Helen’s hybrid SKU session saved us from counting family-plan seats as unique LTV. I wanted more time on Android billing libraries; we only got a short appendix.
Flagship course
Curve to Cash
A twelve-week studio on LTV Forecasting for Apps. You leave with a cohort workbook, a payback clock, and a paragraph that states what would make the forecast wrong.
Learning outcomes
By the last session you should be able to assemble an LTV view from an install cohort, split it by country and surface, and explain the tail without waving at “industry decay”.
- Translate event logs into a retention curve that finance will recognise as observed, not wished.
- Separate payback for UA from a longer LTV used in planning.
- Handle hybrid apps so subscriptions and IAP are not stacked twice.
- Write a limitation that names missing data, promo weeks, and geo mix.
Modules
1. Naming the unit
User versus account versus device. Why restoring a purchase on a new phone is not a new lifetime.
2. Observed curves
D1 through D90 from your own weeks. We refuse to start with a Weibull fit if you have not plotted the raw series.
3. Money on the curve
ARPDAU is a blended fog. We attach revenue to the same cohort that produced the retention points.
4. Payback clocks
Media cost timing versus recognised revenue. Particularly messy for GB iOS after ATT-era signal loss.
5. Tails and honesty
How far you may extrapolate, and the sentence you must put under the chart when you do.
6. Hybrid SKUs
Subscriptions with consumable IAP, trials, and grace periods. The module most teams under-prepare for.
Instructor
Helen Quayle taught analytics inside two consumer app studios before opening this practice. She still sits with finance partners on forecast packs and refuses to “simplify LTV” into a single cell for a pitch deck.
Sessions run in English. Examples often use United Kingdom pricing, VAT-aware revenue, and App Store / Play settlement delays.
People who sat the room
Week two’s observed-curve homework was blunt. Our D7 looked fine until we split Ireland out of the GB install mix.
★★★★★
The limitation paragraph is now a slide we keep. Recruiters kept asking for a certificate; we sent the pack instead.
Questions we actually get
Do you take raw warehouse access?
No. You bring exports or anonymised cohort tables. We will not connect to production.
Is this suitable if we only have eight weeks of data?
Yes, and that is a real limitation: we will not let you publish a D365 LTV from eight weeks without a wide band and a written caveat. Some applicants want a precise number anyway; we will decline those enrolments.
Can finance attend without the growth lead?
They can. The course still uses product language. Expect to ask your growth counterpart for event definitions between sessions.