How we cut a SaaS cost-per-signup from $400 to $15
For a SaaS or ecommerce business, the whole model lives or dies on what it costs to acquire a customer. We took one Sydney platform from $400 per signup to $15. Here is the teardown.
Why cost-per-signup is the number that matters
You can have a great product and still go nowhere if it costs more to acquire a user than they are worth. Most SaaS and ecommerce businesses do not have a traffic problem, they have an efficiency problem: the spend is not tracked properly, the offer is not tight, and nobody is optimising daily. Fix those three and the cost per signup falls fast.
Step one: fix the tracking
You cannot optimise what you cannot measure. We started with server-side tracking and proper attribution so every dollar of spend was tied to an actual signup, not a guess. Once the data was clean, the waste became obvious, channels and audiences that looked fine on the surface were quietly burning budget.
Step two: tighten the offer and the funnel
A lower cost per signup is often a conversion-rate story, not a traffic story. We rebuilt the landing experience and the offer so more of the same traffic converted. Every extra percentage point of conversion drops the cost per signup without spending another cent.
Step three: optimise hourly, not monthly
This is where an AI agent earns its keep. Instead of reviewing the account once a month, the CMO agent watches spend by the hour, kills what is not converting and scales what is, overnight and on weekends. Small daily corrections compound into a very different number by month end.
The result
Cost per signup fell from $400 to $15, a 96% cut, on the same offer to a similar audience (see the SaaS case study, and a related childcare SaaS build). Same market, a fraction of the cost, which means the same budget now buys far more growth.
The takeaway
If your acquisition cost feels high, it is almost never one big problem. It is tracking you cannot trust, a funnel that leaks, and spend nobody watches closely enough. We fix all three, and the AI agents keep it optimised after we do. See how on the AI agents page.