Make the inference visible before asking the institution to accept it.

In 2023 I reviewed a proposal for four new regional business-service teams. The commercial argument was coherent: owner-managed companies in all four regions appeared relevant to the bank's proposition, and a simultaneous launch would make the expansion easier to present as a national step. The proposal nevertheless treated the availability of local management and servicing support as if it followed automatically from selecting the regions. I wanted those conditions examined before the bank converted the map into four undertakings.

Graham Pellivale, our fictional regional-business director, supported the wider launch. He argued that postponement could allow other providers to establish relationships first and that a staggered approach would make the organisation repeat mobilisation work. I accepted both as genuine costs of restraint. The disagreement was not between ambition and caution in the abstract. It concerned whether an attractive opportunity was sufficient evidence that we could support four commitments at the same time, with the people and service arrangements the proposition required.

I recommended that two teams proceed first and that the other two remain deferred until recruitment and servicing evidence justified reconsideration. The responsible executives made the operating and customer decisions; my function assembled the comparative account and challenged the assumption of simultaneous readiness. We retained C$2.8 million in additional assessment and staged mobilisation expense. I did not describe that expenditure as an efficiency saving or claim that the first two regions would necessarily produce the expected return.

The decision gave up the proposed immediate reach across all four regions. It also required a better account of the two that had not proceeded. A deferred proposal could not simply disappear from attention and return later with the same unsupported assumptions. The record identified the conditions that would need fresh examination, while avoiding a public promise about a future launch date. Existing customer commitments continued on their own terms throughout the planning choice.

I would use this case to explain the distinction between selecting a market and accepting a service undertaking. The bank did not need a strategy function to redraw an attractive map. It needed an account of what each additional commitment demanded and what the proposed sequence left undone. The outcome was a staged decision with its expense and opportunity cost visible, not evidence of a universally correct number of regions to enter.

A 2015 assessment concerned a proposed acquisition of an advisory business with 180 client relationships. The initial strategic narrative emphasised access to a larger business-customer population. That was plausible as a starting question, but the relationship count did not itself establish additional reach. Our comparison identified 112 relationships already represented in the bank's existing customer base. Sixty-eight were not matched in that exercise. The proposal needed to distinguish those groups before using the total to support its strategic rationale.

The sponsoring team argued that overlap could strengthen relationships and create a deeper proposition for existing customers. I agreed that a repeated relationship was not necessarily worthless. My objection was to presenting the same relationship as incremental reach while also claiming a benefit from serving it more fully. Those were different arguments with different evidence. Neither could be settled by treating all 180 names as a single addition to the bank's addressable customer population.

I recommended that the strategic case be reconsidered without the unsupported new-reach claim. The sixty-eight unmatched relationships were not assumed to become bank customers, and no conversion rate was invented to preserve the original attractiveness. Financial, legal, risk and transaction specialists retained their responsibilities. My role was to examine the reasoning about institutional fit and the choices available, not to determine a valuation or substitute a strategy opinion for diligence in those disciplines.

We did not complete the proposed transaction. C$740, 000 in diligence expense remained a cost of reaching that conclusion. Some colleagues felt the expenditure made withdrawal harder to defend, particularly after the proposal had occupied management attention. I argued that the cost already incurred should be disclosed, but that disclosure did not make an unsupported rationale stronger. Continuing merely to avoid explaining the expense would have confused the record of the assessment with the merits of the decision still ahead.

The lesson I retained was about the language of reach. Relationship overlap may matter positively, negatively or differently from the original proposition, but it has to be described on its own terms. I do not present the rejected transaction as a quantified loss avoided because that counterfactual was not established. The case shows a recommendation that changed after the evidence narrowed what we could responsibly claim, with the unused opportunity and the assessment cost left visible.

In 2021 the bank's strategic portfolio contained nine approved initiatives. Each had a rationale and a sponsor, and each could be explained individually. When we examined their combined demands, six accountable business leaders appeared repeatedly across the work. Approval of the nine proposals had not established that the same people could give each the judgement and attention assumed in its timetable. I asked for the portfolio to be considered as a set of competing organisational commitments rather than nine separate descriptions of merit.

Several sponsors resisted a sequencing discussion. They had already obtained approval and believed a later start would weaken confidence in their programme. One argued that the appropriate response was to add programme support rather than revisit the order. Additional support could help organise work, but it would not necessarily provide the specific accountable judgement that the business leaders themselves had to exercise. I wanted the discussion to distinguish administrative capacity from decisions that could not simply be delegated to a larger project office.

I recommended that six initiatives proceed while three move to a later sequence. That recommendation followed examination of dependencies and named responsibilities, not a formula limiting a leader to a fixed number of programmes. The relevant executives agreed the revised account. Specialist and risk obligations retained their authority; a strategy portfolio review did not make required work optional merely because its timetable was inconvenient. We retained C$1.6 million in programme replanning expense rather than presenting the change as a cost-free improvement in focus.

The three deferred initiatives lost their original timing advantage. Their sponsors also needed to explain the change to colleagues who had already prepared for it. I considered that consequence part of the decision, not a communications problem to be concealed. Equally, proceeding with six did not prove that all delivery difficulties had been removed. The revised plan made a particular conflict in accountable attention more intelligible; it did not certify the execution of the remaining work.

This case changed how I read a portfolio summary. A list can be complete while the organisation's account of what it has promised remains incomplete. The useful strategy contribution was to show where separately reasonable proposals competed for the same decisions and to support an explicit choice about sequence. The numbers describe this fictional situation only. They are not a staffing benchmark, an assurance of success or a reason to postpone an obligation owned by another authority.

In 2008, while leading business planning, I supported a specialist business-service proposition using 240 expressions of interest. The responses suggested that the subject mattered to prospective users, and I allowed that interest to support a growth assumption beyond what the evidence established. I had not sufficiently distinguished willingness to discuss a problem from willingness to devote time to a defined next step. The mistake belonged in the strategic assessment I endorsed, rather than solely in the way a later team tried to execute it.

A colleague challenged the interpretation after follow-up conversations produced a less uniform picture. Forty-six respondents were willing to enter a defined assessment. Seventy-four remained interested but could not allocate effort, while 120 represented general curiosity without a further commitment. Those three groups reconciled to the original 240. Even the forty-six had not become customers. Calling them conversions would have repeated the earlier error with a smaller number and a more persuasive-looking qualification.

The proposition's advocates argued that early interest was always imperfect and that requiring firmer evidence could delay useful experimentation. I agreed that exploration did not need to begin with a completed customer commitment. The problem was the conclusion I had attached to exploratory evidence. We could justify further learning on one basis without claiming a growth expectation on another. The account needed to make that distinction explicit before it was used to support a wider institutional commitment.

I withdrew the unsupported assumption and supported C$520, 000 in research and proposition rework. That was a real expense within the fictional case, not an estimated benefit from avoiding an imagined failure. The work could continue only on a revised basis that described what remained unknown. The credit, customer and professional decisions associated with any eventual service would retain their own authorities. A strategy research result did not constitute a banking offer or approval for an individual customer.

I keep this example because it makes my own standard open to challenge. It is easy to ask another sponsor whether evidence supports a conclusion and harder to recognise the same gap in a proposal one has helped develop. The correction did not make expressions of interest useless. It restored their actual meaning and required me to explain the cost of having asked them to carry more weight than they could bear.

In 2025 we reviewed whether the bank should continue pursuing new business in a selected segment. The strategic assessment favoured ending that expansion effort, but the decision concerned more than future origination. There were 4500 existing accounts with continuing arrangements and service responsibilities. I wanted the recommendation to distinguish a change in where the bank sought growth from any separate decision affecting the customers already present. A segment label did not erase those relationships or determine their treatment.

Some colleagues questioned why an exit proposal still carried material expenditure. They expected a clearer reduction in the bank's commitment once the strategic direction changed. Iona Brackenhurst, our fictional business-operations head, described the support the existing account population continued to require. I supported keeping that work inside the strategic account. If the decision looked attractive only after its continuing responsibilities had been removed from view, the comparison was incomplete rather than decisive.

The two-year operating plan retained C$3.2 million for support and transition. This was planned operating expense, not an accounting provision, a customer charge or a claim about the cost of terminating arrangements. Contract, conduct, risk and customer authorities remained responsible for the actual obligations. My recommendation did not authorise account closure or allow the strategy function to replace the judgement of those owners. The growth decision and the handling of existing relationships had to remain connected without being treated as identical.

The response gave up continued pursuit of new business in the segment while accepting that the institution's work there would continue. That could appear less dramatic than a clean exit announcement. It was also a more faithful description of the choice. The support plan needed review as circumstances developed, but it could not be dismissed simply because its cost complicated the headline. Nor should a continuing obligation become an excuse to avoid explaining whether the intended change in new-business activity had actually occurred.

I would bring this case to a discussion about what it means for a strategy to stop something. The appropriate evidence depends on what is being stopped and what remains owed. A bank can change its future direction while retaining responsibilities from its previous choices. The leadership account should be able to hold both facts at once, with the authority and expense clear enough for others to examine. All institutions and outcomes in this case are fictional.