A unit economics model
on your real data, not an industry average
A unit economics conversation without a real model is just two people guessing confidently in different directions. We join ad spend, purchases and churn into one model on your actual data, by channel and by country where it is available, so the business decision has a number behind it instead of a feeling.
Where the money leaks today
A business can have strong reviews, growing installs or orders, and a subscription or repeat-purchase model that quietly loses money on every customer it acquires, and nobody notices because the signals that would normally raise alarm, growth, engagement, positive feedback, all look healthy. We modeled exactly this for an app with a 4.8 App Store rating and real installs: churn at 34.5% a month gave a customer lifetime of 2.9 months, a CAC of $72 and an LTV of $44, an LTV to CAC ratio of 0.61 against a healthy benchmark closer to 3.
That gap is invisible without a real model, because CAC and LTV calculated loosely, averaged across all channels, assumed from industry benchmarks instead of your own churn, routinely hide exactly the problem that matters. The same modeling showed that cutting churn to 10% would move LTV to $149 and the ratio to 2.4, turning the one lever that mattered into a specific, actionable number instead of a vague “we should improve retention.”
What we do
We build the model from your real data: actual ad spend by channel joined to actual purchases or signups, and churn or repeat-purchase behavior from your own cohorts, not an industry-average assumption standing in for data you have not looked at yet. Where the data supports it, we break the model down by channel and by country, since unit economics frequently differ enormously by market, as we found directly across four countries on one app, with some markets showing healthy ratios and others showing zero purchases against real spend.
The output is a clear LTV to CAC ratio with the specific lever, churn, CAC on one channel, average order value, that would move it most, modeled with real numbers rather than asserted. We hand over the model itself, not just a summary, so you can rerun it as new data comes in rather than needing us back every quarter just to update a number.
What we need from you
Access to your ad accounts and spend history, your billing or order data, and whatever signal of churn or repeat behavior exists, even an informal one like a last-purchase date. If your numbers live in scattered systems, that is normal; tell us where each piece lives and we will join them.
How we measure
The model’s own LTV to CAC ratio against the benchmark for your business type, and whether the specific lever it identifies, once pulled, actually moves the number the way the model predicted.
Price and timeline
| Option | Price | What it covers | Timeline |
|---|---|---|---|
| Launch or audit | from $1,200 | CAC and LTV model by channel and cohort, rerunnable file | 2 weeks |
| Monthly management | from $600 / month | Model refreshed monthly as new data comes in | monthly, no lock-in |
| Full control, handover to your team | from $2,500 | Full model, documentation and training for your analyst or team | 3 weeks |
Related
This model informs pricing page optimisation and marketing budget allocation and forecasting directly. The full build is on the analytics service page. For the automated version, see cohort LTV modelling. Real examples: the trading app unit-economics audit and the Bali lead-routing project, where channel-level cost differences were modeled against real CRM outcomes.
Growing but not sure if each customer actually pays back their acquisition cost? Get in touch and bring your spend and churn data to the first call.
FAQ
How much does a unit economics model cost?
From $1,200 for a CAC and LTV model built on your real spend, purchase and churn data, delivered in 2 weeks.
What if our economics turn out to be bad?
We report that directly, with the specific lever that would fix it, churn reduction, CAC reduction on a specific channel, a pricing change, since finding this out with a real model is exactly the useful outcome, even when the headline number is uncomfortable.
What data do we need to have for this to be possible?
Ad spend by channel, purchase or signup data, and some signal of repeat behavior or churn, even informal. If your data is scattered across a CRM, a spreadsheet and an ad account, we can still build the model; that is closer to the normal starting point than a clean, joined dataset.
What do you need from us?
Access to your ad accounts, your billing or order data, and whatever churn or retention signal exists, a cancellation date, a last-purchase date, even a rough one.
How do you report the model?
A written readout of CAC, LTV and the ratio between them, broken down by the segments your data supports, plus a rerunnable model file so the numbers are not a one-time snapshot.