How to Cut CAC Without Cutting Growth
When a board asks for lower CAC, the instinct is to reduce spend. That does lower CAC, in the same way that not eating lowers your grocery bill. It also lowers growth, and usually the two fall together in a way nobody intended.
At Penfold we took CAC down by around 60% while revenue tripled over twelve months. Those two things happened together because the work was not a cost-cutting exercise. We changed who we were trying to acquire. That is almost always where the real movement is.
To be clear about what that claim is: those were outcomes during my time as CMO, produced by a team and a set of decisions, not by one channel or one person.
CAC is an output, not a lever
You cannot pull CAC directly. It is the result of who you target, what you offer them, how well you convert, and how much competition there is for that audience's attention.
Which means the useful question is never "how do we lower CAC?" It is "which of those things is producing a number this high?" Different answers lead to entirely different work.
If the audience is wrong, no amount of creative testing rescues it. If conversion is broken, better targeting just sends more people into a leaking funnel. If you are competing for the most expensive audience in your category, efficiency gains at the campaign level are marginal against the structural cost of that auction.
Changing who you acquire
The Penfold situation was this. Around 100,000 customers, a product people liked, and punishing economics, because growth depended on B2C paid acquisition in a category where trust takes time and customer value builds slowly. Ad costs were eating margin.
The move was from B2C to B2B2C, targeting employers rather than individuals. One employer relationship brings many members, which changes the arithmetic of acquisition entirely. Alongside that we built a channel programme around accountants, advisers and IFAs, where a single partner can introduce many companies.
We also analysed the economics by company size and buyer, and that is where a more specific finding emerged. There was a practical ceiling around 500 employees, above which consultants got involved in tendering and the process slowed and complicated. So we retargeted below it.
I got part of this wrong on the way through. I doubled down on HR as the buying centre faster than the evidence justified, and it took longer than it should have to see that HR often could not move the sales cycle forward on their own. In smaller firms the founder or owner is frequently the faster route, because the person who cares is the person who can sign. That correction cost us time.
Where organic changes the arithmetic
At Uncapped the pattern looked different. The macro turned in late 2021, ecommerce performance deteriorated, and Meta and Google costs were climbing. CAC was rising and customer quality was falling.
The reporting on this often gets framed as cutting paid. That is not quite what happened, and the distinction matters. What the data showed was a larger opportunity to scale organic content faster than we had been. We built SEO and content around the funding questions our redefined ICP was actually asking, and that channel had far more headroom than the paid channels we were fighting for.
Paid spend did come down, roughly 30%, and we became more selective. But the decision was driven by where the opportunity was, not by a budget-cutting target. Paid continued to matter. We just ran it more deliberately and stopped treating it as the only engine available.
Revenue doubled in nine months and CAC fell about 33%. Marketing and organic ended up driving around 80% of company revenue.
When paid should stay your primary engine
Organic is not automatically the answer, and the case for it is weaker than its advocates suggest in several common situations.
Paid should stay primary when your sales cycle is short enough that fast feedback is worth paying for, when you have a genuinely large addressable audience that is reachable through targeting, when unit economics already work at current costs, or when you need predictable volume on a timeline that content cannot meet. Organic compounds, but it compounds slowly, and a business with nine months of runway cannot wait for month fourteen.
The honest version is that most companies need both, weighted according to where the headroom actually is. That weighting should be revisited when the market moves, which is precisely what the Uncapped situation forced.
Conversion before acquisition
The cheapest CAC reduction available to most companies is not in acquisition at all. If you double conversion from enquiry to customer, you halve CAC without touching a single campaign.
Look at the whole path. Where do enquiries stall? How fast does sales follow up? How many qualified conversations are lost to slow response rather than genuine objection? Is the offer clear enough that a good-fit prospect understands it without a call?
This is unglamorous work and it is frequently the highest-return thing available. It also requires marketing and sales to look at the same numbers, which is often the actual blocker. That is why attribution rules need to be agreed before the result is known.
What to do first
Segment your CAC before you try to reduce it. Blended CAC hides the answer, because it averages the segment where the economics work with the segment where they do not. Break it down by channel, company size and buyer type. The pattern usually shows up quickly once you look.
Then ask whether the expensive segments are strategically necessary or simply habitual. Businesses persist with acquisition patterns for years because they are familiar rather than because they are efficient.
The largest gains I have seen came from changing the target, not from optimising against the existing one. That is a bigger decision than a campaign change, which is exactly why it tends to get deferred in favour of another round of creative testing.
If your CAC is climbing and the obvious optimisations are producing diminishing returns, the question is probably about who you are acquiring rather than how. Let's talk.
