What a D2C customer actually costs: the arithmetic to run before you buy more ads

Where ad costs actually stand

Meta is where most D2C budgets sit, so start there. Triple Whale's benchmark across more than 40,000 brands, published in August 2026 and covering August 2025 to July 2026, puts the median Meta CPM at $15.06, up 13.2 percent in a year. The averages hide worse category numbers: beauty CPMs rose 33 percent to $18.80, and health and wellness pays the most at $21.80. The median cost per acquisition across the platform sits at $38.99, up 3.1 percent.

Search is not the escape route. WordStream's 2026 Google Ads benchmarks, published May 2026 on data from April 2025 to March 2026, put the average cost per click at $5.42, with 61 percent of industries paying more than the year before.

The longer arc is the part that should concentrate the mind. SimplicityDX calculated in July 2022 that the average merchant lost $29 on each new customer acquired, against $9 in 2013, with acquisition costs and returns accounting for virtually all of the difference. That was before the rises above. There is no single authoritative D2C CAC index, and the platform benchmarks blend brand sizes and categories, so treat any single figure as a market signal rather than your number. Your number comes from your own accounts, which is the point of this article.

The arithmetic to run before approving more spend

Take one representative order and build the stack. Say a £60 average order at a 65 percent gross margin: £39 of product margin. Subtract pick, pack and shipping, typically £8 to £12 for a lightweight product in the UK. Subtract payment processing at roughly 2 percent of the order. Subtract the cost of returns: the NRF and Happy Returns put the online return rate at 19.3 percent of sales in October 2025, and in fashion it runs higher, so a brand that ignores returns in its acquisition maths is overstating margin by a fifth. What remains, perhaps £22 to £28 on this example, is the contribution available to pay for marketing.

Now put your true new-customer acquisition cost against it. If it costs £30 to £40 to acquire, the first order is at or below break-even, which is where a large share of D2C brands now operate. That is not automatically a problem. It becomes one when nobody has checked the second half of the equation: how many first-time buyers come back, how quickly, and at what margin. Two numbers decide whether more spend is defensible: contribution per first order after acquisition cost, and the share of customers who make a second purchase within a window you can afford to wait for. If sixty-day repeat behaviour carries the customer past break-even, spend can scale. If it does not, more budget deepens the hole faster.

Blended numbers hide the problem

Most dashboards report blended figures: total revenue over total spend, returning customers included. A brand with a healthy repeat base can show a comfortable blended ratio while paying more for each new customer than that customer will ever return. The mechanism is worth spelling out: as the repeat base grows, returning revenue pads the numerator every month, so the blended ratio can hold steady or improve while marginal new-customer economics quietly deteriorate. The owner sees a stable dashboard and approves the next increase; the accountant sees the cash conversion worsen and cannot say why. Separate new-customer acquisition cost from the blended number, and new-customer revenue from returning revenue, before reading any ROAS figure. Then look at cohorts rather than months: what a customer acquired in January had returned by April tells you more than any in-month ratio. This is unglamorous reporting work, and it is the difference between a growth decision and a guess.

When the ad problem is not an ad problem

Rising CPMs are a market condition. Every competitor in your auction pays the same rent. What separates brands that can afford the auction from brands that cannot is what happens after the click, and that is where most "our ads stopped working" conversations actually belong.

• The offer. If click-through rates are fine but the page converts poorly, the ads are doing their job and the proposition is not. A landing page converting at 1 percent needs two and a half times the acquisition budget of one converting at 2.5 percent, at identical CPMs. Testing more creative against a weak offer buys more expensive proof of the same problem.
• The positioning. Creative that fatigues in a fortnight is usually creative with nothing distinct to say. If the only lever left is a discount, the auction will be won by whoever has the deepest margin, and that is rarely the owner-led brand.
• The price. If every line of the margin stack is honest and contribution is still negative after a realistic repeat window, no media buyer can fix it. The product is underpriced or the cost base is wrong, and that decision sits with the owner, not the ad account.

What to do

Build the margin stack for one representative order, including returns. Separate new-customer cost from blended cost. Set a CAC ceiling from contribution and your measured repeat rate, not from a target ROAS chosen for comfort. Then, and only then, decide whether the next pound goes into media or into the page, the offer or the price. In our own work the page and the measurement usually come before the budget increase, which is why Paid search and paid social, Landing and sales pages and Analytics and reporting are built as one programme rather than three. The free Business Scan runs this arithmetic on your own numbers, which is worth more than any benchmark in this article.