Every couple of weeks, someone asks me some version of this question: “Should we be looking at ROAS or CPA?” And the honest answer is that it depends — but not in the wishy-washy way people usually mean when they say that. There are clear situations where one metric is vastly more useful than the other, and picking the wrong north star metric can actively hurt your business.
Let me walk through the difference, when each one matters, and the traps I’ve watched smart teams fall into by optimizing for the wrong number.
Quick Definitions
CPA (Cost Per Acquisition): The amount you spend on advertising to acquire one customer, lead, or conversion. If you spent $5,000 on ads and got 50 leads, your CPA is $100.
ROAS (Return on Ad Spend): The revenue generated for every dollar spent on ads. If you spent $5,000 on ads and generated $25,000 in revenue, your ROAS is 5x (or 500%).
Simple enough. But the implications of choosing one over the other as your primary metric are not simple at all.
When CPA Is the Right Metric
CPA makes sense when all your conversions have roughly the same value, or when you’re tracking an action that happens before a sale.
B2B lead generation is the classic example. You’re generating demo requests or consultation bookings. Each lead goes into a sales process, and the revenue comes later — sometimes much later. You don’t have revenue data at the ad platform level, so ROAS isn’t calculable in real time. CPA is what you work with.
If your sales team tells you that a qualified lead is worth approximately $2,000 in eventual revenue, and you close 20% of leads, then each lead is worth about $400 to you. Set your target CPA well below that, account for sales costs, and you have a working model.
CPA also works well for subscription businesses in the early stages where you don’t yet have reliable LTV data. If you know your average customer stays for 8 months at $99/month, that’s roughly $792 in LTV. Your CPA target should be a fraction of that — how much of a fraction depends on your margins and growth ambitions.
The limitation of CPA: it treats every conversion equally. A $500 deal and a $50,000 deal both count as one conversion. If your deal sizes vary dramatically, CPA alone will mislead you.
When ROAS Is the Right Metric
ROAS shines when you can tie revenue directly back to ad spend, and when your transaction values vary.
Ecommerce is the obvious case. If someone clicks your ad and buys a $30 t-shirt, that’s different from someone who clicks and buys a $500 jacket. With CPA, both look the same — one conversion each. With ROAS, you can see that the jacket campaign is producing $500 in revenue per conversion while the t-shirt campaign produces $30. That changes how you allocate budget.
ROAS is also useful for direct-to-consumer brands with online checkout, where purchase data flows directly back to the ad platform via conversion tracking and pixel data.
The limitation of ROAS: it ignores profitability. A 4x ROAS sounds great until you realize your product margins are 20%. That means for every $1 of ad spend, you generate $4 in revenue but only $0.80 in gross profit. You’re losing $0.20 on every sale. This is more common than people admit.
The Trap of Optimizing Purely for ROAS
I’ve seen this play out painfully. A D2C brand sets a ROAS target of 4x. The media buyer hits it consistently. Everyone celebrates. But the business is barely profitable — or worse, losing money — because nobody accounted for cost of goods, shipping, returns, and payment processing fees.
A better version of ROAS is what some people call MER (Marketing Efficiency Ratio) or nROAS (Net ROAS). Instead of raw revenue, you use gross profit (or contribution margin) as the numerator. If you sell a $100 product with $40 in gross profit, and you spent $20 to acquire that sale, your effective ROAS on margin is 2x ($40 / $20), not 5x ($100 / $20).
That’s a much more honest number. It tells you whether the advertising is actually creating value after costs.
The Trap of Optimizing Purely for CPA
CPA has its own pitfalls. The biggest one: teams become so fixated on driving CPA down that they sacrifice volume and quality.
I watched a team bring their CPA from $200 down to $80 and congratulate themselves for months. The problem? Lead volume dropped by 70%, and the leads that did come in were mostly unqualified. Sales close rates fell from 22% to 6%. The effective cost per closed deal actually went up, from $909 to $1,333.
Lower CPA is not inherently better. A $200 CPA on leads that close at 25% is wildly more valuable than a $80 CPA on leads that close at 3%.
This is why CPA should always be paired with a quality metric. For B2B, that might be SQL rate (what percentage of leads become sales-qualified). For ecommerce, it might be repeat purchase rate or average order value.
What About CAC?
People sometimes confuse CPA with CAC (Customer Acquisition Cost). They’re related but different.
CPA is typically just ad spend divided by conversions. It’s a channel-level metric.
CAC is all-in: ad spend, plus sales team salaries, plus tools, plus creative costs, plus agency fees, divided by new customers acquired. It’s a business-level metric.
A $100 CPA might translate to a $350 CAC once you factor in everything else. If you’re only looking at CPA and ignoring the full acquisition cost, you might think you’re profitable when you’re not.
For board meetings and financial planning, use CAC. For day-to-day campaign optimization, use CPA (or ROAS). Just make sure the relationship between the two is understood.
Our Approach: The Blended View
At ROI Sphere, we don’t pick one metric and ignore the other. Here’s the framework we typically use:
For B2B clients: Primary metric is CPA (cost per qualified lead). Secondary metrics are SQL rate, pipeline value generated, and eventually revenue closed. We set CPA targets based on the client’s sales data — what a qualified lead is worth given their close rate and average deal size.
For ecommerce clients: Primary metric is blended ROAS (across all paid channels). Secondary metrics are contribution margin per order, new customer acquisition cost, and customer LTV. We push clients to think in terms of profit ROAS, not just revenue ROAS.
For both: We track efficiency at the channel level (CPA or ROAS by channel) but evaluate success at the business level. Google might have a $150 CPA and Meta might have a $200 CPA, but if Meta’s leads close at 2x the rate, Meta is actually the better channel despite the higher CPA.
This is the kind of nuance that gets lost when you fixate on a single number.
A Practical Recommendation
If you’re not sure where to start, here’s what I’d suggest:
If your transaction values are roughly uniform (B2B lead gen, SaaS signups, single-product businesses), start with CPA as your primary metric. It’s simpler and requires less data infrastructure.
If your transaction values vary significantly (ecommerce with many products, multi-tier pricing), go with ROAS but calculate it on contribution margin, not revenue.
In either case, check your numbers against actual business outcomes monthly. Campaign-level metrics are only useful if they correlate with real profit. If your CPA looks great but the business isn’t growing, something is disconnected — and finding that disconnect is more important than hitting any particular metric target.
Not sure if your campaigns are optimizing for the right metric? Book a free strategy call and we’ll help you figure out what to measure and why.