marketing
ROAS (Return on Ad Spend)
Revenue generated per dollar of paid advertising. Specifically a paid-media metric.
Definition
ROAS is revenue / ad spend. $5K ad spend producing $20K revenue = 4x ROAS. ROAS is the standard inside ad platforms (Google, Meta, LinkedIn) because those platforms can attribute clicks directly to conversions. Important: ROAS is gross revenue, not profit. A 4x ROAS on a product with 25% gross margin is breakeven. The healthy benchmark depends on the product: high-margin SaaS can scale at 2-3x ROAS, low-margin retail needs 6x+ to be profitable.
Platform ROAS versus true ROAS
The ROAS number Meta Ads Manager, Google Ads, or LinkedIn Campaign Manager displays is platform-attributed ROAS, and it is consistently inflated. Meta uses a 7-day click plus 1-day view attribution window by default, which double-counts customers who would have purchased anyway. Google Ads credits the last click in its own ecosystem and ignores LinkedIn or organic touches that did the real work. True ROAS pulls revenue from your source of truth (Shopify, Stripe, HubSpot deal closed-won) and divides by the spend across all paid channels. For US e-commerce brands at scale, true blended ROAS is typically 30 to 50 percent lower than the sum of platform ROAS. Build a weekly reconciliation between platform reports and your ledger before you trust any optimization decision.
ROAS targets by margin and stage
The right ROAS target depends on your gross margin and growth stage. A US e-commerce brand at 40 percent gross margin needs roughly 2.5x ROAS to break even on contribution margin, which means you should target 4x to 5x ROAS for healthy profit. A SaaS company with 80 percent gross margin can profitably scale at 1.5x to 2x first-month ROAS if LTV is high enough, because customer payback happens within 12 to 18 months. A B2B service firm running LinkedIn ads at a 60 percent margin should target 5x to 8x ROAS on the lead-to-customer revenue, not on the cost per lead. Map ROAS to LTV divided by CAC, not to a generic industry number.
Why ROAS drops as you scale
Every paid channel has diminishing returns. The first 10K per month on Meta hits your warmest lookalike audiences and best creative. The next 50K reaches cooler audiences, requires creative rotation, and ROAS slides 20 to 40 percent. Past 100K per month you are buying impressions from people who barely match your ICP, and ROAS can halve. The fix is not lower ROAS targets; it is channel diversification. Add Google Search, Performance Max, LinkedIn, TikTok, programmatic, or affiliate to maintain blended ROAS as you scale. Brands that try to scale a single channel past its efficient frontier always burn cash.
ROAS in HubSpot and Salesforce attribution
For B2B with long sales cycles, ad-platform ROAS is meaningless because the platform never sees the closed deal 90 days later. HubSpot Marketing Hub and Salesforce both support multi-touch attribution that ties ad clicks to closed-won opportunities, but only if you pipe UTM parameters cleanly through every ad, every landing page, and into the CRM contact record. Set up first-touch and last-touch attribution as separate reports, then run linear or W-shaped attribution monthly. True B2B ROAS often takes 60 to 120 days to read because revenue lags. Build a dashboard that shows pipeline-ROAS at 30 days and revenue-ROAS at 90 days side by side.
FAQ
What is a good ROAS for a Shopify store?
A profitable Shopify store typically targets 3x to 5x ROAS on prospecting campaigns and 6x to 10x on retargeting. The number depends on your gross margin: a 30 percent margin product needs 3.5x just to break even on ads alone, while a 70 percent margin product can profitably scale at 1.5x. Use this formula: minimum ROAS for profit equals one divided by gross margin percentage. So a 50 percent margin product needs at least 2x ROAS to cover ad cost.
Why does my Meta ROAS look great but my bank account is empty?
Three usual reasons. First, Meta attribution counts purchases from people who saw your ad but would have bought anyway (view-through and brand-search cannibalization). Second, you are likely not subtracting Shopify fees, COGS, shipping, and returns from the revenue Meta reports. Third, scaling spend rapidly produces revenue lag: you book the spend in month one but revenue and reorders unfold over months two through six. Reconcile platform ROAS against true contribution margin weekly.
Should I optimize ad campaigns for ROAS or for purchase volume?
It depends on your stage. Pre-1M revenue, optimize for purchase volume with a max-CPA cap, because you need data and ROAS-bidding is starved of conversions. Post-1M revenue and 50+ conversions per week per ad set, switch to target ROAS bidding in Meta and Google. ROAS bidding requires statistical signal that smaller accounts cannot produce reliably, and turning it on too early causes campaigns to under-deliver.
How is ROAS different from ROI?
ROAS is revenue divided by ad spend. ROI is profit divided by total investment including overhead, content production, and sales team cost. ROAS is a marketing efficiency metric; ROI is a financial return metric. A campaign with 5x ROAS and 30 percent gross margin produces 50 cents of gross profit per dollar of ad spend, which is 50 percent ROI before subtracting other costs. Always translate ROAS to ROI before making budget decisions.
Does Apple's iOS privacy update still affect ROAS reporting?
Yes, materially. iOS 14.5 plus and subsequent ATT changes cut Meta's deterministic attribution by an estimated 15 to 30 percent in US markets. Apple's SKAdNetwork and Meta's Aggregated Event Measurement provide partial signal, but most US D2C brands now run server-side tracking via Meta Conversions API and Google Enhanced Conversions to recover signal. Without server-side tracking, reported ROAS understates real performance by roughly 20 to 35 percent, which means you may be over-cutting profitable campaigns.
In your business
- →Translate ROAS into profit ROI before scaling - 4x ROAS at 25% margin is breakeven
- →Healthy benchmark for B2B service: 4-6x ROAS minimum
- →Track per ad set, not just per channel