There is a structural problem in how multi-title publishers allocate budget, and it has nothing to do with whether the games are any good.
Executives measure revenue realised inside the financial year. Free-to-play mobile pays back over something like two years. Premium releases book most of their revenue at launch.
So when portfolio budget is allocated against in-year revenue, mobile is competing on a scoring system that structurally disadvantages it. Not because the returns are worse. Because most of them arrive after the measurement window closes.
Everyone involved knows this. Almost nobody has a defensible way to adjust for it, so the adjustment happens through argument and seniority instead.
Why the usual fix does not work
The obvious answer is to present lifetime value alongside in-year revenue. Most teams do. It rarely changes the decision, for two reasons.
Lifetime value is a forecast and in-year revenue is a fact. Or at least it is treated that way. When a number that will be audited sits next to a number that was modelled, the modelled one loses, regardless of which is more relevant to the decision.
The two numbers are not commensurable. Saying "this campaign returns 1.3× over 24 months" and "this release books $40m this year" does not tell anyone how to split a budget between them. They are answers to different questions.
The fix is not to argue harder for lifetime value. It is to produce the in-year number for the cohort, so both sides of the comparison are measured on the same clock.
Split the curve at the year boundary
Every acquisition cohort has a revenue curve that extends well past the fiscal year it was acquired in. That curve can be cut at the year end.
For a cohort acquired in month three, most of the first-year revenue lands inside the fiscal year. For a cohort acquired in month eleven, almost none of it does. Same cohort quality, same payback profile, wildly different contribution to the number the executive committee is looking at.
The calculator's defaults make it concrete. Each install returns $1.49 of net revenue in its first year and $0.37 in its first 30 days. A cohort acquired in the last month of the year books at most that first $0.37 inside the year, about a quarter of its first-year revenue, and the rest lands as carry. Payback there is day 207, so on a calendar fiscal year any cohort acquired after early June has not paid back by the year end, whatever its quality. Those are the calculator's published defaults, not a customer result.
Which produces three figures rather than one:
- In-year realised revenue. What this spend contributes to the year being measured. Directly comparable to a premium release.
- Lifetime value. What the spend returns in total.
- The carry. Revenue generated by this cohort that lands in the following year, and therefore arrives as apparently free performance in a period where it was not spent.
The third is the one that goes missing. A portfolio running consistent UA has a standing carry from prior cohorts propping up every year's numbers, and nobody books it as a return on last year's spend. Cut UA and the following year looks inexplicably soft. Scale it and the year after looks unaccountably strong.
Making the carry explicit changes the conversation. The question stops being "why does mobile want money for revenue that arrives later" and becomes "how much of this year's mobile revenue is carry from last year's spend, and what happens to next year if we cut now".
Acquisition timing becomes a real variable
Once the curve is cut at the year boundary, something falls out that most planning ignores: when you spend inside the year materially changes the in-year return, with no change in cohort quality at all.
Front-loaded spend has more of its curve inside the window. Back-loaded spend pushes the return into the carry. If your organisation is measured on in-year revenue, spend pacing is not an operational detail, it is one of the larger levers you have on the number you are judged by.
This is not an argument for front-loading. Sometimes the launch calendar decides it, sometimes media costs do. It is an argument for pacing being a deliberate decision with a modelled consequence, rather than a consequence of when budget happened to get approved.
The public-market version is worse
For listed publishers, the year is not the only clock. Quarterly expectations mean anything long-tail or risky is hard to justify unless it contributes inside the window. The effective horizon shortens from twelve months to three.
That pressure compounds when a portfolio is ageing. Declining titles produce a revenue gap that has to be filled from somewhere, and the things that fill a gap inside a quarter are rarely the things that build the following year.
There is no modelling fix for that. It is a governance problem. But the same split helps, because it names what is being traded. A decision to protect the quarter at the cost of next year's carry is a legitimate decision. A decision made without knowing that is the trade is not.
What to actually put in the plan
Three numbers per cohort or per campaign, on the same page:
- In-year realised revenue, comparable to every other line item competing for the same budget.
- Lifetime value, at whatever horizon your board uses.
- The carry into the following year, stated explicitly.
Plus the pacing assumption, named, because it moves the first number substantially.
That is enough to have an honest argument. It does not make mobile win the budget, and it should not. What it does is stop the argument being about which number to believe, and make it about which horizon the business is optimising for. That second argument is the one worth having.
Running it
Kadar models cohorts forward day by day and reports payback day and return on spend at D7, 30, 90, 180 and 365, which is the curve the split is made from.
Run your own numbers in the free calculator. No sign up.