There is a familiar moment in most growth businesses when customer acquisition begins to feel uncomfortable. Media costs rise, conversion softens, the easiest audiences have already been reached, and finance starts asking why the organisation is spending more to create roughly the same amount of growth.
The instinctive response is predictable: optimise harder. Improve creative, refine targeting, renegotiate media, increase conversion, test another channel, revisit attribution. All of that is sensible, and sometimes it is exactly what the business needs.
But I think CAC is often diagnosed too narrowly.
Customer acquisition cost is easy to isolate because the spend is visible and the transaction arrives relatively quickly afterwards. What happens beyond that point is much less tidy. Retention may sit with CRM, returns with operations, service with customer care, margin with commercial, loyalty with marketing, and product experience somewhere between digital, stores and technology.
The P&L, inconveniently, receives all of them.
So when acquisition gets more expensive, I would resist starting with the question of how to make the next customer cheaper. The more useful question is whether the customer becomes valuable enough once acquired.
A low CAC can hide a poor customer
Digital commerce made acquisition wonderfully measurable. Spend this much, generate this traffic, convert this percentage, acquire a customer at this cost. There is something reassuring about a metric that can be displayed neatly beside a target and discussed every Monday morning.
The problem is that a cheap acquisition is not necessarily good growth.
Consider two customers. One costs $25 to acquire, purchases only during promotion, returns part of the order, contacts service twice and never buys again. Another costs $75, purchases mostly at full price, comes back several months later, gradually expands into other categories and stays with the brand for years.
The first customer looks better on the acquisition dashboard. The second looks better almost everywhere else.
This is not an argument for ignoring CAC. Businesses that acquire customers inefficiently eventually run into trouble, and nobody should disguise poor media performance behind an optimistic lifetime-value model. But CAC is the entry price into a relationship, not the economic value of the relationship itself.
The distinction becomes particularly important when companies are growing quickly. New-customer numbers can look healthy while the underlying cohorts quietly become less attractive. More discounting is required, repeat behaviour weakens, full-price mix falls, and the business needs ever more acquisition simply to replace customers who never developed into much of a relationship.
Growth is still happening. It is just becoming more expensive to sustain.
The customer P&L is usually fragmented across the organisation
One reason this problem persists is that customer economics rarely belong to one team.
Marketing is asked to acquire. CRM is asked to retain. Commercial teams protect margin. Operations manage fulfilment and returns. Customer service focuses on resolution and cost. Each function can therefore make a perfectly rational decision inside its own mandate while producing a poor result for the customer relationship as a whole.
I have seen versions of this repeatedly. A marketing team hits a new-customer target through a strong promotional campaign, and the initial results look excellent. Three or six months later, the cohort behaves poorly: low repeat, heavy discount dependence, high returns or weak contribution.
By then, those outcomes may sit in somebody else’s report.
This is not a marketing problem so much as an operating-model problem. People optimise for the measures the organisation gives them. If the target is new customers, they will find new customers. If the target is revenue, they will find revenue. If nobody is accountable for whether those customers become profitable relationships, that part of the economics has been left to chance.
A stronger growth model therefore needs some shared view of customer value across functions. Not necessarily one department owning everything, but one economic truth about what constitutes a good customer.
That sounds obvious. It becomes less obvious the moment different teams have different targets, budgets and definitions of success.
Rising CAC is often a symptom, not the disease
Acquisition costs do rise for straightforward reasons. Competition intensifies, media becomes more expensive, privacy changes reduce targeting efficiency, or a business moves beyond its earliest adopters into audiences that are naturally harder to persuade.
But sometimes CAC is telling you something about the business rather than the media market.
Perhaps the proposition has become less distinctive. Perhaps organic demand is weak and performance marketing is carrying too much of the burden. Perhaps the business is repeatedly paying to bring back customers it should have retained. Perhaps promotions have trained customers to wait rather than buy. Perhaps the post-purchase experience quietly destroys some of the value created before checkout.
If that is the case, improving acquisition efficiency treats the symptom.
A better campaign may reduce CAC temporarily, but it cannot make a forgettable proposition memorable. Better targeting cannot compensate indefinitely for weak repeat behaviour. Another attribution model will not repair a service experience that makes customers reluctant to return.
There is a point where every additional dollar spent acquiring demand becomes more expensive because the business has not created enough reasons for that demand to persist.
That is where growth teams need to look beyond the top of the funnel.
Retention changes how much CAC you can afford
The most useful way to think about rising CAC, in my view, is not simply “How do we bring the number down?” It is “How do we improve the economics around it?”
Retention is central to that.
If customers stay longer, purchase more frequently and buy across more categories, the original acquisition cost is spread across a larger relationship. If full-price mix improves, the same customer produces more contribution. If returns fall, revenue becomes more valuable. If service recovers a relationship that would otherwise have churned, the economics of the original acquisition improve without touching the original media cost at all.
These levers are less theatrical than acquisition. Nobody puts a countdown clock on a modest improvement in repeat rate. Reducing unnecessary incentives rarely generates the internal excitement of a major campaign launch.
But from a P&L perspective, they can be far more powerful.
This is why I like asking a slightly different question in growth conversations: where should the next unit of investment go?
Sometimes the answer is acquisition. Sometimes it is retention, conversion, inventory availability, service recovery, loyalty, product experience or better merchandising. The discipline comes from comparing those opportunities against one another instead of assuming growth begins and ends with media.
The answer will not always belong to marketing.
That is exactly the point.
Customer intelligence should decide where we invest, not just what we communicate
This is where customer data and AI can become commercially useful rather than merely impressive.
Retailers have spent years using data to personalise communications. We reorder products on a homepage, change an email, alter recommendations and create increasingly narrow audiences. There is value in that, but it still treats personalisation largely as a messaging problem.
The larger opportunity is to personalise investment.
Two customers can look superficially similar and deserve very different commercial treatment. One may need an incentive to convert; another would have purchased anyway. One customer may be worth paying significantly more to reacquire because their future value is high. Another may appear valuable because of gross spend but become much less attractive after discounts, returns and cost-to-serve are included.
A customer who has just experienced a service failure may need acknowledgement rather than another promotion. A loyal customer may value access, convenience or recognition more than a percentage discount. A dormant customer may not be worth winning back simply because a model says they are likely to respond.
This is where AI should help move the business from broad segmentation toward better resource allocation. The useful question is not only who is likely to act, but whose behaviour is economically worth changing.
That distinction matters because propensity is not value.
The customer most likely to use a discount is not necessarily the customer who needs one. The customer most likely to respond to reactivation may not be the one whose return creates attractive economics. Better models can help, but only if the organisation is willing to act on conclusions that may conflict with traditional campaign targets.
Technology can improve the recommendation. Incentives still determine whether anyone follows it.
LTV is useful, but it can become another comforting fiction
Once businesses recognise the limits of CAC, the natural response is often to embrace lifetime value. That is directionally right, but LTV can become just as misleading if treated as a fact rather than an estimate.
Lifetime value depends on assumptions about retention, frequency, margin, discounting, returns and cost-to-serve. A model can tell you that a customer is worth $1,000 over three years while quietly assuming the customer remains loyal, margins stay stable and the organisation behaves sensibly throughout.
Those assumptions deserve scrutiny.
Gross revenue is not customer value. Repeat purchasing driven entirely by discounting is not necessarily healthy retention. A high-spend customer with unusually high returns can look very different once contribution is considered. Likewise, a lower-spend customer who buys consistently at full price and requires little intervention may be more attractive than the headline revenue suggests.
The point is not to abandon LTV. It is to make the model honest enough to inform decisions.
A useful customer economics framework should force the organisation to connect acquisition source, margin, repeat behaviour, returns, service cost and retention over time. When those pieces sit together, the conversation becomes much more interesting than whether CAC was two dollars above target last month.
It also becomes much harder for individual functions to declare victory in isolation.
The strongest acquisition strategy may be a business customers want to return to
There is a slightly unfashionable conclusion to all of this.
One of the best ways to improve acquisition economics is to build a proposition that reduces how much persuasion the business needs to purchase.
Good product matters. Reliable service matters. Trust matters. Distinctiveness matters. Availability matters. A genuinely useful loyalty proposition matters. So does a brand people remember before the advertising platform reminds them.
None of these is as easy to optimise as a campaign bid.
They are also harder for competitors to copy.
A strong brand lowers the amount of paid persuasion required to generate interest. A good experience increases the probability that the customer returns without another acquisition campaign. A distinctive proposition attracts demand that is less dependent on discounting.
Marketing can amplify all of that. It cannot manufacture it indefinitely.
AI will certainly make the acquisition machine more efficient. Creative will become faster to produce, targeting more dynamic, media allocation more intelligent and customer prediction more precise.
But everybody else will have access to much of the same capability.
When the tools become widely available, advantage moves back into the things that are harder to commoditise: proprietary customer understanding, proposition, brand, operating execution and the ability to identify which customers are genuinely worth acquiring.
The question is therefore not how cheaply the business can buy its next customer.
It is how much profitable value the business can create from the relationship once it does.
That is a broader definition of growth.
It is also a more honest one.
CAC may continue rising. In many categories, it probably will.
The answer is not necessarily to spend less acquiring customers.
It may be to build customers worth spending more to acquire.