The CFO conversation every UA lead has

Every UA lead eventually has some version of the same meeting: a CFO or finance business partner looking at a ROAS number that looks healthy and asking why the studio's cash position doesn't reflect it. The gap is rarely dishonesty on either side. It's two teams describing the same spend in two different vocabularies, and the meeting goes better once someone translates between them rather than defending their own number harder.

ROAS is not cash, and payback is not profit

Blended ROAS at day thirty or day ninety describes projected revenue against spend. For a live-service game most of that projected revenue arrives well after the spend that generated it, sometimes over a year or more for genres with long monetisation tails. Finance is modelling cash flow. It needs to know when the money actually lands, not when the model expects it to eventually total up to a healthy multiple. The blended ROAS calculation itself has known places it quietly goes wrong, such as mixing organic and paid cohorts or using a modelled rather than observed later-period number, and a finance partner who has been through one of those before will reasonably ask how you built the figure, not just what it says.

Payback period is the more useful bridge metric. It marks the point at which cumulative revenue from a cohort equals what you spent to acquire it, so it answers a cash question directly; it also needs a stated confidence band. A payback estimate built from six weeks of data on a cohort with a genuinely twelve-month monetisation tail is an extrapolation rather than an observation, and presenting it as the latter is where trust between UA and finance erodes fastest.

The capitalisation debate

Some studios capitalise a portion of UA spend on the balance sheet, treating it as an asset that will generate future revenue rather than an immediate expense, which changes reported profitability in the period the spend happens. That's a genuine accounting judgment with reasonable arguments on both sides. It's also not the UA team's decision. What the UA team can control is giving finance a defensible, consistently applied basis for whatever treatment the studio chooses, typically tied to a cohort's demonstrated payback trajectory rather than an aspirational one, and flagging early when a channel or campaign's actual performance is diverging from the assumption the capitalisation policy rests on.

Forecasting error bands, stated honestly

A UA forecast that presents a single number for next quarter's spend efficiency invites finance to hold that number to account as if it were a commitment. It's an estimate. Present a range instead, built from the actual variance the studio has observed in its own forecasts over the last several quarters rather than a generic confidence interval; that sets a more honest expectation and gives finance something they can plan a reserve against. The width of that band should itself be a tracked metric. A UA team that consistently forecasts inside a narrow, accurate range earns more budget latitude over time than one whose forecasts are directionally right but wildly imprecise, even if the second team's average forecast turns out closer to correct in hindsight.

Shared definitions before shared numbers

A surprising share of these disagreements trace back to the two teams using the same word to mean different things, not to any real disagreement about the underlying business. Finance may define a quarter's cohort by billing period while UA defines it by install date. One side may be looking at gross revenue and the other at net of platform fees. Neither team notices until the numbers fail to reconcile in a review. Agreeing a short shared glossary before the first joint report goes out, covering what counts as a cohort, which revenue basis applies and how international spend converts to the reporting currency, saves considerably more time than resolving it retroactively once two conflicting dashboards already exist and each team has started to trust its own.

The report format that actually ends the argument

The format that tends to work is one page, updated monthly, showing three things together rather than in separate documents: spend and blended ROAS by cohort age, a payback curve against the studio's agreed target window, and a short reconciliation line explaining any material gap between last month's forecast and what actually happened.

That reconciliation line is what makes the budget conversation, how to defend a UA line in a flat market, considerably easier. It converts the recurring argument about whether the numbers deserve trust into a smaller, specific conversation about one identified gap, and that is a meeting finance and UA can actually resolve rather than relitigate every month.

Related archive reading

These articles provide related context and remain subject to their stated review status.

• ROAS targets versus payback windows
• Profitable UA can still destroy a studio