ROAS vs MER vs CAC: Which Metric Should Run Your Ads?
ROAS, MER and CAC each answer a different question. Here is what each one measures, where it misleads, and how to combine all three into one decision system.
Ask three marketers which number matters most and you will get three answers. The media buyer says ROAS. The founder says MER. The finance lead says CAC. All three are right, because each metric answers a different question.
The trouble starts when you use one metric to answer a question it was never built for. This guide defines each one, shows the same business through all three lenses, and gives you a simple hierarchy for which number gets the final say.
The three metrics in one table
| Metric | Formula | Question it answers | Time horizon |
|---|---|---|---|
| ROAS | Attributed revenue / Ad spend | Which ads and campaigns are working? | Daily to weekly |
| MER | Total revenue / Total marketing spend | Is marketing profitable overall? | Weekly to monthly |
| CAC | Marketing spend / New customers acquired | What does a new customer cost? | Monthly to quarterly |
ROAS: the optimization metric
Return on ad spend divides the revenue a platform attributes to an ad, ad set or campaign by what you spent on it. If Meta says a campaign drove $8,000 in purchases on $2,000 of spend, platform ROAS is 4.0.
Where ROAS shines
ROAS is granular. You can compare two ad creatives, two audiences or two bid strategies within the same platform, because they share the same attribution model. That makes it the right tool for day-to-day decisions: which ad to pause, which to scale, which landing page to keep.
Where ROAS misleads
Each platform measures with its own rules and its own attribution windows. A shopper who sees a Meta ad on Monday, a ChatGPT chat card on Wednesday and clicks a Google search ad on Friday may be counted as a conversion by all three. Add their reported revenue together and you will exceed your actual sales.
ROAS also ignores margin. A 3.0 ROAS is excellent for a brand with 75% margins and a loss for one with 30% margins. Before you judge any ROAS, you need your break-even point. Our guide to break-even ROAS walks through that calculation.
MER: the business health metric
Marketing efficiency ratio divides all revenue by all marketing spend. It cannot be double-counted, because it does not care which platform claims the sale.
Where MER shines
MER tells you whether the whole system works. If MER holds at 4.0 while you add a new channel, the business is healthy even if attribution is fuzzy. It is the best metric for overall budget decisions: should you spend more next month or less?
Where MER misleads
MER includes revenue from repeat customers, email, organic search and word of mouth. A loyal customer base can keep MER looking healthy while your paid acquisition quietly deteriorates. And MER cannot tell you which ad or channel caused the change. For a deeper treatment, read how to calculate and use MER.
CAC: the unit economics metric
Customer acquisition cost divides marketing spend by the number of new customers you acquired. If you spent $30,000 last month and gained 600 new customers, CAC is $50.
Where CAC shines
CAC connects marketing to the long-term value of a customer. If an average customer generates $180 in gross profit over two years, a $50 CAC is very healthy, even if first-order ROAS looks thin. That is why subscription businesses, consumables and anything with strong repeat purchase should lean on CAC.
Where CAC misleads
Blended CAC hides channel differences. A $50 average might be $30 from one channel and $110 from another. CAC also says nothing about order value; a $50 CAC on a $40 first order is a very different business from a $50 CAC on a $200 order. And it depends on correctly identifying who is genuinely new, which requires clean customer data.
One business, three lenses
Say you run a coffee subscription brand. Last month:
- Total revenue: $150,000
- Revenue from first-time customers: $54,000
- Total marketing spend: $36,000
- New customers: 900
- Platform-reported revenue: Meta $70,000 on $22,000 spend, Google $38,000 on $9,000, ChatGPT ads $11,000 on $5,000
Here is what each metric says:
| Metric | Calculation | Result |
|---|---|---|
| Meta ROAS | $70,000 / $22,000 | 3.18 |
| Google ROAS | $38,000 / $9,000 | 4.22 |
| ChatGPT ROAS | $11,000 / $5,000 | 2.2 |
| Sum of platform claims | $70k + $38k + $11k | $119,000 |
| MER | $150,000 / $36,000 | 4.17 |
| New-customer MER | $54,000 / $36,000 | 1.5 |
| Blended CAC | $36,000 / 900 | $40 |
Notice what the platforms claim: $119,000 of revenue, against only $54,000 of first-time customer revenue. Some of those claims are repeat buyers clicking an ad on their way back, and some are overlap.
Now add customer value. Say an average subscriber stays six months and generates $22 of gross profit per month, or $132 over their lifetime. With a $40 CAC, the LTV to CAC ratio is 3.3, which most consumer brands would be very happy with. Our article on the LTV to CAC ratio explains how to calculate this properly.
If you only looked at ChatGPT's 2.2 ROAS, you might cut it. If you looked only at MER, you might scale everything. CAC plus LTV tells you there is room to spend more to acquire subscribers, and channel ROAS helps you decide where.
A decision hierarchy
Use each metric at the level where it is reliable:
- CAC and LTV set the ceiling. They tell you the maximum you can afford to pay for a new customer and still hit your profit goals. Revisit monthly or quarterly.
- MER sets the overall budget. If MER and new-customer MER are above target, you have room to grow spend. If they are below, cut. Review weekly and monthly.
- ROAS allocates within channels. Inside each platform, shift budget toward the ads and campaigns with stronger ROAS relative to that platform's own baseline. Review daily or every few days.
When the metrics disagree, the higher level wins. A campaign with great platform ROAS is not worth scaling if overall MER is falling and CAC is climbing past what LTV supports.
Setting targets for each
You can derive all three targets from the same margin math.
- Break-even CAC equals the gross profit you expect from a customer over the period you are willing to wait to recover the cost. If you want payback on the first order, it is first-order gross profit. If you accept 90-day payback, it is 90-day gross profit.
- Target MER equals 1 divided by the share of revenue you are willing to spend on marketing. Spending 25% means a target MER of 4.0.
- Target ROAS for a channel starts from break-even ROAS (1 divided by contribution margin percent), adjusted for how much that platform tends to over- or under-report against your MER.
The free CAC calculator and ROAS calculator make these quick to run with your own numbers.
Special case: new channels with weak attribution
Newer channels often look worse in platform ROAS than they really are, because tracking is less mature and many buyers convert later through other paths. ChatGPT ads are a current example. Someone might see your chat card while asking ChatGPT for a gift idea, then search your brand name on Google two days later. Google gets the credit.
When testing a channel like that, judge it on incremental MER and new-customer counts over several weeks, not on its own dashboard alone. Hold other spend steady, add the channel at a fixed budget, and watch what happens to total revenue and new customers. Our guide to attribution for AI channels covers this testing approach step by step.
Quick reference: which metric for which decision
| Decision | Primary metric | Supporting metric |
|---|---|---|
| Pause or scale an individual ad | Platform ROAS | CTR, conversion rate |
| Shift budget between platforms | Incremental MER test | Channel ROAS trend |
| Raise or cut total monthly budget | MER, new-customer MER | CAC |
| Decide how much to pay per customer | LTV to CAC | Payback period |
| Judge a brand-new channel | Incremental revenue and new customers | Platform ROAS |
| Report to investors or owners | MER and CAC | LTV |
Where SecondWin fits
SecondWin runs ChatGPT ads for consumer brands, and we encourage customers to judge the channel the way this article recommends: by new customers and incremental revenue, not just Ads Manager's reported ROAS. Campaigns run in your own OpenAI ad account with OpenAI's pixel connected, so you see every number yourself, and the ad spend stays on your card while our fee is a flat monthly rate shown on the pricing page.
If you want to see the buyer questions and proven messages we would test for your store, start with a free URL analysis of your brand.
FAQ
Is ROAS or MER more important?
They serve different jobs. ROAS is better for optimizing within a platform because it compares ads on the same attribution model. MER is better for overall budget decisions because it cannot be inflated by platforms double-counting the same sale. If the two disagree, trust MER for the question of whether to spend more overall, and use ROAS to decide where within a platform that spend should go.
What is the difference between CAC and cost per acquisition?
Cost per acquisition, or CPA, usually means the cost per conversion event a platform reports, which can include repeat buyers or non-purchase actions like signups. CAC specifically means the cost to acquire a genuinely new customer, ideally measured from your own customer data across all channels. CPA is a campaign metric; CAC is a business metric used alongside lifetime value.
Can ROAS be good while MER is bad?
Yes, and it happens often. If several platforms claim credit for the same sales, each can report healthy ROAS while total revenue barely moves relative to total spend. Retargeting campaigns are a common cause, since they reach people who were likely to buy anyway. When platform ROAS looks great but MER falls, reduce spend on the most overlapping campaigns and test incrementality.
Which metric should a small brand focus on first?
Start with MER and break-even ROAS. MER is easy to calculate from your store revenue and ad bills, and it keeps you honest about profitability. Break-even ROAS gives you a threshold for judging individual campaigns. Once you have a few months of customer data, add CAC and a simple LTV estimate so you can decide how aggressively to invest in acquisition.