ROAS stands for Return On Ad Spend. Most brands track it religiously. Most agencies optimize for it relentlessly. Most board decks lead with it.
Most of those brands are also quietly losing money on otherwise "profitable" accounts. Their ROAS is fine. Their cash flow is bleeding. The metric is wrong.
This post is why ROAS is misleading, what it leaves out, and which number to track instead.
The simple ROAS math.
ROAS = revenue from ads / spend on ads. If you spent Rs 1 lakh and got Rs 4 lakh back, your ROAS is 4x. Easy.
That math is correct. The problem is what it does not include.
ROAS does not include cost of goods sold. ROAS does not include shipping. ROAS does not include refunds. ROAS does not include payment processing fees. ROAS does not include the salaries of the people fulfilling, supporting, and onboarding those customers. ROAS does not include time-to-payment for B2B accounts.
A 4x ROAS account can be deeply unprofitable. A 1.7x ROAS account can be a cash machine. The metric is structurally incomplete.
The accounts that look great and are bleeding.
We have audited 350+ accounts. The pattern repeats. A D2C brand reports 4.7x ROAS on Meta. CFO is happy. Bank balance shows revenue dropping every month.
The breakdown:
Revenue: Rs 14,73,200 from ads. Reported ROAS: 4.7x.
COGS: 47 percent of revenue = Rs 6,92,404.
Shipping: 7 percent = Rs 1,03,124.
Refunds and returns: 11 percent = Rs 1,62,052.
Payment fees: 2.4 percent = Rs 35,357.
Customer support: Rs 87,000 fixed.
Fulfillment ops: Rs 124,000 fixed.
Net contribution: Rs 1,32,663 on Rs 3,13,447 in ad spend. The actual blended return on the ads is 1.42x. Not 4.7x.
The CFO sees 4.7x. The bank account sees 1.42x. The bank account is correct. The dashboard is lying because the dashboard does not look at the operations stack.
The one number that matters: contribution margin per acquisition.
Skip ROAS. Track contribution margin per acquisition (CMPA).
CMPA = (revenue per customer) - (variable costs per customer) - (CAC).
For a D2C brand: revenue minus COGS minus shipping minus refunds minus payment fees, then minus the cost to acquire that customer. Whatever is left is what actually pays for fixed costs and profit.
For a SaaS brand: monthly revenue per customer x average contract length minus support costs minus payment fees minus CAC. Watch this number. Optimize ads against this number.
When you optimize against CMPA, your "best" channels often change. The cheap CPL channel that drove a 4.7x ROAS may produce a CMPA of Rs 47. The expensive CPL channel that drove a 1.7x ROAS may produce a CMPA of Rs 1,847. The expensive channel is 39x more profitable per acquisition. Your dashboard never told you.
CAC payback period.
The second metric that beats ROAS is CAC payback period. How many days from spending the marketing dollar to getting it back as cash from the customer.
A B2B SaaS that has 14-month CAC payback can scale efficiently. A B2B SaaS that has 24-month CAC payback cannot, because it runs out of cash before payback compounds.
A D2C brand that has 30-day CAC payback has a license to print money. A D2C brand with 90-day payback is a cash-burning machine even if its ROAS looks great.
The brands that get this run their ads against payback period, not ROAS. They optimize for the channel that pays back in 47 days, not the channel that has the highest ROAS over 12 months.
Why agencies will not switch.
Agencies are paid on retainer or percentage of spend. Both models reward optimization for ROAS, not for CMPA. ROAS is reportable. CMPA requires the agency to know your COGS, your support costs, your refund rate. They do not have that data. They do not want that data. The reporting friction is real.
This is structural, not malicious. The agency model is built around ad-platform metrics. Your operating reality is built around bank-account metrics. The two will never line up unless you push them to.
The 3-question audit.
Run this on your own account this week. Pull last 90 days of paid acquisition. Then answer:
- What is my contribution margin on the average new customer acquired through paid ads? Not revenue. Margin.
- How many days from ad spend to first dollar of cash from that customer? Not "first sale." Cash.
- If I doubled my paid spend tomorrow, what would happen to my cash position over the next 90 days?
If you cannot answer all three within 17 minutes, your dashboard is lying to you. The lie is not malicious. It is just incomplete.
The Prudent example.
Prudent is a B2B cybersecurity firm in our case study vault. Their LinkedIn CPL is $93. Their Meta CPL would be roughly $19 to $24. ROAS comparison favors Meta by approximately 4x.
CMPA tells the opposite story. LinkedIn CMPA: $245,000 - $4,371 CAC = $240,629. Meta CMPA: $14,000 - $4,693 CAC = $9,307.
LinkedIn is 25x more profitable per acquisition. Their ROAS-optimized agency would have killed LinkedIn 18 months ago. The CMPA-optimized agency tripled it. Same business, same ad spend, opposite outcomes.
What you do.
Stop reporting ROAS at the company level. Track CMPA per channel. Track CAC payback period per channel. Track these monthly. The first time you do, you will likely discover that one of your "best" channels is actually losing you money and one of your "expensive" channels is the one printing.
The free 30-minute audit walks through this calculation live on your account. Most operators leave the audit knowing exactly which channel to scale and which to kill. About 1 in 4 walk away saying "I cannot believe we have been optimizing the wrong number for 14 months." The other 3 in 4 say it but more politely.
