# Confidence _The risk of false precision_ --- You're asked to sign off a proposal for a new product launch that looks good on paper: - Six months of upfront development, costing £500k - 50 new sales at £30k each within the next 18 months, totalling £1.5m You're not naïve - you know business cases are usually optimistic. But at a 3x payback, surely you have enough room for error? --- On the cost side, the estimate assumes six months of development. The variability is lop-sided - extremely unlikely to be finished in 1 month, but - let's be honest, as issues are uncovered, scope is refined - 12 months is not out of the question. So £500k is really a lower bound, with a reasonable range being perhaps £500k-£1m. Let's take £750k - 9 months of development - as a cautious case. On the sales side, the original assumption of £1.5m over 18 months works out to £250k per quarter (assuming a flat profile). With an extra 3 months of development, you've already lost one quarter of the original window - down to 15 months - 50 sales goes down to 40. You also know that in practice, two things tend to be true about the first few deals of a new product. Firstly, they always take longer to close than planned. And secondly, they rarely get off the ground at book price - early customers expect a discount for being guinea pigs and references. So rather than 40 deals at £30k each, you plan for 30 deals at £25k each: 30 * £25k = £750k All of a sudden, the proposal looks breakeven - £750k of development cost delivering £750k of new sales over the next two years. The 3x return has evaporated. --- Of course - the second 'cautious' set of assumptions are no more true than the first. The point is that the variance in assumptions matters. The business case is targeting a 3x return, but the honest range is probably somewhere between breakeven and 2x. Instead of "3x return over 2 years" a slide that says "between 1x and 3x, depending mainly on how long development actually takes" changes the conversation from approve-or-reject towards a constructive debate on what needs to be true, and which metrics matter to track progress. Some assumptions matter more than others. Price discounting may cost a fixed amount across the early deals: worth watching but likely not fatal. A development delay is a different animal. It adds direct cost, it takes months off the selling window, and the revenue you do collect arrives later, which matters if the project is being funded from cash flow rather than retained reserves (The grow-or-optimise question is explored further [[Beyond margins|here]]). It's impossible to remove uncertainty - better to face into it, rather than pretend it isn't there.