IGNITE EXTENDED / LESSON 23
When is waiting or staging an investment worth more than committing now?
A staged-investment tree and a decision trigger.
A static investment calculation assumes a specified commitment and subsequent cash flows. A real option gives management a future choice, such as waiting, expanding or abandoning. The future action can depend on information that arrives after the first commitment.
The original note uses project examples and payoff diagrams to show the asymmetry. Management can decline an unattractive later investment while retaining the ability to proceed in a favorable state. Value depends on the flexibility actually existing, the information arriving in time and the costs of preserving the choice.
Consider a teaching example with no discounting or delay cost. Investing 80 immediately yields 140 in a high state and 40 in a low state, each with probability one-half. Expected net value is one-half of 60 plus one-half of negative 40, or 10.
A test costing 10 reveals the state before the investment. Invest only in the high state. Expected net value is −10 + 0.5 × 60 + 0.5 × 0 = 20. The increase comes from avoiding the low-state commitment, after paying for information.
A static analysis commits to a specified action now. A staged analysis includes the ability to select a later action after observing information.
Values cash flows for a specified course of action.
Allows a later decision to change after information arrives; the choice can alter future cash flows.
Illustrative application: building the whole facility immediately incurs its cost in every outcome. Testing first incurs a smaller cost now and the facility cost only in branches where the later evidence justifies building. Include failure branches and test expense when comparing the plans.
The option is the ability to respond, not simply the act of postponing a payment.
Illustrative, zero-discount example: a $10 test reveals a high or low state before an optional $80 investment.
A learning experiment supplies information. An option supplies the ability to choose a later action. They often reinforce each other, but information is not valuable for the decision if no action can change after it arrives.
The source note also introduces option-pricing reasoning. A probability used for a risk-neutral valuation is a pricing device under the model’s assumptions and should not automatically be read as management’s forecast probability. A decision tree using real-world expected cash flows needs a consistent treatment of risk.
A milestone should identify the evidence available when the next commitment is due. Define the favorable, unfavorable and ambiguous outcomes and the actions feasible in each. Include the resources needed to maintain the option until that date.
Waiting has a cost when competitors move, capacity disappears or the market window narrows. Staging can also increase total cost if repeated setup is required. Compare the benefit of adapting to information with these costs rather than treating flexibility as automatically valuable.
The staged path is worth $20 versus $10 in this simplified model.
Time passing and information improving are separate mechanisms. A delay can preserve flexibility, but it can also consume cash or allow an opportunity to disappear.
Changes a future decision enough to justify its cost.
Can destroy value if it consumes cash, loses an opportunity or never resolves the important uncertainty.
Illustrative application: a field trial may reveal whether a device meets operating requirements. Waiting three months without collecting relevant evidence may reveal little. The staged decision should include information quality, test cost, delay cost and the possibility that access or terms change.
Specify what becomes known during the delay and which later action can change because of it.
A real option exists when new information can change a later commitment. Design the milestone around a decision that the team can still choose to defer, change or stop.
Identify what could change the investment decision. Example: whether the prototype can meet the buyer’s required performance in the actual operating environment.
Spend a bounded amount on a test that can reveal that information. Include the time cost and specify what the test represents. A laboratory result may leave deployment uncertainty unresolved.
Agree how different findings affect the decision. Passing a threshold may justify the next stage; an ambiguous result may call for redesign only if another test is worth its cost.
Continue, revise, defer or stop while the choice remains available. Include the larger investment only in branches where it is made and include the test cost in the overall comparison.
Learning creates option value only if the future action can respond to it. A milestone with automatic funding regardless of the result does not provide the same flexibility.
Selected reading: PDF pages 1, 2, 3, 4, 5, 6, 7, 8
An investment opportunity can include a valuable choice about later action. A valuation must represent how management can respond to new information, including the ability to avoid an unattractive commitment.
The opening motel example first values a commitment that proceeds under both favorable and unfavorable outcomes.
The alternative lets management observe demand before deciding whether to build. The later investment occurs only in the branch where it is worthwhile.
The option allows participation in an attractive outcome while leaving some unfavorable outcomes unchosen. That asymmetry changes the opportunity’s value.
The note connects financial option concepts to real decisions. Exercise cost, time, uncertainty and the underlying opportunity help describe the choice.
Selected reading: PDF pages 1, 2, 3, 4, 5, 6, 7, 8
Management needs a real ability to defer, expand, stop or change the commitment. Contractual or competitive constraints can remove that freedom.
Include test expense, delay, lost cash flows and the possibility that the opportunity disappears. Waiting does not automatically create value.
Compare complete alternative policies with consistent timing and risk assumptions. Do not add a generic “strategic premium” to cash flows that already represent flexible action.
The Real Options reading examines managerial flexibility that a passive cash-flow plan can miss. A later choice can limit downside or preserve upside, but it has value only if the choice is actionable and the cost of keeping it open is understood.
What makes a pilot a real option rather than merely a smaller expense?
Use the toy packet: invest $80 before uncertainty resolves; future payoff is $140 or $40 with equal probability. A $10 test perfectly reveals the state before the investment. Assume no timing cost, no discounting and no loss of access.
Calculate both paths. Then change one idealized assumption and explain how the conclusion could change.
Relay's NPV of a deep integration with a second EHR ($300k now, $150k/year incremental gross profit for 3 years) was +$15.97k at 20% and -$27.58k at 30%. The $150k is an average of two illustrative, equally likely states: clinics on the second EHR fill gaps like Relay's core base ($240k/year), or the vendor's API limits (batch-only slot updates) cap fill rates ($60k/year). Daniel proposes an $80,000 pilot integration at 3 clinics over 10 weeks, scaling only if fill rates clear a threshold.
Draw the commit-now versus pilot-first decision tree at 20% and 30%. State the fill-rate threshold that triggers the full build, the most Relay should pay for the pilot, and one fact that would make committing now the better choice.
The $80k is spent whether the pilot succeeds or fails. Leaving it out of the low branch inflates Relay's staged value from +$22.78k to +$62.78k.
If the pilot pushes the high-branch cash flows back a year, the staged value falls to +$5.65k, below committing now. Time value and competitive windows belong in the tree.
If hires, partner contracts or funding for the full build are locked before the pilot reads out, or the threshold is set after seeing results, there is no option, only a slower commitment.
A two-branch decision tree for Relay's second-EHR integration at 20% and 30%, with a pre-set fill-rate trigger, the pilot's maximum justified cost, and the sensitivities that would flip the choice.
Next lesson: Founders, equity splits & compensation
Check your understanding
Question 1
Relay's second-EHR integration has two equally likely states with NPVs of +$205.56k and -$173.61k at 20%. An $80k pilot reveals the state before the $300k build, with no delay. What is the expected NPV of piloting first?
Pay $80k in every branch, build only in the high state: -80 + 0.5 x 205.56 = +$22.78k. +$102.78k forgets the pilot cost; -$64.03k charges the pilot on top of committing anyway, which removes the reason to pilot.
Question 2
Same tree, no delay. What is the most Relay should pay for a pilot that perfectly reveals the state before the build decision?
Value of information = value with the choice before paying for it (+$102.78k) minus commit now (+$15.97k) = $86.81k, which equals 0.5 x $173.61k, the expected loss avoided. $173.61k ignores that the low state occurs only half the time.
Question 3
Toy model: invest $80 to receive $140 or $40 with equal odds. A $10 test gives a good or bad signal that is correct 80% of the time. If you invest only after a good signal, what is the expected net value?
Good signal (probability 0.5): 0.8 x 60 + 0.2 x (-40) = +40, invest. Bad signal: 0.2 x 60 + 0.8 x (-40) = -20, skip. Value = -10 + 0.5 x 40 = $10, the same as committing now. Noise erased the perfect test's $20.
Question 4
The second EHR vendor's partner contract would oblige Relay to fund the full $300k build once any pilot starts, whatever the results. What is the pilot worth for the build decision?
Option value comes from the ability to act differently after learning. If the build is contractually compelled, the information cannot change the action. Negotiating the right to decline the build is worth more than improving the pilot.
Question 5
At a 30% required return, committing now is worth -$27.58k and piloting first -$12.07k (no code reuse). What should Relay do?
The relevant comparison includes doing nothing. A staged path that is less negative is still negative. With reusable pilot code the staged path becomes +$17.93k at 30%, so the decision depends on that engineering fact, not on staging alone.
Question 6
Relay's pilot-first plan is worth +$22.78k versus +$15.97k for committing now at 20%. Which development would most likely make committing now the better choice?
A cost of waiting (a lost channel or window) can exceed the value of learning. A more accurate pilot and reusable code both favor staging; a higher required return makes both paths negative rather than favoring commitment.
Workbook
Module output: A linked financial analysis, operating forecast, valuation and staged-investment decision.
Sources
Original pages and full documents are in the Library edition.