Making It Stick
Money already spent gets no vote.
The method meets the organisation
Everything that has been described in this book is, in a technical sense, straightforward. The logic of the Investment Loop is not difficult to understand. The stages are not obscure. The instruments are practical. The settings are clearly defined. And yet most organisations that attempt to run a portfolio management discipline — that build the reference framework, stand up the board, run the first comparison, establish the cadence — find, within twelve to eighteen months, that the method has eroded. The cadence slips. The comparison becomes a formality. The benefits tracker is maintained but not reviewed. The board approves what arrives rather than comparing it to what else could arrive. The stop decisions are deferred.
This is not because the people are bad at their jobs. It is because applying the method consistently, in a real organisation with real political pressures, real sponsors with real stakes, and real data that is imperfect and sometimes gamed, is harder than understanding it. Making it stick is a different challenge from designing it, and it requires its own thinking.
Communicating the decisions
The first and most important thing the portfolio can do to make itself durable is communicate its decisions openly: what was funded and why, what was not funded and why, what was stopped and what that means for the people involved.
Opacity protects decisions in the short run. If the criteria are not published, the weights are not shared, and the comparison is not visible, then no individual decision can be individually challenged — because the basis for it is unknown. This feels like a management advantage, and it is not. It is an advantage for protecting bad decisions and indistinguishable from an advantage for protecting good ones. Opacity teaches everyone that the portfolio process is a black box, which teaches them to work around it rather than through it, which is the beginning of the end of any portfolio discipline.
A portfolio that cannot explain its decisions is a portfolio that will not be trusted. A portfolio that is not trusted will be bypassed. A portfolio that is bypassed has ceased to manage the portfolio.
Transparency does not require publishing every number from every scoring session. It requires, at a minimum: which investments were funded and which were not; that the same criteria and weights were applied to both; and what the funded set means for the organisation’s strategic priorities, expressed in terms that those priorities’ owners can understand. The communication should happen before people find out through other channels — because they will — and should be direct enough to address the investments that were declined, not just the ones that were approved.
Carrying people with you
Stopping an investment — or even declining to start one — affects real people. The team that built the business case spent months on it. The sponsor advocated for it and made commitments based on it. The people who were going to work on it had plans. Treating the stop or decline as a purely financial decision, communicated in portfolio language without acknowledgement of the human consequences, produces resistance that is often more durable than the decision that caused it.
Carrying people with you does not mean softening the decision or being unclear about it. It means doing three things consistently. First, explain the decision in terms of the portfolio logic — the forward value, the alternatives, the capacity — not in terms of the investment’s quality or the team’s capability. A declined investment is not necessarily a bad investment or a bad team; it is an investment whose forward value, at this moment, does not exceed the alternatives competing for the same pool. Second, be specific about what the decision means practically: what happens to the team, the work, the commitments that were made on the basis of the expected approval. Third, do this early, directly, and with the same information given to everyone with a stake in the outcome.
“A decision communicated clearly and early is a decision people can plan around. A decision communicated late, or communicated differently to different stakeholders, is a decision that teaches people the process cannot be relied upon.”
The resistance to stopping work is intense and predictable enough that it is worth treating as a design problem, not a management problem. If the process is designed well — pre-agreed criteria, visible comparisons, decisions communicated openly — the resistance has less to attach to. If the process is opaque, inconsistent, or communicated poorly, the resistance has legitimate grievances alongside the illegitimate ones, and it cannot be distinguished from them. Design the process to deprive the resistance of its legitimate arguments.
The usual failure modes
Twelve months in, the failure modes look like this.
The cadence slips. The quarterly review becomes semi-annual, then annual, then irregular. Each slip is justified by a specific, local, reasonable reason: the board has too many things on that cycle; a significant programme is at a critical stage; the organisation is in the middle of a restructuring. Each individual slip is defensible. The cumulative effect is a portfolio that is not actually being managed on a rhythm, and where drift — the very condition the method was designed to prevent — has re-established itself between the increasingly infrequent review points. The discipline at this point is not to review whether slipping is sometimes necessary — it is — but to hold the cadence as a default that requires explicit exception, rather than as an aspiration that yields to any pressure.
The comparison becomes a formality. The demand log is maintained, the One-Pagers are submitted, the scoring is done, the comparison table is produced — and then the board approves everything on the list because declining anything would require a difficult conversation. The comparison exists in form and not in effect. The signs are: every investment submitted scores above the funding line; no investment has ever been declined at the comparison stage; the board’s discussion of the comparison lasts less than the time it took to prepare it. The fix is to require that the comparison produce at least one decision — at minimum, a funding line that is clearly defined and leaves something below it — and to track whether the decisions actually differ from the submissions.
The benefits tracker is maintained but not reviewed. The template is completed for each investment at go-live. The post-delivery review dates are in the calendar. When those dates arrive, the item is quietly removed from the agenda, or pushed to “next meeting,” until it disappears. Eighteen months later, nobody has looked at whether the value arrived. The benefits tracker has become a document-management exercise rather than a learning instrument. The fix is to make the post-delivery review a mandatory agenda item at the board, with an owner accountable for producing the finding.
Status stays green too long. Self-reported status from investment teams systematically lags reality — not usually through dishonesty, but through optimism, through the reluctance to deliver bad news, and through the genuine difficulty of knowing, from inside, whether the problems you are experiencing are temporary or structural. A portfolio that has relied entirely on self-reported status for six months has a portfolio health picture that is approximately as accurate as the investment teams’ collective optimism, which is, in most cases, considerably more accurate than it should be. Steer’s external view — the dependency map, the resource-contention picture, the escalation log — is the corrective, and it must be maintained actively, not as a backup to be consulted when problems become undeniable.
Sponsors find the workaround. When the portfolio process produces decisions that sponsors disagree with, the instinct is to find a route around it: a budget exception, a direct executive approval, a reframing of the investment as something that doesn’t require portfolio governance. These workarounds are individually small and collectively corrosive. Each one signals that the portfolio process is optional for people with enough seniority or leverage, which teaches everyone that seniority and leverage are the real allocation mechanism. The fix is not to punish the sponsor — it is to close the workaround, by ensuring that every route to funding flows through the portfolio process, and by making the single front door a governance commitment, not merely an expectation.
Data traps
The method depends on data: forward value estimates, cost-to-complete figures, benefit claims, performance data, risk assessments. Most of this data is not known precisely; it is estimated, by people who have reasons to estimate it in particular directions. Understanding the ways in which portfolio data is systematically distorted is part of running the method honestly.
Guessed estimates presented as analysis. A forward value estimate that is built on three assumptions, each uncertain, and presented as a number to two decimal places, is a guess with more significant figures than it deserves. The method does not require precise numbers — it requires honest ones. An estimate of “between five and fifteen million, depending on adoption rate” is more useful than a number like “eight-point-seven million” that implies a precision the analysis cannot support. The portfolio should prefer honest uncertainty to spurious precision, and the comparison process should treat estimates of similar quality as comparable and estimates of widely different quality as incomparable.
Vanity metrics used as value proxies. Measures that are easy to collect and look impressive but do not correspond to the value the portfolio cares about: page views, deployment frequency, satisfaction scores that are never connected to retention, cost-per-transaction without reference to whether the transaction should have happened. These metrics are not without value, but they are not the same as the benefit claims the investment was funded on, and substituting them for those claims produces a portfolio that is measuring its own activity rather than its outcomes.
Gamed status. A portfolio that penalises the bearers of bad news will receive very little bad news. A portfolio that makes it professionally costly to report a delay will see no delays reported until they are undeniable. The design of the Steer and Review processes should make honest reporting the path of least resistance: no-surprises culture means early bad news is welcomed, not punished; the external view from Steer provides an independent check on self-reported status; and the portfolio owner explicitly creates space — in meeting culture, in the framing of reviews — for problems to surface before they become crises.
Making the rules credible
A reference framework that is sometimes bypassed is not a framework — it is a decoration. A set of stop criteria that is applied to investments whose sponsors lack power, and waived for investments whose sponsors have it, is not a set of criteria — it is a power map that has been given the name of criteria. Rules that are inconsistently applied lose their authority faster than rules that are visibly wrong, because inconsistency destroys the basis on which anyone can trust the process.
The credibility of the rules is the credibility of the portfolio. Rules applied inconsistently do not merely fail to govern the exceptions; they undermine every decision that was made under them.
Making the rules credible requires three things. First, the rules must be visible: the reference framework, the weights, the stop criteria, and the decision-rights framework must be documented and available to everyone who participates in or is governed by them. Second, they must be applied consistently across the full population of investments — the powerful sponsor’s investment and the less powerful one, the favourite programme and the unpopular one — with no exceptions that are not themselves governed by a rule. Third, the people responsible for applying them must be accountable for doing so: a portfolio board that routinely overrides the comparison without stated reasons has an accountability gap that the process owner must address.
Sustaining the rhythm
The hardest thing to sustain is not the process or the tools. It is the commitment to the cadence — the regular moment of re-decision that the method requires. The cadence is what distinguishes a portfolio that is genuinely continuously managed from one that runs a big review once a year and calls it portfolio management. Sustaining it means treating the review as a standing commitment of the people in the governance structure, not as a discretionary activity that yields to schedule pressure. It means keeping the process light enough that the review is not so burdensome that it becomes a reason to skip it, while keeping it rigorous enough that the review produces real decisions.
Over time, the discipline compounds. A portfolio that has consistently run its cadence for two years has two years of comparable data, two years of learning about what the reference framework actually selects, two years of benefits-after-delivery observations. It can see its own patterns, improve its own criteria, and make steadily better decisions. A portfolio that has run an annual review with significant interruptions has none of this; it is making each decision with the memory of the last one and no more.
The promise of this book is not that the method makes portfolio management easy. It is that it makes it possible — consistently, at any scale, and with the discipline that the organisation’s resources actually deserve. The method works if it is worked. That is the only condition.