What is a good customer acquisition cost?

Acquisition cost has no correct value on its own. It only means something next to two things: what a customer is worth to you, and how long you can wait for that money.

Start with gross profit rather than revenue. If a first order leaves 30 dollars of profit and a typical customer buys twice in their first year, the first year is worth about 60 dollars of profit. Spending 50 dollars to acquire that customer is viable in principle and awful in practice if you need the cash back this quarter, because most of the return arrives months later.

So the practical question is the payback window. How many months of a customer's purchases does it take to recover the acquisition cost. A store funded by its own cash usually needs that inside the first order or two. A store with funding or with a strong subscription base can wait longer and outbid the first store for the same customer.

That is why two competitors can pay very different amounts for the same customer and both be right.

Write down the window you can actually survive, then work backwards to the cost you can pay. That number is your target, and it will be specific to your store, your margin and your bank balance.

The most common error is dividing total spend by total orders. That mixes repeat orders into the denominator and produces a figure far lower than the truth.

Count new customers only. Take the marketing spend aimed at acquiring them and divide by the number of first time buyers in the period. Every store platform can tell you how many customers placed a first order.

Decide what goes in the numerator and keep it consistent. At minimum, paid media. Most stores should also include the fees for running that media and any creative production that scales with it, because those costs are part of what acquisition really takes. Tools and salaries can be included or excluded, so long as you never change your mind halfway through a year.

Be careful with attribution here as well. Assigning new customers to specific channels using platform reporting will double count, because two platforms will both claim the same person. For channel decisions, use the platform numbers knowingly. For the business figure, use total acquisition spend over total new customers, which cannot be double counted.

Then segment by channel where you can, because a blended figure hides the channel that is quietly expensive and the one that is cheap enough to grow into.

The comparison that matters is acquisition cost against the gross profit a customer actually produces in a defined window, usually twelve months.

Build it from your own cohorts rather than a projection. Take customers who first bought a year ago. Add up everything they have spent since, take out the variable costs, and divide by the number of customers. That is a real first year value, and it will differ from any model you build.

Then compare it with what you paid to acquire that same cohort. If the value comfortably exceeds the cost, you have room to spend more. If it barely covers it, acquisition is not funding the business, it is renting revenue.

Segment the comparison in two ways that usually change decisions. By channel, because some sources bring customers who return and some bring customers who never do, and the cheap channel is often the worse one on this measure. And by first product, because the item someone starts with predicts how likely they are to come back. Stores frequently find one entry product that produces loyal customers and another that produces one time buyers, which changes what they advertise.

This is also how you decide whether to accept a higher cost. Not because lifetime value theoretically justifies it, but because the cohort from last year proves it.

Four routes, in the order they usually pay.

Raise the conversion rate. Acquisition cost is what you pay for traffic divided by the share that buys. Improving the second number lowers the cost without touching the auction, and the improvement applies to every channel at once.

Raise the value of the first order. A larger first purchase brings the payback window forward, which is often more useful than a lower cost. A sensible bundle or a shipping threshold does this directly.

Bring the second order closer. If customers return in eight weeks rather than six months, the same acquisition cost is affordable at a much lower risk. This is retention work, and it changes what you can pay for traffic.

Then fix the media. Cut the campaigns and products that consistently deliver expensive first orders, tighten the matching so you stop paying for searches that never convert, and move budget towards the sources whose cohorts come back.

If all four are done and the cost still exceeds what a customer is worth, the honest conclusion is that the price, the margin or the product mix cannot support paid acquisition at that scale. That is a business conversation rather than an advertising one, and having it early is cheaper than having it in a year.