What to do when customer acquisition cost keeps climbing
Before cutting spend, work out which kind of rising CAC you have. Degraded tracking, audience saturation, creative fatigue, a shift in channel mix and more expensive auctions look identical on a dashboard and call for opposite responses. Then judge the answer on payback period against contribution margin, not CAC alone.
First, check whether CAC actually rose
Measured CAC is a ratio between spend you know exactly and conversions you only partly observe. When observation degrades, CAC worsens on paper while nothing changed in the business. Falling consent rates, a duplicated purchase event, a checkout change that dropped a parameter, a UTM convention somebody edited quietly: each moves reported conversions without moving orders. The check is cheap. Compare total orders and revenue in your store platform against the same window last quarter and last year, then compare that shape with the ad platform reports. If store level orders are steady while platform conversions have fallen, the problem is measurement, and cutting spend would be the wrong response. Fix tracking before diagnosing anything else, because every step below depends on trustworthy numbers.
The four causes to rule out in order
Once the numbers are trustworthy, four causes remain. Audience saturation shows up as frequency climbing and reach flattening at the current budget: more money to reach the same people. Creative fatigue looks similar but sits at the ad level, where top performers have run for weeks with nothing new behind them. A shift in channel mix is the quiet one: total CAC can rise while every channel holds steady, because budget moved to a more expensive channel or prospecting grew as a share of spend. Genuinely more expensive auctions are the last explanation, not the first. Each needs a different response, which is why the order matters: saturation needs new audiences or a new offer, fatigue needs production capacity, mix needs a budget decision, and auction pressure needs a conversation about what a customer is worth. That sequence is where we start at plaidtesting.com/services/paid-media/ before touching a budget.
Why CAC alone cannot tell you whether to keep spending
CAC is a cost with no time dimension and no margin inside it. Two brands with the same CAC can be in very different positions: one sells a thin margin item that is rarely reordered, the other a product with healthy contribution margin and a real repeat purchase habit. The second brand can pay far more for a customer and still be right. So rising CAC is not automatically bad news, and falling CAC is not automatically good. CAC that drops because you retreated into your warmest audiences usually means growth is about to stall. The decision needs a measure carrying both time and margin. That is the CAC payback period: how long a new customer’s contribution margin takes to cover what you paid to acquire them.
How payback period changes what you will pay
Once the decision runs on payback period, the question is not whether CAC went up but whether the payback window still sits inside what the business can fund. A brand financing growth from cash flow needs a short window. One with capital behind it can accept a longer one and outbid rivals who cannot. Set the window with your finance side, using contribution margin after cost of goods, shipping and payment fees, counting repeat orders only where the data is reliable. Rising CAC then becomes answerable: at this cost, does the payback window still clear the threshold. If it does, keep spending and fix the cause in parallel. If not, pull back to the level where it clears.
Working with a womens fashion brand, we grew sales 99% while efficiency improved at the same time: new customer CPA down 21%, new-customer ROAS up 58% and marketing efficiency ratio down 6.7%. Net profit rose 136%. Net margin improved 119% and is still around 3%, which is the honest context.
If CAC has climbed for months and nobody has ruled out measurement, start there. A free thirty minute growth audit from Plaid Testing starts with exactly that check. Tracking first, then account structure, then three fixes you can act on immediately.
