Six Assumptions That Kill New Business Launches
(07)
Overview
ExCo approved the launch. The business case cleared the hurdle rate. Six months later the product is late, over budget, and under-earning. The cause wasn't execution — it was six operating-model assumptions nobody tested before saying yes.
Year
2026
Industry
Capital Markets / Operating Model
Challenge
ExCo approved the new product three months ago. The market opportunity was clear, the revenue case was strong, the front office wanted it live by quarter-end. A program was stood up. Workstreams were assigned. The slide said "on track." Then the calls started. The booking model hadn't been decided. The front office assumed the product would book in one legal entity; operations discovered it needed to book in another, because the risk limits in the assumed entity couldn't absorb the new exposure without a capital reallocation nobody had requested. That decision, which should have taken a week in month one, took six weeks in month four, because by then, three downstream workstreams had built against the wrong assumption and had to unwind. This is not an unusual story. It is the most common story in new business launches across Capital Markets, and it happens because the business case that ExCo approved was a financial document, not an operating-model document. The revenue projection was tested. The cost estimate was tested. The six assumptions underneath them, the ones that determine whether the product actually launches on time, were not. I have spent most of my career at this seam, standing up new business initiatives in investment banking where the COO function sits between the business decision and the operating reality. The pattern is consistent. The same six assumptions go untested, the same six assumptions surface late, and the same six assumptions cost the launch a quarter or more of schedule. They are worth naming, because naming them is the first step toward testing them. Assumption 1: The booking model. Which legal entity books the product, in which jurisdiction, against which capital and liquidity framework. This decision cascades into every downstream system, control, and reporting requirement. When it is deferred — and it is almost always deferred, because it requires coordination between the front office, finance, risk, and operations that nobody wants to force in week one, every workstream plans against a guess. Some guess right. Some don't. The ones that don't discover it in month four, and the rework is expensive. Assumption 2: The control model. Who signs off on what, at which thresholds, against which limits. Most launches draft this late and socialize it with risk and compliance even later. Risk and compliance then ask sensible questions, the questions they are paid to ask, that require design changes nobody budgeted. The delay is not because risk is slow. It is because risk saw the design after it had hardened into commitments. Assumption 3: Capacity absorption. The business case assumes operations, technology, and the middle office will absorb the new product within their current headcount. Sometimes they can. Often they cannot, and the team that was supposed to absorb the work either drops balls on the existing book or staffs up reactively at a higher cost than a planned build would have required. Assumption 4: Regulatory capital and the ramp-up economics. This is the assumption that kills launches not by delaying them, but by changing their economics entirely. Regulators, FINMA among them, can impose capital requirements on a new product line that make the ramp-up punishing. A product that looks profitable at scale can be uneconomic in its first eighteen months if the capital charge during ramp-up is assessed against a small, immature book with limited diversification benefit. The COO work here is not waiting for the regulator to respond. It is proactive: structuring the initial deal pipeline to demonstrate a conservative approach, smaller notionals, tighter risk parameters, booking models that signal discipline, so the regulatory conversation starts from a position of credibility rather than defense. The institutions that treat the regulatory capital dimension as a timeline item ("how long until we get approval?") lose quarters. The institutions that treat it as a negotiation, showing the regulator a conservative ramp-up thesis, demonstrating that the booking and trading models are built for prudent growth, get to market faster with fewer conditions attached. This is senior work. It requires the COO to sit with the business, with risk, and with the regulatory relationship simultaneously, and to shape a ramp-up narrative the regulator can say yes to. It cannot be delegated to a project manager, and it cannot be started in month three. By month three, the regulator has already formed a view based on whatever the business filed initially, and changing that view is harder than setting it correctly the first time. Assumption 5: Cross-LE accounting. When the product touches more than one legal entity, and in a global bank, it almost always does, the transfer-pricing, inter-company accounting, and P&L attribution need to be designed, not inherited from an adjacent product that works differently. This surfaces as a month-three discovery, usually when finance asks where the revenue is being recognized and nobody has a clean answer. Assumption 6: Client lifecycle ownership. Who owns the client relationship after the trade is done. Origination, servicing, lifecycle events, complaints, off-boarding. In most launches, the front office owns origination and assumes someone else will own everything after. That someone else was never consulted and discovers their new responsibility when the first client issue arrives.
Impact
Each of these is knowable in week one. None of them requires the product to be built before the question can be asked. They are operating-model decisions, decisions about how the business will run, and they have answers that can be researched, debated, and locked before a single workstream kicks off. The cost of not testing them is real. In Capital Markets, where a successful new product can deliver run-rate revenue in the tens of millions annually, a single quarter of delay subtracts eight figures from the first-year P&L. That number never appears in the launch governance reporting, because it is lost revenue, invisible in a way that overrun spend is not. But it is the number the business case was built on, and when the launch slips, the business case no longer holds. The regulatory capital dimension deserves particular attention because its cost is not delay, it is margin erosion. A product that ramps under punitive capital requirements earns less per unit of risk than the business case assumed. If the capital charge is not negotiated proactively through conservative initial structuring, the product may launch on time and still underperform its economics for the first year or more. That gap between projected and actual return is the quietest way a launch fails, no one calls it a failure, because the product is live, but the P&L never reaches the trajectory the business case promised. The discipline to catch these assumptions is not complicated. It is a structured review, done in the first two weeks after ExCo approval, that asks six questions, one for each assumption, and requires a named owner and a committed answer date for each. The review takes two days of senior time. It is not a governance burden. It is the cheapest intervention available in a launch program, and it is the one most consistently skipped. Why is it skipped? Because testing assumptions feels like slowing down. The front office wants momentum. The program manager wants to show progress. Asking "have we decided the booking model?" in week one feels like bureaucracy when the energy in the room is about building. But the choice is not between momentum and discipline. It is between testing the assumption in week one when the answer costs nothing, or discovering it in month four when the answer costs a quarter. The COO role exists for this work. It sits at the intersection of business strategy, operations, technology, risk, and regulatory readiness, the exact intersection where untested assumptions live. The COO who forces the six questions in week one is not slowing the launch. They are compressing it, because every week of ambiguity deferred is a week of rework earned. Across the institutions I have worked with and observed, the launches that hit their first-revenue date are not the ones with the best technology, the most enthusiastic front office, or the biggest budget. They are the ones where someone, usually a COO or a COO-equivalent, tested the operating-model assumptions before the program plan was written. The assumptions were the same every time. The difference was whether anyone named them. The business case gets the launch approved. The operating model determines whether the launch delivers what the business case promised. The six assumptions sit between the two, and they are either tested early or paid for late. There is no third option.