How We Value
Money already spent gets no vote.
The comparison problem
Demand has assembled the pool: new candidates framed on the One-Pager, running investments re-entered, uncertain bets flagged. The pool is now visible. It is not yet useful, because the investments in it are not yet comparable. A data platform modernisation, a new market entry bet, a mandatory compliance upgrade, and a cost-reduction programme are all described in the language of their own logic, on their own terms, with their own units of value. The portfolio cannot choose between them until they can be placed on the same scale.
Valuation is the act of translating every investment into comparable terms, using the criteria and weights the reference framework set in Direction. This chapter sets out how to do that — which models are available, when to use each, what forward value means in practice, and how to value the investments whose return genuinely cannot be known in advance. It ends with the numbers the next stage needs: a comparable value for every candidate in the pool, arrived at honestly.
What this chapter deliberately does not produce is a ranked list or a funding decision. Those are the work of the following chapter. Valuation’s output is the inputs to a choice, not the choice itself.
The models and when to use them
There is no single correct way to value an investment. There are several models, each well-suited to a different kind of investment and a different kind of organisation. The mistake is to use one model for everything — to push every investment through a financial appraisal because the finance team is comfortable with spreadsheets, or to use weighted scoring for everything because it is easy to explain. The skill is knowing which model to reach for, and why.
Four families of model cover the portfolio’s needs.
Weighted scoring against the reference framework
The most direct application of Direction’s work. Each criterion in the reference framework is scored, the scores are multiplied by the strategy-set weights, and the totals become the basis of comparison. The appeal is transparency: the scores are visible, the weights are agreed and published, and the outcome of changing a weight is immediately observable.
Weighted scoring works well when the investments being compared are genuinely heterogeneous — different in shape, size, sector, and type — and where the portfolio needs a single comparable number that encodes strategy, not just financial return. It is the default for most portfolios because it can be applied to anything the reference framework can assess.
Its weakest points are that scores for criteria like “strategic fit” or “value” depend on consistent judgement across scorers — a criterion that means different things to different assessors produces noise, not signal — and that it can become a mechanical ritual that produces whatever number was wanted before the scoring started, if the review process around it is weak. The model is only as rigorous as the conversation that produced the weights and validates the scores.
Cost of Delay and WSJF
Cost of Delay is the answer to a question that financial appraisal rarely asks: what is the cost to the organisation of not doing this investment now? Every day an investment is not running is a day its value is not being created. That delay has a cost — in revenue foregone, in competitive position not captured, in user problems not solved — and that cost can be expressed as a rate: the value lost per unit of time the investment is deferred.
Weighted Shortest Job First (WSJF) turns that insight into a ranking mechanism: divide the Cost of Delay by the duration or effort required to deliver the investment, and the result is a measure of the rate at which value is created relative to the time the delivery takes. A small investment with a large Cost of Delay and a short delivery duration should, in most cases, be done before a large investment with a moderate Cost of Delay and a long delivery duration — even if the large investment’s total value is greater — because the small one creates its value faster and frees the resource sooner.
Cost of Delay makes time visible. A portfolio that never asks “what does delay cost us?” is implicitly treating delay as free — which it is not, and which leads to queues of high-value work waiting behind large, slow incumbents.
Cost of Delay and WSJF are most powerful in portfolios with a high rate of demand — technology and product portfolios, especially — where the sequencing of work and the time-to-value of individual items is a significant driver of outcome. They are less useful where most investments are long-running, structurally similar, and where the comparison is across investment categories rather than within a queue of comparable work items.
Value-versus-effort (the simple portfolio map)
A plotting of investments on two axes — value (the return expected) against effort or cost (the resource required) — is the most accessible of the valuation tools and one of the most useful as a portfolio-level view. High value, low effort: do these first. Low value, high effort: these are the candidates for stopping. The middle is where most of the interesting decisions live.
The value-versus-effort map does not pretend to precision. Its power is transparency at the portfolio level — it makes the shape of the whole set visible in a way that numbers in a spreadsheet do not. Used well, it is a conversation starter that reveals outliers, clusters, and imbalances that nobody had previously named. Used badly, it is a scatter chart produced in a workshop and ignored by Tuesday.
It works best as a complement to a more rigorous valuation for individual investments, and as the view the governance board uses to interrogate the whole set rather than the basis on which individual investments are funded or stopped.
Financial appraisal — ROI, NPV, payback
The traditional toolkit of investment finance. Return on investment expresses the expected return as a proportion of the cost. Net present value discounts expected future cash flows to their present value, accounting for the time-value of money and the organisation’s cost of capital. Payback period states how long before the investment recovers its cost.
These models are powerful when the value of an investment is primarily financial and can be estimated with reasonable confidence — when the investment will generate or save measurable money, in a timeframe that can be modelled, against assumptions that can be defended. They are the right tool for investments where finance has the data to run them honestly.
They become a liability when they are required everywhere, including where the value is not primarily financial (a culture change programme, a research experiment, a strategic capability) or where the estimates cannot be honest (a five-year NPV built on three layers of assumption, each plausible and collectively fantasy). A confident number produced by a bad model is worse than an honest acknowledgement of uncertainty, because the confident number goes into a spreadsheet and gets compared on equal terms with numbers that actually mean something.
“Financial appraisal is the right tool for investments whose value is financial and estimable. Required everywhere else, it teaches people to produce the number that justifies the investment they have already decided to make.”
Which model for which setting
The choice of model should match the investment and the organisation’s capability to use it honestly.
| Model | Best for | Weakest when |
|---|---|---|
| Weighted scoring | Heterogeneous portfolios; where strategy alignment matters as much as financial return | Scorers apply criteria inconsistently; weights are gamed |
| Cost of Delay / WSJF | High-flow product and technology portfolios; sequencing within a queue | Investments are long, heterogeneous, or not sequencing-dependent |
| Value-versus-effort | Portfolio-level view; conversation with the board; identifying outliers | Used as the primary valuation for individual funding decisions |
| Financial appraisal | Investments with measurable financial returns and defensible estimates | Value is not financial; estimates are multi-layered assumption |
In practice, most portfolios use weighted scoring as the primary comparison model and supplement it with Cost of Delay for sequencing decisions within a category, financial appraisal where the numbers are genuinely available, and the value-versus-effort map as the governance view of the whole set. The principle is not model purity but model honesty: use the approach that best represents what this investment’s return actually looks like, and resist the institutional pressure to use a single model for everything because it makes the spreadsheet tidy.
Applying the strategy-set weights
Whichever model is primary, the strategy-set weights from the reference framework must be applied. This is the mechanism, established in Direction, by which the organisation’s strategic intent enters the valuation. Without it, valuation is optimising for the wrong thing — typically, the easiest-to-defend financial return — regardless of what the strategy says should matter.
Applying the weights is not complicated, but it requires discipline. The score for each criterion is multiplied by the criterion’s weight before the totals are summed. A criterion weighted at thirty-five per cent has more than twice the influence on the outcome of a criterion weighted at fifteen per cent. Changing the weights changes who wins — which is exactly what was established in Direction, and is the point. The person running the valuation does not adjust the weights to get the desired outcome; that conversation happened, under different conditions, when the weights were set.
The integrity of the comparison depends on the weights being applied consistently across all investments in the pool. An investment that gets its strategic fit scored differently from its competitors because its sponsor is more articulate, or because the scoring session for it was chaired by someone with a different view, is not competing on the same terms. Consistent application is a process discipline, and it is one the intake log and the comparison meeting must enforce.
Forward value: the only number that matters for the decision
Whatever model is used, the number it produces must be a number about the future, not the past.
Forward value is the value still to come, measured against the cost still to spend. It is the only relevant quantity for a funding decision, because a funding decision is about whether to continue committing the organisation’s limited pool to this investment going forward. The money already spent is gone. It cannot be recovered by continuing, and it cannot be avoided by stopping. It gets no vote.
This is one of the most important ideas in the method, and one of the most routinely violated in practice. The violation looks like this: an investment has consumed eight million over three years. It has delivered little of what it promised. The forecast to complete has grown twice. In the valuation review it is nonetheless compared against candidates with much lower total investment, because “we’ve already spent eight million — we can’t walk away now.” The eight million is not evidence of the investment’s future value. It is evidence of its history. Forward value says: set the eight million aside. What is this investment worth, from today, against what it will cost, from today? If that number is worse than the alternatives in the pool, the investment does not earn its place — regardless of what has already been spent on it.
The money already spent is gone and gets no vote. Forward value — value still to come versus cost still to spend — is the only basis for a funding decision. Sunk cost is a fact about the past, not a claim on the future.
Forward value for a running investment is not the original business case. It is an updated estimate of the remaining benefit, based on what has actually been learned about the investment’s performance, against the remaining cost to complete. The Demand stage structured the update; the valuation models translate it into a comparable number. The How We Value stage is where sunk cost thinking is confronted and, with discipline, overcome.
Valuing uncertain bets as options
The investments flagged at the demand stage as uncertain bets cannot be honestly valued by the standard models, because their return depends on information that does not yet exist. Forcing them through a financial appraisal or a weighted score produces a number, but not a meaningful one: it encodes assumptions about a future that the bet itself is designed to test.
The right conceptual frame is option value. An uncertain investment is not a commitment to a known return; it is the purchase of the right to learn something, and then to make a better-informed decision on the basis of what was learned. The question is not “what is this investment worth if it works?” — that question cannot be answered honestly. The question is: “what is the next increment of learning worth, and what does it cost to acquire?”
In practice this means valuing uncertain bets in two parts. First, the small amount of money and time required to do the next piece of learning — the prototype, the experiment, the market test, the technical spike. Second, the option that learning creates: if the learning is positive, the right to commit to the next stage; if negative, the ability to stop having spent the minimum. The value of the option is not the full projected return of the investment if it goes well; it is the information value of knowing which it is likely to be.
“Funding an uncertain bet to learn is not a commitment to deliver. It is the purchase of better information. The decision to deliver comes after the learning, not before it.”
This framing resolves a common governance problem. A governance process that treats every investment as a commitment to a deliverable, and therefore demands a full business case, will never fund uncertain bets honestly — because an honest uncertain bet cannot produce a credible full business case. The organisation ends up either declining the bet (and missing what might have come next) or funding it on the basis of fabricated confidence (and being surprised when reality differs from the case). Valuing bets as options creates a third path: fund the learning, genuinely, without demanding the pretence of certainty, and reserve the larger commitment for after the information is in.
How performance data updates the value
Running investments are valued forward, as established above. But the forward value is not simply what the original case said it would be, updated for time. It is updated by what has actually happened: whether the value being created is tracking as expected, whether costs have changed, whether the assumptions the case depended on have held.
Performance data — the health and progress of a running investment — enters the valuation through a specific gate. It does not, by itself, change the funding decision. It updates the forward value and forward cost of the investment. Once updated, that investment re-enters the comparison on the same terms as every other candidate in the pool, with its revised numbers. The comparison, using the same reference framework and weights, then determines whether it still earns its place.
This is the mechanism that was named at a glance in the canon, now made operational. A project that is performing well — on time, on budget, delivering as promised — will have its forward value confirmed and its forward cost unchanged: good performance is reflected in the numbers, and the numbers go back into the comparison. A project that is performing badly will have its forward value downgraded and its forward cost increased: bad performance is also reflected in the numbers, and the numbers go back into the comparison. Neither “performing well” nor “performing badly” by itself decides the funding; the comparison decides it, using the updated numbers as inputs.
How We Value at three settings
| Aspect | Lean | Managed | Enterprise |
|---|---|---|---|
| Primary model | Weighted scoring, lightweight; or owner’s structured judgement | Weighted scoring with defined criteria and a scoring session | Weighted scoring as the standard; supplemented by financial appraisal where data supports it |
| Forward value | A pragmatic estimate: what is realistically left to gain, what does it cost to get there | A documented forward value calculation for each investment | A formally updated estimate each period, reviewed by the finance function |
| Uncertain bets | Owner decides the learning investment and the stop criteria | Flagged at intake, valued as a learning tranche with an explicit decision point | A governed fund-to-learn pathway with tranche approvals and learning milestones |
| Performance update | A conversation about whether the value claim still holds | A structured update on the standard template each period | A formal reforecast reviewed before each funding cycle |
The temptation at the Lean setting is to skip the structured comparison and rely on the owner’s intuition about what is most valuable. Intuition informed by the right criteria and honest forward value is excellent. Intuition uninformed by those things is habit, applied with more confidence than it deserves. Even a simple, quick scoring session against the reference framework does more than it might appear to do: it forces the criteria to be named, the weights to be applied, and the comparison to be explicit — which means it can be challenged and improved, rather than reconstructed after the fact to justify a decision already made.
With each investment in the pool carrying a comparable value — and each uncertain bet identified and sized for what its first learning tranche is worth — the portfolio can now do the hardest thing: choose. That is the subject of the following chapter.