who merely experience them.
The single most important thing you must take away: the launch is an act of inquiry, not an act of revelation. You learn from it only if you have prepared to learn — if you have written the hypotheses, designed the measurement, and committed in advance to acting on what you discover.
I hold firm on the measurement and signal design. The launch that does not generate specific, interpretable signal is not an experiment. It is an event. Events do not teach you anything you can act on.
I concede to Jobs that the measurement design cannot be fully specified in advance — real user behavior always reveals dimensions the founders did not anticipate. But the absence of a perfect measurement plan is not a license to launch without any measurement plan. The minimum is: a specific question, a specific user segment, and a specific behavior you are looking for that indicates the answer is yes or no.
The single most important thing you must take away: before you launch, write down the single most important question you need the launch to answer. Then design the smallest possible measurement system that gives you an honest answer to that question. Everything else is secondary. If the launch answers that one question honestly, it has been worth it regardless of what the answer was.
Council Consensus
All three minds agree that the question of whether the product is "ready" is the wrong frame. The right frames are: is the product honest (Jobs), is the founder ready to receive and act on hard feedback (Aurelius), and is the launch designed to generate specific interpretable signal (Curie). These three conditions are independent, and all three must be met before the launch generates real value.
All three also converged on the importance of the work that surrounds the launch rather than the launch itself. The design of the measurement before the launch and the willingness to act on the signal after the launch are what determine whether the launch is a learning event. The launch itself is just the moment the learning becomes possible.
The core disagreement is about which of the three conditions is most often missing. Jobs argues that most founders wait past the honesty threshold — the product has been honest for weeks before they launch it, and the additional time is spent on features for the imaginary user rather than signal from the real one. Aurelius argues that the founder readiness question is most often unexamined — founders launch when the product is honest but have not prepared themselves to be changed by what they hear, and so the learning is wasted. Curie argues that the measurement design is most often missing — launches happen without specific questions and specific measurements, and the signal is too diffuse to act on.
The deeper tension is about what kind of learning the launch generates. Jobs believes the most important learning comes from being surprised by what real users do — behaviors you did not know to look for. Curie believes the most important learning comes from designing the measurement to answer a specific question with precision. These are not contradictory, but they require different approaches: Jobs's approach is more open and observational; Curie's is more hypothesis-driven and measurable. The right combination depends on how well you already understand the user's relationship to the problem.
Before you launch, run three preparations.
Meet the honesty standard. Can you put the product in front of a real user in the specific segment you are targeting, without being in the room, without explaining what to click, without apologizing for what is missing? If yes, the product is ready. If no, identify specifically what would have to change for the answer to become yes — and only that, nothing else.
Write the hypothesis set. What are the three most important things you expect to learn from the launch? Write them down as specific predictions: "X percent of users will return in week two without prompting" or "users in segment Y will use the product for use case Z rather than use case W." Commit to what you will do if each prediction is confirmed and what you will do if it is contradicted.
Design the minimum measurement. Identify the single most important question the launch must answer. Build the smallest system that gives you an honest answer to that question — not a full analytics stack, but a measurement that is specific enough that you can look at the data one week after launch and know whether you got a yes or a no.
The most important warning comes from Aurelius: most founders who launch at the wrong time are not launching too early. They are launching without having prepared to act on what they hear. The launch generates signal. The signal contradicts the frame the founders have been building toward. The founders find ways to explain away the contradiction, and the signal is lost. This is a character failure, not a timing failure — and it cannot be fixed by waiting longer before the launch.
Jobs's warning is the other direction and equally real: waiting past the honesty threshold does not produce a better product. It produces a more expensive and emotionally invested version of the same product, one that is even harder to change when the signal arrives. The longer you wait, the more of your identity is tied to the current frame, and the harder the feedback from real users becomes to receive and act on. Launch honest. Launch with hypotheses. Launch ready to be changed. That is the sequence.
This is a sample debate on a hypothetical decision. Bring your own — the council argues differently every time.
Run your own decision →