You have been sitting on the country expansion for two months. The team reads it as hesitation. You would say you are waiting for the regulator’s guidance in September.
One of those readings is correct and the test is not how you feel about it. A wait is worth something only if something arrives, and you should be able to name the date.
Why these three models
The decision is whether to commit now or hold the choice open, and how to tell the difference between the two versions of waiting. The features that fire are a commitment that is costly to reverse, uncertainty that may or may not resolve on its own, and a window that may not stay open while you think.
Three lenses. Options produce a random answer about what flexibility is worth. The value of information produces an equilibrium answer about whether waiting actually buys anything. Regime states produce a cycle answer about whether the opportunity survives the wait. Most founders run the first intuitively, skip the second entirely, and discover the third afterwards.
1. Flexibility has a value, and it comes from irreversibility
An option is the right to do something without the obligation to do it. Holding one is worth something whenever the future is uncertain and the commitment is hard to undo, and the value rises with both.
That second condition is the one founders skip. If a decision is cheap to reverse, deferring it buys you almost nothing, and the correct move is to decide immediately and change your mind later if needed. The value of waiting comes entirely from the cost of being wrong in a way you cannot undo. A three-year office lease and a month-to-month desk arrangement present the same decision and only one of them is worth deferring.1
So the first question is not how uncertain am I. It is what does reversal cost. Rank your open decisions by reversal cost and you will usually find that the ones you are agonising over are cheap to undo, while the ones you signed quickly are not. That inversion is common and it has a simple cause: hard-to-reverse commitments often arrive with time pressure attached, which is not an accident on the other side of the table.
The corresponding move is to buy reversibility rather than to buy time. A contractor before an employee. A pilot before a rollout. A one-year lease at a higher rate. Each of these costs money and each converts an irreversible commitment into a reversible one, which is the same thing as manufacturing an option where none existed. The premium is visible on the invoice and the thing it buys is not, which is why finance-minded teams cut exactly these line items first and discover their absence only in the quarter they needed to change direction.
Price the reversal in cash and in weeks rather than in feeling. Take each commitment and write what it would cost to be out of it in ninety days. A distributor agreement with six months of exclusivity costs you the margin on every unit you could have moved through anyone else, plus the relationship you burn getting out. A two-year warehouse lease on the industrial road out of Kampala often carries key money and a landlord who expects the balance of the term whether you occupy it or not, so the reversal cost is the remaining rent, not the notice period. Four people on contract cost you a month of notice. Those three numbers are nowhere near each other, and they should buy three different amounts of patience.
When you cannot afford the premium on reversibility, shorten the commitment rather than soften it. One quarter of the distributor agreement, non-exclusive, at a worse unit rate. Half the warehouse with the option to take the rest. That buys less flexibility than a properly structured option and it costs almost nothing, and founders who say optionality is out of reach have usually only priced the expensive version of it.
2. Does the wait actually resolve anything
Holding an option is only valuable if something happens while you hold it. This is where most deferral fails, and the test is precise enough to apply in a meeting.
Information has value only when it could change what you do. If you would take the same action under every plausible outcome of the thing you are waiting for, then the information is worth nothing regardless of how interesting it is, and the wait is pure cost.2
Run it as two sentences. If the September guidance says yes, I will expand. If it says no, I will expand anyway through the partner structure. Written down like that, the wait has just been revealed as theatre, and the two months were spent buying an answer that changes nothing.
The version that passes looks different. If the guidance permits direct licensing, I hire four people and open in Q1. If it does not, I do not enter this market at all this year. That is a decision with a genuine branch in it, and waiting for the branch to resolve is correct rather than cowardly.
If you cannot write two different actions for two different answers, you are not waiting for information.
There is a further check worth applying, because the two-sentence test can be passed dishonestly. Ask what would have to be true for you to decide before the date. If no answer exists, the decision is genuinely gated. If the honest answer is that you would move now if a competitor did, then the constraint was never the regulator.
The two-sentence test has one failure mode worth naming, because it passes the check while breaking the logic. You write two genuinely different branches and one of them is a branch you would never execute. “If it says no, I do not enter this market at all this year”, written by a founder who has already told the board that this market is next year’s plan, is not a branch. It is a sentence. The check that catches it is whether you could say the losing branch out loud, in those words, to the person who would have to carry it out. If you would soften it in that conversation, rewrite it now, before the wait borrows its credibility.
The same test keeps three things apart that all feel identical from the inside. Waiting for information means an answer is coming that changes what you do. Waiting for permission means the answer is already known and someone else has to sign. Waiting for capacity means you know exactly what to do and cannot staff it. Only the first is deferral. The second is a follow-up problem with a name and a phone number attached, and the third is a hiring plan. Filing either of them under strategy is how a two-week task becomes a two-month silence.
3. The window may not wait for you
The first two models treat the opportunity as sitting still. It usually does not, and the way it moves is not smooth.
Markets and financing conditions tend to persist in a state for a while and then switch, rather than drifting continuously. Which means the risk of waiting is not that the opportunity gets gradually worse. It is that it is roughly the same for months and then materially different, and the transition is not announced in advance.3
That changes what a deferral costs. If conditions decayed smoothly, waiting two months would cost roughly two months of decay and you could price it. Under regime behaviour, the expected cost is small and the actual cost is either nothing or the whole opportunity. Averaging those two produces a number that describes neither outcome.
The practical response is not to hurry. It is to hold the option and watch the transition signals rather than the calendar. Write down the two or three observable events that would tell you the state had changed: the licence granted to a competitor, the funding window closing in your category, the partner you were relying on signing with someone else. Then treat any of them as a trigger that forces the decision immediately, ahead of your original date.
The signals have to be things you would actually see. Where most of what matters moves informally, the observable events are rarely announcements. A competitor advertising for a country manager. A clearing agent mentioning that a rival’s container came through last week. A supplier asking, lightly, whether you would mind them serving someone else in your category. Those all arrive before any press release, and each one is a state change you can act on. Name the person who would hear each signal and ask them for it monthly. A signal nobody is assigned to watch is just a date with extra steps.
Decide in advance what happens when the counterparty moves the other way. Sometimes the guidance lands early, or the partner improves the terms if you commit this week. The instinct is to read a favourable surprise as confirmation and sign. It is a state change like any other, so rerun both branches against the new state first. An offer that arrives ahead of schedule was designed by the person offering it, and the deadline attached to it is part of the design.
Watching signals rather than dates is more than a preference. The state of the thing you care about and the information you have about that state are two different quantities, and the second lags the first by however long it takes news to reach you.4 A calendar date measures your own patience. A named signal measures the world, and it is the only one of the two that can arrive early. Set a date so the wait ends, and set signals so it can end sooner.
The combination is the actual discipline. A deferral with an expiry date and a trigger list is a strategy. A deferral with neither is procrastination wearing the vocabulary of strategy, and the vocabulary is convincing enough that it usually survives several board meetings.
What the three say together
- Price the reversal, not the uncertainty. Cheap to undo means decide now.
- Write the two branches. Different answers must produce different actions or the wait is worthless.
- Set an expiry date on the deferral, chosen in advance and written down.
- List the two or three events that would force the decision early, and treat any one of them as the trigger.
Where they disagree
The options view and the regime view give opposite instructions under high uncertainty.
The options view says uncertainty raises the value of waiting, so the more unclear things are, the more valuable flexibility becomes. The regime view says that uncertainty about which state you are in raises the chance that the state switches while you hold, so high uncertainty is precisely when waiting is most likely to cost you the opportunity outright.
Both follow correctly from their own assumptions, and the assumptions differ on one thing: whether the underlying opportunity is stable while you deliberate. When you are waiting on something internal, like a hiring decision or a technical choice, the options view governs and patience is rewarded. When you are waiting on something several other parties are also watching, the regime view governs, because their responses are what move the state. The test is whether anyone else is making the same decision at the same time. If they are, your deliberation window is shorter than your calendar suggests.
What none of them contain
None of the three prices what deferral does to your team. A held decision is experienced by everyone downstream as a blocked decision, and the cost shows up as people planning around uncertainty, working on things that may be discarded, and reading the delay as an absence of conviction. That cost is real and none of these models sees it. Announcing the expiry date to the team removes most of it, which is a strong argument for setting one even when you do not need it yourself.
None of them handles the option that is destroyed by being observed. Some choices lose value once the other party knows you are considering them: a competing supplier who learns they are being compared, an employee who learns their role is under review. The option framing treats holding as costless and here it is not.
And none of them tells you what to do when the trigger fires and you still do not know. That is the common case, and the honest answer is that the trigger converts the decision from an informed one into a bet, which is worse than deciding well and better than deciding never.
The one action that survives the ignorance: take the decision you have been holding longest and write the two sentences, one action per branch. If both sentences describe the same action, decide today, because the wait has already been shown to be worthless and every further week is pure cost.
Who has to move
This is the founder’s and nobody else can do it, because only the founder can see which decisions are being held and which are simply slow. The instinct under uncertainty is to gather more input, which feels like diligence and usually adds observations that could not change the action. The cheapest first test is the two-sentence branch written for each of your three oldest open decisions. It takes fifteen minutes, and in most cases at least one of the three turns out to have been decided months ago by everything except the announcement.
Sources and notes
- Sanjit S. Dhami, The Foundations of Behavioral Economic Analysis, Oxford University Press, and the real-options literature it references. The claim used in section 1 is structural: the value of deferring a decision rises with the irreversibility of the commitment and with the uncertainty to be resolved, and falls to near zero when reversal is cheap. No option value is calculated here, since the inputs required for a valuation are not available for the decisions this article addresses.
- Jeffrey Carpenter and Andrea Robbett, Game Theory and Behavior, MIT Press. The value of information, and the result that information has positive value only when it can change the chosen action, are developed in the treatment of decision-making under uncertainty. Used in section 2 as the test rather than as a quantity to be computed.
- Thomas R. Bensignor, editor, New Thinking in Technical Analysis, Bloomberg Press, and the regime-switching literature more broadly. The property used in section 3 is that many market series persist in a state and then transition, rather than drifting smoothly. Cited for the qualitative shape only. Note the caution the wider evidence supports: identifying which regime you are in is far easier after the fact than during, and this article makes no claim that transitions can be predicted, only that they should be watched for with named signals.
- Donella H. Meadows, Thinking in Systems: A Primer, Chelsea Green Publishing. The treatment of delays and of the difference between a system’s state and the information available about its state supports the section 3 recommendation to watch observable transition signals rather than the calendar. Used for framing.
A note on a number this article does not give. There is no correct length for a deferral. It depends on the reversal cost, on when the information actually arrives, and on how contested the opportunity is. What transfers is that the length should be chosen in advance and written down, because a deferral without an expiry date does not end, it just stops being discussed.
Joshua Agonya Pi’Rwot, Founder.