Customer Acquisition
Growth is not a traffic problem. It is a payback problem.
Acquisition economics modelled to payback period, then media bought to fit.
-34%
Reduction in new-customer acquisition cost
The problem
Most acquisition targets are set backwards. A CPA number gets agreed in a planning cycle, usually derived from last year’s CPA plus an improvement, and then media is bought to hit it. Nobody has checked whether that CPA is affordable at current margins, how long the customer takes to repay it, or whether the cohorts acquired at that price behave like the cohorts the model assumed. Businesses discover the problem when cash gets tight, which is exactly when acquisition spend has to be cut, which is the worst possible time to find out the model was wrong.
Process
How we approach it
Build the unit economics before the media plan
Contribution margin per first order, repeat rate and repeat interval, gross margin by product mix or plan tier, and the resulting payback curve. This is finance work, not marketing work, and we do it with your finance team rather than around them. It produces the only number that matters: what you can afford to pay for a customer and still be solvent at scale.
Segment by cohort value, not by channel
Two customers acquired at the same CPA can be worth very different amounts depending on entry product, discount depth at first order, and acquisition channel. We track cohorts by acquisition source and month, and report payback by cohort rather than blended, so a channel that acquires cheap low-value customers stops looking efficient.
Set a bid target the algorithm can actually use
Once payback tolerance is agreed, it becomes a target CPA or target ROAS per channel, tiered by expected cohort value. That target goes into the bidding platforms with the conversion values to match. Most of the work of acquisition is making the economics legible to a machine that is going to spend your money at four in the morning.
Review the model against reality
Payback assumptions decay. Repeat rates move, margins compress, discounting creeps, and exchange rates move underneath any target set in a currency you do not sell in. We re-run the cohort analysis against actual behaviour on an agreed cycle and adjust targets, rather than defending a model built at the start of the engagement.
Deliverables
What you get
Acquisition economics model covering contribution margin per first order, repeat rate, repeat interval and payback curve, built with your finance team
Maximum affordable acquisition cost by segment, with the assumptions stated explicitly so they can be challenged
Cohort reporting by acquisition month and channel, showing payback progress rather than blended lifetime value
Channel-level target CPA or target ROAS derived from the payback model and loaded into bidding platforms with matched conversion values
New-customer versus returning-customer split reported as standard across all paid channels
Re-forecast of the acquisition model against actual cohort behaviour, with target changes recommended and justified
A documented stop-loss position: the performance level at which we would recommend reducing spend rather than defending it
Industries
Where we run this
Ecommerce
Paid acquisition managed against contribution margin and breakeven efficiency, not platform-reported return.
SaaS
Demand capture and creation for B2B software, measured on payback duration rather than lead volume.
Finance
Acquisition for regulated firms, where the advertisement is itself a regulated document.
Gaming
User acquisition for free-to-play titles, measured on retention curves and payback windows.
B2B
Paid acquisition for considered purchases with long cycles, small keyword universes and offline conversions.
Frequently asked
We do not have clean lifetime value data. Can you still do this?
Yes, and most clients do not have it at the start. We begin with a payback window you can observe — often first 90 or 180 days — rather than a modelled lifetime value that nobody believes. Short observed windows are less impressive and considerably more useful.
Is this different from what our media agency already does?
It is upstream of it. A media agency optimises towards the target it is given. This work sets the target, and it is frequently the reason a well-run media account is still unprofitable. If your existing agency is doing this properly, you do not need us for it.
How does this work if we have long or irregular purchase cycles?
Longer cycles make the observed-payback approach more important, not less, because modelled lifetime value gets less reliable the further out you project. We extend the observation window and lean harder on leading indicators — second-purchase rate, activation, and in subscription models, early retention curves.
Will you tell us to spend less?
Sometimes. If the model says the current CPA is above affordable and cannot be brought down, the recommendation is to reduce spend or change the offer. We would rather say that early than defend an unaffordable position for a year.
Start with the audit.
It has a defined scope and a defined deliverable, and it is deliberately separable from anything that follows. If the audit says your current setup is fine, that is a legitimate outcome and we will say so.