Cohort and LTV modelling on autopilot:
know what a customer is worth before they churn
Lifetime value usually gets calculated once, in a spreadsheet, for a board deck, and then goes stale the moment the next acquisition channel changes the mix. We build a model that tracks retention by cohort and projects LTV on a schedule, so the number driving your acquisition spend reflects this quarter, not last year.
The process today
Lifetime value gets calculated the way most teams calculate anything hard: once, carefully, by someone who is good at spreadsheets, usually ahead of a board meeting or a budget planning cycle. The number is often blended across all customers regardless of when or how they were acquired, which hides the fact that a cohort acquired through one channel three months ago is behaving very differently from a cohort acquired through another channel a year ago.
Between those calculations, the business keeps acquiring customers and keeps spending against an LTV number that is already out of date the moment the acquisition mix shifts, a new channel ramps up, a pricing change goes live, a product change affects retention. Nobody notices the number is stale until the next quarterly exercise, by which time several months of acquisition spend has already been allocated against an assumption nobody was actively checking.
The deeper issue is that a single blended LTV number hides the channel-level truth that actually matters for a budget decision: one channel might have a short payback and modest ceiling, another a long payback and a much higher ceiling, and averaging them together tells a growth team nothing useful about where the next dollar should go.
What the agent does
The model tracks retention by acquisition cohort, typically by month and by channel, reading from your billing, subscription or transaction data, and projects lifetime value for each cohort using the retention curve it has actually observed rather than a single industry rule of thumb. Mature cohorts with enough history get a tight projection; newer cohorts get a wider confidence range that narrows automatically as they age and more real data comes in.
LTV by channel sits next to that channel’s CAC on the same dashboard, so payback period is visible without a separate manual calculation, and a growth team can see in one place which channels are paying back fast, which are slower but larger in ceiling, and which are not paying back at all. The whole view refreshes on a schedule, weekly or monthly depending on your volume, instead of being rebuilt from scratch for each board deck.
Because the model is trained on your own retention history, it reflects what your product and pricing actually do to customer behaviour, not a generic SaaS or e-commerce benchmark that may not apply to your specific business.
What stays with humans
Budget allocation across channels, pricing decisions, and any strategic call about accepting a longer payback period for a channel with a higher ceiling stay with your growth and finance teams. The model provides the cohort-level number; it does not move spend or set a target CAC on its own.
Guards
Every projection is logged alongside the retention data that produced it, and the model’s track record is checked against cohorts that have since matured, so you can see how close past projections came to reality. New-cohort projections are always shown with a confidence range rather than a single confident number, and the model is retrained on a schedule as new cohort data comes in.
Price and timeline
| Option | Price | What it covers | Timeline |
|---|---|---|---|
| Single automation | from $900 | Main acquisition channels, cohort tracking, LTV projection | 7 to 14 days |
| Department package | from $2,800 | LTV modelling plus channel-level CAC dashboards for growth and finance | 3 to 5 weeks |
Running cost is usually $25 to $90 a month depending on transaction volume.
Related
Pair this with contract renewal risk scoring so an at-risk high-LTV account gets priority attention, and with ad budget allocation so spend actually follows the channel-level payback this model reveals. The full package breakdown is on the AI agents service page and the automation-everything overview; for a real analytics build, see the two-brand analytics hub case study and the trading app unit economics audit.
Ready to see LTV by cohort instead of one blended guess? Get in touch and we will look at your retention data in the first call.
Tired of doing this by hand? We can take the whole routine off your team, not just this step: Routine takeover, from $400 →
FAQ
How much does cohort and LTV modelling cost?
From $900 for tracking and projection across your main acquisition channels, live in 7 to 14 days. A department package adding channel-level CAC comparison and dashboards usually starts at $2,800.
How accurate can an LTV projection really be for a new cohort?
A projection for a brand-new cohort is necessarily less certain than one for a cohort with two years of history, and the model says so with a confidence range rather than presenting a guess as fact. Projections sharpen automatically as each cohort matures.
What data does this need?
Transaction or subscription history tagged with acquisition date and channel, typically from your billing system, CRM or e-commerce platform. The cleaner the channel tagging, the sharper the by-channel breakdown.
Can it tell us which channel is actually worth the spend?
It shows LTV by channel next to that channel's CAC, so payback period and long-run value are visible side by side. The budget decision based on that data stays with your growth team.
Does this replace our finance team's unit economics model?
No, it feeds it. This automates the tracking and projection layer so finance gets a current, cohort-level number instead of rebuilding the analysis by hand every quarter.