ROAS vs MER vs CAC: Which Should Drive Bids

The ROAS column in your Google Ads dashboard says 4.8. Your finance team’s spreadsheet says you’re barely breaking even. Both are correct. That gap — the chasm between a platform metric and a business outcome — is where most paid-media programs quietly bleed money.

I’ve spent the last several years stitching together server-side pixels, CRMs, and media-mix models for brands spending anywhere from $30k to $4M per month on ads. The pattern is the same everywhere: teams pick a single number, hardcode it into a bid strategy, and then wonder why the P&L diverges from the dashboard six months later. The fix isn’t picking a “better” metric. It’s understanding what each metric can and cannot tell you — and matching the metric to the decision you’re actually making.

This is a working guide to the three numbers that should drive paid-media decisions: ROAS (return on ad spend), MER (marketing efficiency ratio), and CAC (customer acquisition cost). When to use which, what they hide, and how to wire them into bid strategy without lying to yourself.

Why Channel-Level ROAS Misleads at the Account Level

ROAS is the simplest definition in this whole conversation: revenue attributed to a campaign, divided by what you spent on that campaign. Spend $1,000, attribute $4,000 of revenue, ROAS is 4.0. Done.

The problem is the word attributed. Every platform — Google, Meta, TikTok, LinkedIn — runs its own attribution model. Meta credits view-through conversions for up to 7 days. Google uses data-driven attribution by default and pulls credit across multiple touchpoints. TikTok counts a click that happened 28 days before purchase. Add all of these reports together and you’ll find your “platform-attributed revenue” exceeds your actual revenue by 30% to 60%. I have seen it hit 90% on brands running aggressive prospecting across four platforms.

So when a Performance Max campaign reports ROAS 5.2, what that number really says is: “Of the orders we touched and could plausibly take credit for, our cost-to-revenue ratio looks like this.” It is not a claim about incremental revenue. It is not a P&L number. And it absolutely should not be the only input to a bid decision.

A common mistake I see: an account manager sees branded search ROAS at 18x and pushes more budget there. But branded search captures demand that already exists — most of those buyers would have arrived via organic if the ad weren’t running. The platform happily reports the conversion, the bid strategy happily scales spend, and the brand pays for clicks it would have gotten for free. Incremental ROAS on branded search routinely lands between 0.3 and 1.5 in geo holdout tests, not 18.

Channel-level ROAS is useful for one thing: comparing campaigns or ad sets within the same platform, on the same attribution window. It tells the bid algorithm which auctions are worth bidding harder on. That is where its usefulness ends.

MER — The Metric That Survives Attribution Loss

MER (marketing efficiency ratio), also called blended ROAS or eROAS depending on who you ask, sidesteps the attribution mess entirely. The formula is brutally simple:

MER = Total Revenue ÷ Total Marketing Spend

Total revenue means everything — paid, organic, email, direct, wholesale, retail. Total marketing spend means every dollar going out the door to acquire customers: ads, agency fees, influencer payments, affiliate commissions. No attribution model. No platform self-reporting. Just the two numbers your accountant already trusts.

The trade-off: MER tells you nothing about which channel is working. A 3.2 MER could be 80% organic plus a few inefficient paid campaigns, or a healthy paid program lifting weak organic. You need other tools (incrementality tests, media-mix models, geo holdouts) to decompose it.

What MER does well is anchor your business in reality. Common Thread Collective’s analysis of 150+ DTC brands found a clear pattern: brands growing 30%+ year-over-year ran median blended ROAS around 3.8, while plateaued or declining brands sat near 2.1. The number itself isn’t magic — your healthy MER depends on your gross margin and operating cost structure — but the principle holds: MER moves with the business, channel ROAS moves with the algorithm.

Practical rule of thumb for DTC: if your gross margin is 60%, an MER of 2.5 means you’re spending 40% of revenue on marketing, which leaves a thin sliver for fixed costs and profit. An MER of 4.0+ usually means you have headroom to invest more. If MER drops two weeks in a row while paid ROAS holds steady, your incremental ROAS is collapsing even though the dashboard looks fine. That divergence is your early warning system.

When CAC Beats Both (and What Payback Period Adds)

ROAS and MER are revenue-based ratios. They are blind to two things finance cares about: who the revenue came from (new vs. returning customer) and when the money actually arrives. Customer Acquisition Cost solves the first. Payback period solves the second.

CAC = Total Acquisition Spend ÷ Number of New Customers Acquired

For a subscription business, a B2B SaaS, or anything with high repeat purchase, CAC is the metric that matters. Revenue-based ROAS rewards you for re-targeting buyers you already had — which often produces beautiful dashboards and stagnant growth. CAC forces you to ask: “What did it cost to get a person who has never bought from us into the customer base?”

CAC payback period takes it one step further: how many months of gross profit from that customer does it take to repay the acquisition cost? Industry data across 35+ DTC engagements suggests roughly: food & beverage 1–3 months, beauty 2–4, pet 2–4, supplements 3–6, fashion 3–6, electronics 6–12+. The general benchmark is under 12 months, with under 6 months considered healthy for a cash-funded operation.

Why payback period matters more than raw CAC: a CAC of $80 looks identical on a spreadsheet whether the customer recovers it in 3 months or 18 months, but the cash-flow consequences are completely different. Bootstrapped brands need short payback. Funded brands can stretch it. Pick the wrong target and you either starve growth or run out of cash.

The Math: Same Campaign, Three Numbers, Three Decisions

Let’s run one month of a real-shaped DTC campaign through all three metrics. Brand sells a $90 supplement with a 70% gross margin. Repeat purchase rate is 45% within 6 months.

Ad spend (paid only): $50,000
Total marketing spend (ads + agency + affiliate): $65,000
Paid-attributed revenue (platform-reported, sum of Google + Meta): $185,000
Actual total revenue (Shopify): $210,000
New customers acquired: 920
Returning customer revenue: $42,000

Three calculations, three stories:

Channel-blended ROAS: $185,000 ÷ $50,000 = 3.7. Platforms claim a healthy return.
MER: $210,000 ÷ $65,000 = 3.23. Reality is tighter once you include agency fees and count only revenue that actually hit the bank.
New-customer CAC: ($50,000 ad spend, assume 90% targets acquisition) $45,000 ÷ 920 = $48.91. Gross profit per new customer first order: $90 × 70% = $63. Payback: 0.78 of an order — roughly 4 weeks if average reorder happens then.

The decisions diverge. A bid algorithm staring at ROAS 3.7 might pull back if your target is 4.0. The finance team looking at MER 3.23 sees a campaign that is acceptable but not great. The growth lead looking at $49 CAC with a fast payback sees room to push spend harder — because each new customer recovers cost within their first order and the back-end (returning revenue) is a free bonus on top. All three views are correct. Each drives a different action.

Bid Strategy Matched to Metric

Here is the comparison table I draw on whiteboards more than any other:

Dimension	ROAS (channel)	MER (blended)	CAC (new customer)
Signal type	Platform-attributed revenue ratio	Total revenue ÷ total spend	Acquisition cost per net-new buyer
Attribution dependency	High — fully tied to platform model	None — accounting-level	Low — needs customer-list dedup, not pixel
Best decision use	Within-platform bid adjustments, auction-level optimization	Budget allocation, profitability, board reporting	Channel mix, payback modeling, unit economics
When it lies to you	Cross-channel comparisons, branded vs. prospecting, iOS post-2021	When organic shifts mask paid performance change	When LTV varies wildly by acquisition source
Best-fit business	Mature programs on single platform	Multi-channel DTC, omnichannel retail	Subscription, SaaS, high-LTV consumables
Refresh cadence for decisions	Daily / per auction	Weekly	Monthly + cohort tracking

The bid-strategy mapping that has worked best for the brands I advise:

Target ROAS bidding (Google, Meta): Use it for catalog-driven campaigns where conversion values genuinely vary order to order. Google’s own docs recommend 15+ conversions in the prior 30 days; in practice, you want 30–50+ conversions per month for stable behavior. Set the ROAS target 15–25% below what you actually need at the business level — the algorithm will overshoot on the easy auctions and underperform on the hard ones, averaging close to your true target.
Target CPA bidding: Use it for lead-gen or fixed-price products. Calculate it as: (gross margin per sale × close rate × payback tolerance). For a $90 product at 70% margin with a 1-month payback rule, target CPA caps at roughly $63.
Max conversions with budget cap: Use it for new campaigns under 30 conversions/month, when you have less data than tROAS/tCPA need. Manage to MER and CAC at the program level instead.
Manual bidding: Use it for surgical control on tiny budgets ($5k/month and below) or in B2B niches where every conversion is a sales-qualified meeting and the algorithm has nothing to learn from.

The most common error is setting a Target ROAS that matches your break-even MER. They are not the same number. Break-even MER is a P&L truth. Target ROAS is a platform-attributed instruction. Because platforms over-credit themselves by 30–60%, a Target ROAS equal to break-even MER will starve volume. Reverse it the wrong way and you’ll over-spend.

Pitfalls When Your Sales Cycle Is Long

Everything above assumes a relatively short purchase cycle — days, not months. For B2B, considered-purchase consumer goods (mattresses, furniture, financial products), or SaaS, all three metrics have lag issues.

If your average sales cycle is 90 days, a campaign launched in March doesn’t show its true ROAS, MER, or CAC contribution until June at the earliest. By then, you’ve made three months of bid decisions on incomplete data. Two patterns help:

Optimize to lead-stage values, not final revenue. Assign a dollar value to each pipeline stage (MQL = $10, SQL = $80, demo-booked = $200) based on historical close rates. Feed those values back into Google or Meta as conversion values. The bid algorithm needs some revenue signal within its learning window. Made-up values that are directionally accurate beat true values that arrive 90 days late.
Track CAC by cohort, not by month. Group customers by acquisition month and track their cumulative revenue at 30, 60, 90, 180, 365 days. Compare cohorts to each other. This is the only honest way to know if your CAC is improving or just looking better because of seasonality.

For long-cycle businesses, MER becomes nearly useless as a steering metric — too much noise from organic and existing-pipeline revenue. CAC by cohort plus pipeline-velocity metrics are what actually work.

Common Mistakes That Wreck Budget Decisions

Summing ROAS across platforms. If Google reports 4.2 and Meta reports 3.8, your blended platform ROAS is not the weighted average — it’s whatever falls out when you stop double-counting the same buyer. Always trust MER over summed channel ROAS.
Including branded search in target-setting math. Branded clicks should usually live in their own campaign with their own efficiency standard. Mixing them with cold-prospecting ROAS makes both numbers meaningless.
Treating returning-customer revenue as ad-driven. Most retargeting and DPA campaigns hit existing customers. If you don’t exclude buyers from the prior 90 days, your ROAS inflates and your CAC vanishes — because you’re paying to re-acquire people who would have come back anyway.
Setting Target ROAS without testing the floor. Algorithms scale spend up to whatever ROAS you target. Set 6.0 and the algo finds 6.0 — but only by drastically reducing volume. Test 4.0, 5.0, 6.0 in 2-week increments and watch absolute new-customer count, not just efficiency.
Ignoring agency and tech fees in MER. If your CRM, ESP, attribution tool, and agency together cost $20k/month, leaving them out of “marketing spend” makes MER look 25% better than reality. Pretend you’re presenting to your CFO. Include everything.
Confusing efficiency improvements with growth. Cutting spend usually improves ROAS and MER on the surface — but only because you’re harvesting the most efficient demand and starving the top of funnel. Watch new-customer volume and CAC in parallel. Efficiency without growth is decay in slow motion.

FAQ

Is MER the same as blended ROAS?
They’re closely related but not identical. Blended ROAS typically divides paid-attributed revenue across all channels by total paid spend. MER divides all revenue — including organic, email, and direct — by all marketing spend including agency and tech fees. MER is always higher and closer to a P&L truth.

What’s a “good” MER?
There is no universal number. Your minimum MER is roughly 1 ÷ (gross margin − fixed cost ratio). For a 60% gross margin DTC brand with 20% fixed costs, break-even MER is around 2.5. Healthy growth typically requires 3.0–4.0+. Common Thread’s data on growing DTC brands clusters around 3.8 blended ROAS, suggesting MER often sits 4.5–5.5 for the same cohort.

Should I use Target ROAS or Target CPA in Google Ads?
Target ROAS when conversion values genuinely vary (ecommerce, marketplaces). Target CPA when leads are roughly interchangeable or product prices are fixed. Both require 15+ conversions per month minimum; 30–50+ for stable behavior. If you have less data, start with Maximize Conversions and graduate later.

How often should I recalculate CAC?
Monthly for the headline number, weekly for trend monitoring, and always by cohort for honest analysis. If your CAC drops 20% in a single month, check whether new-customer volume held — a drop in CAC with a drop in new customers usually means you stopped prospecting, not that you got more efficient.

Can I use a single metric for my whole business?
No. Use ROAS for within-platform bid optimization, MER for budget allocation and board reporting, CAC for unit-economics and channel-mix decisions. They answer different questions. Trying to compress them into one number is how brands end up with dashboards that look great while the bank balance drops.

Bottom Line

ROAS, MER, and CAC are not interchangeable. They sit at different altitudes of the business: ROAS at the auction, MER at the P&L, CAC at the unit economics. A mature paid-media program watches all three weekly and uses each one for the decision it’s actually equipped to answer.

If you’re running an ecommerce brand right now and you only have time to change one thing this week: stop steering your bid strategy with platform ROAS alone, and add a weekly MER calculation pulled directly from your accounting system. The first time MER and channel ROAS diverge by more than 0.5, you’ll find money you didn’t know you were losing — or, just as often, headroom to invest you didn’t know you had.

Both outcomes pay for the spreadsheet.