What I chose when the simplest number did not settle the budget

In February 2022, two acquisition channels presented a similar opening account: each had activated 2,400 lines in the selected joining window. A renewal proposal treated that equality as a reason to preserve the allocation. The channels had different offers, commissions and customer mixes, so I was not comfortable allowing the activation total to carry the decision alone. Sales colleagues wanted an early commitment to retain placement for the next period.

I requested a ninety-day cohort account before recommending the next spend. At that cutoff, 1,800 lines remained active in the first group and 2,220 in the second. Across the original 4,800, 780 were no longer active. These were observed groups, not a randomised comparison. CFO Elise Norwyn required the contribution working to treat introductory support and channel commissions consistently before we discussed any ranking.

The decision was to redirect C$1.2 million of prospective channel spend within my existing authority. I accepted the risk of losing some familiar placement while retaining a smaller allocation to the first route for further observation. The channel team disagreed with reducing a route that had delivered the requested opening volume. Their objection deserved an answer: the original goal had measured a different event from the investment decision now being made.

The completed result was a changed allocation and a recurring cohort review. It was not evidence that the second channel caused stronger retention or that the move generated a stated return. Offer and customer differences still needed examination. The enterprise gained a more explicit account of the period it was buying, while accepting that a decision could not wait for perfect causal evidence.

I would establish the review window before the joining campaign next time. Introducing it at renewal made a reasonable question feel like a retrospective change to the team's scorecard. The people responsible for the opening goal should know in advance how the later investment decision will be judged.

A September 2023 portfolio proposal estimated C$900,000 in annual communication savings from combining two mobile brands. The smaller proposition focused on prepaid choice; the larger used a different commercial relationship. The operating attraction was clear: one campaign calendar, fewer assets and a more concentrated media account. I was asked to recommend whether marketing should support the combination.

My concern was the customer's reason for selecting the smaller proposition. Research with sixty selected participants found that thirty-seven emphasised control over replenishment and commitment, while twenty-three placed greater weight on other features. Those responses were not a forecast of churn. They did show that the proposed simplification needed a more precise account of what would remain recognisable to the customer.

I recommended retaining the distinct proposition for six further months, with C$340,000 additional communication authority. Finance challenged the expenditure against the projected savings, and the CEO asked for a decision date rather than an indefinite defence of brand heritage. We agreed to examine the combined proposition and its customer explanation before returning to the portfolio decision. The board accepted the bounded extension.

During that period, the team prepared three alternative proposition accounts and reviewed them with twenty-four further selected participants. Two alternatives blurred the replenishment distinction; one retained it more clearly. That finding informed the next commercial discussion without establishing future behaviour or the financial superiority of separate branding. The C$340,000 was a real cost of keeping the option open, not an investment whose return had already been measured.

I would bring the commercial terms into the first portfolio paper more explicitly. A name and visual identity can make a merger look simple while the underlying relationship remains materially different. My recommendation was strongest when it explained the customer's choice and the expenditure needed to investigate it, rather than treating a familiar brand as something that deserved preservation for its own sake.

In April 2024, a distribution review proposed directing the next acquisition increment entirely to digital enrolment. The online route was easier to count and its cost per completed application appeared lower than the assisted alternative. Distribution colleagues argued that some customers needed a conversation about commitment and payment timing before choosing. Neither claim established what additional business the company would gain from another dollar in either route.

I authorised a C$480,000 pilot across twelve locations for six months, working within my marketing remit while distribution retained responsibility for its staff and operating arrangements. The trade-off was explicit: this money could not also buy the planned digital acquisition increment. The digital team questioned why a less attributable route should survive when the online report already showed completed applications.

We recorded 360 choice sessions during the pilot. Of these, 228 ended in a completed application and 132 did not. Staff also recorded the question discussed, using categories agreed before the pilot. The application count did not establish activation, ongoing use or incrementality. A person might have applied digitally without the conversation, and the selected locations were not a controlled representation of the whole market.

At the review, I recommended retaining a smaller assisted programme while narrowing the questions it was intended to answer. The immediate result was a bounded continuation, not a claim that assistance had beaten digital acquisition. Finance kept the cost account separate from the application tally. Distribution used the recorded questions to revise the briefing provided to colleagues handling the conversations.

I would specify the later commercial follow-up earlier. We had enough evidence to understand the encounter more clearly, but less than we wanted about the relationship after application. The lesson was not that every difficult-to-measure route deserves protection. It was that a convenient attribution account should not become the sole definition of customer need before competing explanations have been considered.

The 2025 planning round included a proposal to move C$2.6 million from brand activity into acquisition channels ranked by their last recorded click. The proposed transfer had an appealing simplicity: fund the routes attached to completed applications and reduce the activity whose contribution was harder to assign. I agreed that the brand account needed challenge, but not that the ranking answered the whole allocation question.

I asked the analysts to state what the model observed and which influences it could not assign. Finance then helped distinguish a recorded conversion path from an estimate of additional business. The exercise did not produce a new causal model. It made the limitations of the existing account visible enough to discuss alternatives without treating a software ranking as a complete investment case.

I retained C$1.8 million of the disputed authority for a six-month brand programme and moved C$800,000 into the proposed acquisition routes. The CEO questioned whether this was merely a compromise protecting an established budget. My answer had to include a review date, specific evaluation questions and a willingness to reduce the remaining programme if its rationale no longer held. I could not defend it with a broad assertion that every brand investment pays back eventually.

At the six-month review, the activity and spend were reconciled, but the evidence still did not support an attributed incremental return. We continued a narrower programme under fresh authority rather than carrying the original amount forward automatically. The meaningful result was a budget process that recorded the judgement and its uncertainty, including the portion moved, the portion retained and the next decision required.

I would involve finance in defining the contested decision before the planning deadline. A late argument about attribution can sound like a refusal to be measured. The stronger position is to agree what each measure can support, then accept responsibility for the investment judgement that remains after the measurement work is complete.

In April 2026, I began a twelve-month review of how customers understand the period after an introductory offer. The programme has C$1.8 million authority and sixteen planned customer sessions. It examines the language, sequence and commercial explanation used when the opening arrangement changes. Pricing decisions remain with the appropriate commercial and finance colleagues; the review does not authorise me to alter an agreed customer contract.

The initial proposal favoured a quick communication refresh. I argued for examining the transition before choosing the new presentation, accepting a longer programme and fewer immediate creative outputs. The customer-experience team wanted changes sooner because colleagues were already answering questions. We agreed that urgent corrections to known inaccuracies should proceed through normal review, while the broader proposition work followed its own schedule.

By October 2026, nine sessions have taken place and seven remain. Forty-eight selected participants attended the completed sessions. Their accounts identified questions about the timing of the change, the later amount and the choices available. These are qualitative observations from selected participants. They are not a measured rate for the customer base or evidence that a revised explanation has changed renewal behaviour.

The current output is a set of issues for commercial, service and communication colleagues to examine together. No final proposition has been approved through this programme, and the full authority has not been described as spent. I report the work as open because an unfinished investigation should not become a success story simply when a website needs a recent example.

My concern for the remaining months is whether the team can distinguish an explanation problem from a proposition the customer understands but does not prefer. Better wording cannot resolve every commercial disagreement. I would judge the programme by the quality of the eventual decision and its supporting account, while reserving any claim about customer outcomes for evidence gathered after a defined change has actually occurred.