Rising CAC Is a Symptom, Not the Disease

#Commerce Infrastructure 6 min read
Two feedback loops side by side, one brightly instrumented and one left dark, a brand watching only the loop it can measure.

A survey of more than a hundred Indian D2C founders this year surfaced a number that looks like a marketing problem and is actually something else. Sixty-two percent named creative fatigue as their biggest bottleneck: more spend, more ad variations, flatter and flatter returns. Most are stuck between one and fifty crore in revenue, repeat rates sitting somewhere between ten and thirty percent, and more than half admit they under-invest in keeping the customers they already paid to acquire.

When acquisition costs climb, the instinct is to treat it as an acquisition problem. Make better creative. Find fresh audiences. Shift budget to whatever channel is working this week. That instinct is completely reasonable. It is exactly what the dashboard is pointing at, and for a while it works.

But rising customer acquisition cost is often a symptom. The disease tends to sit somewhere the dashboard cannot see.

What the dashboard actually measures

Look at the metrics a growth team lives inside. Cost per mille, click-through rate, return on ad spend, cost per acquisition. Every one of them describes the mechanics of the ad: how many people saw it, how many clicked, what it cost to get a click to convert. They are precise, they update fast, and they are genuinely useful.

Notice what none of them describes. Whether the product actually landed with a mainstream buyer. Whether the person who bought will still be using it in three months. Whether they would tell a friend, or quietly never reorder. The dashboard is exhaustively instrumented on one side of the business and almost silent on the other. It measures the ad in fine detail and the product barely at all.

Two loops, and only one of them has telemetry

There are really two feedback loops running in any consumer brand, and they could not be more different in how well you can see them.

Acquisition is a fast loop with rich telemetry. You change a creative in the morning and by lunch you know the click-through rate. The signal is immediate, quantified, and sitting right there on the screen.

Product-resonance is a slow loop with almost no telemetry. Does this genuinely fit the person's life? Will they come back without being retargeted into it? The signal takes weeks or months to form, arrives faint, and gets tangled up with everything else happening in the market. There is no clean dashboard tile for it.

So teams do the very human thing. They optimize the loop they can measure and starve the one they cannot. They pour effort into the fast, bright, quantified loop, and read the resulting cost inflation as a marketing problem to be solved with more marketing. Creative fatigue at sixty-two percent is the tell. Teams are iterating on the ad, not on the understanding. It is a reasoning failure driven by acting on the variable you can see instead of the one that decides the outcome, the same failure I keep coming back to in how I evaluate any system that has to act on incomplete state.

The mirror problem

Here is the part that makes the loop hard to escape.

The signal a brand uses to decide whether it is winning, return on ad spend, engagement, conversion, is produced by the same channel it is optimizing on. The platform that sells you the ad is also the one that tells you how the ad did. You are grading the campaign with the campaign's own report card.

This is an old distributed-systems trap wearing a marketing costume. A component's report about its own performance is a claim, not a measurement, and the moment the thing being evaluated is also the source of the evaluation, you have lost the ability to check it. A strong return-on-ad-spend number can be perfectly real and still tell you nothing about whether the product will survive contact with a mainstream buyer at scale. It describes the ad's efficiency, not the product's truth.

Founders think they have observability. What they often have is a mirror.

The channel that drives volume shows you the least

It gets harder still, because of where the volume increasingly comes from. Quick commerce has become a default distribution channel for a lot of Indian D2C, and it is precisely the channel where a brand sees its own customer least. The platform owns the shelf, the ranking, the transaction, and the relationship. The brand gets a sale and a set of fees, and very little of the "who bought this, and will they buy again" that was supposed to be the whole advantage of going direct.

This is not an accusation. It is the structural shape of the arrangement. The platform is optimizing its own surface, honestly, and the brand's visibility into its own demand is a casualty of that structure, not a conspiracy inside it. But the effect is real: the faster a brand leans on the channels that scale, the thinner its own read on why anyone is buying.

What this is not

To be precise, because it is easy to overshoot. This is not an argument that performance marketing is broken or that CAC does not matter. Acquisition telemetry is real, and the fast loop deserves to be measured well. And not every stalled brand has a resonance problem; some genuinely have a pricing or distribution problem instead.

The failure is narrower and more specific. It is mistaking a rich measurement of the ad for a measurement of the product, and letting the loud, fast, well-instrumented loop set the strategy while the quiet, slow, decisive one goes unwatched. The survey's shape, creative fatigue, thin repeat rates, under-investment in retention, is what that blind spot looks like from the outside.

The invoice arrives late

You can feed the loop you can see for a long time before you notice that the loop you cannot see is where the business was actually being decided. Rising CAC is, in a sense, the invoice for that blind spot, and it always arrives late, after the budget has already gone into iterating on ads for a product the market was quietly declining to adopt.

The founders in that survey are not short on data. On one side of the business they are drowning in it. The question none of the dashboards answer is the one that decides everything: does the product actually resonate, and would you know in time if it did not?

A dashboard can tell you what your ad did. It cannot tell you whether your product landed. Those are different questions, and only one of them is on the screen.

Sources

The figures here are reported by the survey and its coverage, not independently verified. Read them as directional.