ROAS, MER and CAC explained: the complete guide

Each number answers a different question, so start by fixing the question rather than the formula.

Return on ad spend asks: for the money spent inside this campaign, how much revenue did the platform credit back. It is a campaign management number. Its strength is that it is available quickly and per campaign, so it can guide daily decisions. Its weakness is that it uses the platform's own view of which sales it caused, and every platform counts generously within its own walls.

Marketing efficiency ratio asks: for all the marketing money the business spent this period, how much total revenue did the business make. It ignores attribution entirely. Nothing can be double counted because nothing is being assigned. Its strength is that it cannot be gamed. Its weakness is that it tells you nothing about which channel to change.

Customer acquisition cost asks: what did it cost to win one new customer. It is the number that connects marketing to the value of a customer over time, and it is the only one of the three that treats a customer as something more than a single order.

The relationship between them is the useful part. Return on ad spend is the microscope, efficiency is the thermometer, and acquisition cost is the link between marketing and the rest of the business. A store that watches only the first will optimise campaigns while the business gets worse. A store that watches only the second knows something is wrong but not what. A store that watches only the third will not notice a campaign breaking for weeks.

Write the three definitions down for your own store, including what counts as marketing cost and what counts as revenue, because the rest of this guide depends on those two choices being stable.

Agree those definitions with anyone who reports on your marketing, including outside help. Most disputes about performance are definition disputes, and they are settled once rather than argued every month.

The formulas are simple. The decisions inside them are where stores go wrong, so make each decision consciously and then keep it.

Return on ad spend is the revenue attributed to a campaign divided by the spend on that campaign. Decide first whose revenue figure you use, since the platform and your analytics will disagree, and decide whether the revenue includes tax and delivery. Excluding both gives a cleaner number for judging profitability.

Marketing efficiency is total revenue divided by total marketing spend for the same period. The decision here is what counts as marketing spend. The honest version includes everything: ad spend across every platform, agency or freelancer fees, software subscriptions, creator and influencer payments, discount codes given to partners, and the cost of tools that exist only for marketing. A version that counts only ad spend will look better and will mislead you when you are deciding whether the whole function pays.

Acquisition cost is marketing spend aimed at winning new customers divided by the number of new customers won. Two decisions matter. Which spend counts, since money spent on retention email to existing customers is not acquisition. And who counts as new, which your store admin can tell you by first order date.

Use the same period, the same time zone and the same revenue definition for all three, then put them in one table together. Almost every argument about whether marketing is working turns out to be two people using different windows or different revenue definitions.

Worked example, so the arithmetic is clear. Say a store spends ten thousand dollars on ads in a month, another two thousand on fees and tools, and makes sixty thousand dollars in revenue with four hundred new customers. Efficiency is five dollars of revenue for each dollar of marketing. If eight thousand of the ad spend targeted new customers, acquisition cost is twenty five dollars, once the relevant fees are included.

A return on ad spend figure means nothing until you know what you needed. The same number can be excellent for one store and ruinous for another, and the difference is margin, not skill.

Start with contribution margin: for every dollar of revenue, how much is left after the cost of goods, the payment fee, the packing, the outbound delivery you pay and an allowance for returns. That fraction is what advertising has to be paid from.

Your break even return is one divided by that fraction. If forty cents of every revenue dollar survives to contribution, you break even when each advertising dollar brings in two dollars fifty of revenue. Below that, every sale costs you money. Above it, you are earning something.

That gives you a floor rather than a target. Your target has to leave room for the costs advertising does not pay for: rent, wages, software, the owner's income and whatever you want to reinvest. Decide what share of contribution advertising is allowed to consume, and set the target accordingly.

Do the calculation per product group, not once for the store. A store selling both a high margin accessory and a low margin machine has two very different break even points, and a single target will overspend on one and starve the other. This is also why campaign structure and margin should line up: it lets each group carry its own target.

And recalculate when your costs move. Carrier increases, supplier price rises and a shift in returns all change the floor. A target set two years ago against an older cost base can quietly turn a profitable account into an unprofitable one without any change in the advertising itself.

Write the break even and the target on the same page as your reporting, so nobody has to remember them.

Acquisition cost is the number most often calculated wrongly, and the errors all point the same way: they make acquisition look cheaper than it is.

The first error is dividing by orders rather than by new customers. If existing customers are ordering, the denominator is inflated and your acquisition cost looks low. Use first order date from your store admin to count genuinely new customers in the period.

The second is counting only ad spend. Fees, tools, creative production and creator payments are all part of what it costs to win a customer. Leaving them out understates the figure by a meaningful amount in most stores.

The third is including retention spend in the numerator while counting only new customers in the denominator, which is the reverse error and makes acquisition look expensive. Split your spend into acquisition and retention as best you can, accept that some is genuinely mixed, and be consistent about how you treat it.

Once you have an honest figure, look at it by channel and by first product. Both cuts change decisions. A channel with a higher acquisition cost that brings customers who buy repeatedly may be your best channel. An entry product that produces cheap first orders and no second order is expensive in the way that matters.

Then compare acquisition cost against what a customer is worth. The comparison should use contribution over a stated window, not revenue and not a lifetime projection. Contribution over twelve months against acquisition cost is a number you can act on, and it directly answers the only question that matters: can you afford to buy another customer.

Track the trend as much as the level. Rising acquisition cost with flat contribution is the early warning that a growth plan is running out of room, and it usually shows months before revenue reacts.

Blended efficiency is the number that cannot be argued with, which makes it the referee whenever the platforms and the analytics disagree. Total revenue for the month, total marketing cost for the month, one ratio.

Its value is that it includes everything and assigns nothing. If a platform is claiming sales that would have happened anyway, blended efficiency will not move when that platform's reported return improves. If a channel is genuinely adding customers, blended efficiency improves even when the platform under reports.

Use it as a level and as a trend. The level tells you whether the business can afford its marketing today. The trend tells you whether growth is getting more or less efficient, which is the real question when you are deciding whether to spend more.

Watch for the two ways it misleads. It flatters a store with lots of repeat business, because returning customers add revenue without adding marketing cost, so a store can show improving blended efficiency while new customer acquisition quietly fails. And it lags, because spend today produces revenue over following weeks, so a sudden increase in budget looks worse than it is for a month.

The fix for the first is to look at the blended number beside new customer count and acquisition cost. If efficiency is improving while new customers fall, you are harvesting rather than growing. The fix for the second is to compare quarters as well as months when spend has changed sharply.

Calculate it monthly, plot it over a rolling twelve months, and put it at the top of your reporting. When a platform report and your blended trend disagree about whether the last quarter went well, the blended trend is describing the business and the platform report is describing itself.

Keep the calculation simple enough that anyone can repeat it. A blended figure that depends on a spreadsheet only one person understands stops being the referee the moment that person is away.

A customer who repays their acquisition cost over two years is profitable and may still be unaffordable, because you have to pay for the advertising this month. Payback period is what turns marketing numbers into a plan the bank account can support.

Calculate it simply. Take contribution from the first order, then from each subsequent order in the months that follow, and find the point at which the cumulative total passes the acquisition cost. That point is your payback period, and it varies enormously between a business selling one large item and one selling a small consumable.

Use it to decide how aggressive you can be. If a customer pays back within the first order, you can spend as fast as demand allows. If payback takes several months, growth is limited by cash rather than by opportunity, and spending as though it were not is how growing stores run out of money.

It also settles the argument about whether a channel with a high acquisition cost is acceptable. If the customers from that channel repay quickly and keep buying, the cost is an investment with a known return. If they repay slowly, the same cost is a cash flow problem regardless of the eventual profit.

Segment it. Payback by first product, by channel and by whether the first order used a discount will usually differ enough to change what you promote. Customers acquired through a heavy discount often pay back more slowly, both because the first order contributes less and because they are more likely to wait for the next offer.

And hold the window fixed when comparing. Contribution to ninety days, or to twelve months, chosen once and used everywhere. A payback figure that quietly extends its window every time it is reported is a way of hiding a problem rather than measuring one.

Every one of these numbers can improve while the business gets worse, so they are only trustworthy as a set. Four common patterns show why.

High return on ad spend with falling revenue. This is usually a campaign that has narrowed onto people who were going to buy anyway, most often searches for your own brand and retargeting of existing customers. The efficiency looks excellent because the sales were already yours. The tell is a good campaign number beside a flat or falling new customer count.

Improving blended efficiency with rising acquisition cost. The repeat business is carrying the numbers while new customer acquisition deteriorates. It can continue for a long time and then stops suddenly, because the customer base you are harvesting was bought earlier.

Falling acquisition cost with falling order value. You are buying cheaper customers who buy less, which can be fine or can be the beginning of a slide towards a discount driven audience. Watch order value and second order rate for the customers from each channel.

All three numbers steady while contribution falls. Costs have moved, not marketing. Supplier prices, carrier rates or returns have changed the margin under a target that has not been updated.

The practical answer is one small table, monthly, with return by main channel, blended efficiency, acquisition cost, new customer count, average order value and contribution. Six numbers side by side, twelve months of history. Nearly every important marketing question is answerable from that table, and it takes an hour a month to maintain once it exists.

When two of the six disagree about how the month went, the disagreement is the finding. Go and explain it before making any decision about budget.

Add a short written note each month explaining what the table shows. A number without a sentence beside it gets reinterpreted every quarter, usually to fit whatever decision somebody already wanted to make.

There is no correct return on ad spend, and a target borrowed from somewhere else is worse than no target because it feels authoritative. Your target comes from your margin, your growth appetite and your cash position, and nobody outside your business can calculate it.

Build it in four steps. Start at break even, from your contribution margin. Add the overhead that advertising has to help pay for, expressed as a share of contribution. Adjust for growth: a business willing to accept slimmer returns in exchange for more customers sets a lower target, provided payback and cash allow it. Then adjust for cash, because a slow payback period means the target has to be higher no matter what the long term maths says.

Set different targets for different jobs. Campaigns aimed at people who already know you should return more than campaigns reaching strangers, because they are cheaper sales to make. Holding both to the same target will push all your spend to the easy audiences and stop the business growing. Making the prospecting target explicitly lower is how you decide to buy new customers on purpose rather than by accident.

Set an acquisition cost ceiling too, tied to contribution over your chosen window. That ceiling is the sentence that stops a good month turning into an expensive one: we will pay up to this much for a new customer, and above it we stop.

Review targets quarterly and whenever costs change. A target is a statement about your economics, and your economics move.

And write the targets down where the person managing the spend can see them. Numbers that live in someone's head get reinterpreted every month, usually in whichever direction the results happened to go.

Give the targets a review date rather than an indefinite life. A target with no expiry becomes a rule, and rules built on last year margin are how a profitable account turns unprofitable without anyone changing anything.

The value of these numbers is in the actions they trigger, so write the actions down before the month when you need them. Decisions made calmly in advance are better than decisions made in the week revenue drops.

When campaign return falls, check the boring causes first: a feed or tracking problem, a product out of stock, a competitor promotion, a landing page change, or a seasonal shift that happens every year. Only after those should you change bids or budgets. Most emergency budget cuts are made in response to a broken tag.

When blended efficiency falls while campaign returns hold, look for spend that is not in the ad platforms, or revenue that has fallen for a reason unrelated to marketing, such as stock. This pattern often means the platforms are claiming a larger share of a shrinking total.

When acquisition cost rises steadily, treat it as a structural signal rather than a campaign problem. The usual causes are audience saturation, tired creative, or a channel that has been pushed past its efficient volume. The responses are new creative, new audiences or a new channel, none of which are quick, which is why a rising trend needs to be caught early.

When everything looks fine but contribution falls, go to the cost side. Recalculate margin, check carrier invoices and returns, and reset the targets.

Set thresholds that trigger a review rather than an immediate change, because acting on a single week of data usually makes things worse. A campaign needs enough conversions to be judged, and daily changes prevent any of them from being measurable.

Then record what you did and what followed. Over a year that log becomes the most valuable marketing document you have, because it is the only record of which responses actually worked in your store.