How do I reduce customer acquisition cost?

Customer acquisition cost is a fraction: what you spent to win customers, divided by the number of first-time buyers in the same period. Most stores only ever touch the top of that fraction. They cut bids, pause campaigns, argue about creative costs, and the number barely moves.

The bottom of the fraction is where the room usually is. If two visitors in every hundred buy, then every improvement in that rate cuts acquisition cost by the same proportion, on traffic you have already paid for. A product page that answers the sizing question, a shipping cost shown before the cart, a checkout that does not force people to create an account: each of those adds buyers without adding spend.

So before you touch a budget, walk your own store on a phone, starting from an ad click and ending at the payment step. Note every place you had to guess, wait, or go hunting for information. Those moments are inflating your acquisition cost right now, and they are cheaper and faster to fix than a bidding strategy. This is also why we look at conversion before spend in almost every plan we write.

Not every product in your catalogue can afford the same cost per click, and blended CAC hides that. A store with a healthy average acquisition cost often has a group of products losing money on every order and another group carrying them.

The fix is to price the ceiling per product before you set bids. Take the gross margin on one order of that product, after the cost of goods, payment fees, packaging and the shipping you actually pay rather than the shipping you charge. That number is the most you can spend to win the order and still be level. Anything above it is a decision to buy the customer for later, which is fine if you mean it and expensive if you do not.

Once you have that, group the catalogue into a few margin bands and let the campaign structure follow the bands rather than the collection names. Products with room get the budget and the aggressive targets. Thin-margin products get a lower target, a bundle requirement, or no paid support at all. That single change often lowers CAC without lowering revenue, because the spend moves from products that could never pay it back to products that can.

Blended acquisition cost drops when part of your new-customer volume comes from places you are not billed per click for. Those customers still count in the denominator, but they add nothing to the top of the fraction.

The reliable sources are organic search on product and category pages, free product listings in Google's shopping surfaces, email and SMS to people who signed up and never bought, and referrals from customers you already have. None of them is free to build. All of them keep working after the work stops, which is the difference that matters over a year.

Start with the pages that are closest to a purchase. A category page that explains the choice between two similar products earns visitors who are already deciding. A product page written so an AI assistant can quote its specifications gets pulled into answers that a competing page cannot. A referral or review request in the post-purchase email costs one message.

Set the expectation honestly: this part is slow. Give it two or three quarters before judging it, and keep paid running while it builds. The point is not to replace ads. It is to stop paid from being the only way a stranger can find you.

Acquisition cost only means something next to what a customer is worth. Two stores with the same CAC can be in completely different positions, because one sells a consumable that is reordered every couple of months and the other sells something bought once.

So work out what a customer spends over their first year, not their first order. If the repeat rate rises, the amount you can afford to pay for a new customer rises with it, and you can outbid a store that only counts the first sale. That is a real competitive position, and it is built with post-purchase email, replenishment timing, subscribe-and-save on the products people rebuy, and a win-back that gives a reason rather than a discount.

There is a trap here worth naming. Raising your allowable CAC on the strength of a lifetime value you have not yet earned is how stores talk themselves into losses. Use the repeat behaviour you can already see in your order data, from customers who bought a year ago, not a projection. When that number genuinely improves, raise the ceiling then.

Measure acquisition cost on new customers only, by channel, and over a period long enough to mean something. Three details cause most of the confusion.

First, separate new from returning. A platform reporting cheap conversions is often being paid to reach people who would have come back anyway. Ad platforms can report new-customer conversions, and your ecommerce platform knows which orders were a customer's first. Compare the two.

Second, keep a blended number next to the channel numbers. Blended CAC is total marketing spend divided by all new customers, whatever the source. It is the honest one, because it cannot be improved by moving credit between reports. When blended CAC and channel CAC move in opposite directions, believe the blended number.

Third, choose a window that fits your buying cycle. A store selling a considered purchase will see clicks in one month and orders in the next, so a weekly CAC reading will swing wildly for no reason. Monthly is enough for most stores. Then set a target you actually act on: the ceiling from your margin work, not a number borrowed from someone else's store.