...
...
We help businesses scale with data-driven strategies and world-class design.
Get Started rocket_launchReturn on Investment (ROI) and Return on Ad Spend (ROAS) are the two most commonly used numbers to judge whether Google Ads and Meta Ads spend is actually working — but they answer different questions, and mixing them up leads to genuinely wrong decisions.
ROAS tells you how much revenue came back for every rupee spent on ads. A ROAS of 4 means ₹4 in revenue for every ₹1 spent. It does not account for the cost of goods, overheads, or any other business expense — it is a pure advertising-efficiency number.
ROI is a profit-based number. “Total Cost” includes ad spend plus the cost of delivering the product or service — not just the media spend. A campaign can have a strong ROAS and still lose money once real costs are included, which is exactly why relying on ROAS alone is risky for any business with meaningful cost of goods or service delivery costs.
| ROAS | ROI | |
|---|---|---|
| Measures | Revenue per ad rupee | Profit after all costs |
| Accounts for product/service cost | No | Yes |
| Best for | Comparing campaigns/channels quickly | Understanding real business impact |
| Risk if used alone | Can look profitable while losing money | Requires accurate cost data to be reliable |
A campaign generates ₹1,00,000 in revenue from ₹20,000 in ad spend — a ROAS of 5. That looks excellent. But if the cost of delivering that revenue (product cost, fulfilment, service delivery) is ₹75,000, the actual profit is ₹5,000 on a ₹95,000 total cost — an ROI of roughly 5%, not the impressive number ROAS alone suggests. Both numbers are useful together: ROAS for fast campaign-level comparisons, ROI for the real business decision.
ROI tracking is only as good as the conversion data feeding it. For Google Ads, this starts with conversion actions configured correctly for what the business actually needs to measure.
Without this level of specificity, Google Ads’ own reported conversion value — and therefore any ROAS/ROI calculated from it — will be systematically inaccurate.
Meta Ads tracking has become materially less reliable through the Pixel alone since iOS privacy changes reduced browser-based tracking accuracy. The Conversions API (CAPI) sends conversion events directly from a business’s server to Meta, bypassing browser-level blocking.
The Pixel and CAPI are meant to work together, not replace each other. Running both, with Meta’s deduplication (via a shared event ID) enabled, produces the most complete picture of which conversions are genuinely attributable to Meta Ads.
How ROI is calculated genuinely differs by business model — treating both the same way is a common source of misleading numbers.
Revenue is usually directly measurable at the transaction level. The main risks are attribution windows that are too generous (crediting ads for purchases that would have happened anyway) and failing to exclude returns/refunds from reported revenue.
Revenue is not known at the moment of conversion — a form submission is not revenue, it is a lead that may or may not close. This means lead-gen ROI tracking depends on connecting ad platform data to what actually happens after the lead is generated, which requires CRM integration (covered next) rather than relying on the ad platform’s own conversion value alone.
For most lead-generation businesses, the single biggest accuracy gap in ROI tracking is that the ad platform never learns what happened to a lead after it entered the CRM.
This closes the loop between ad spend and real revenue, rather than stopping at “form submitted.” Without it, ROI reporting for a lead-gen business is really just a lead-volume report wearing an ROI label.
UTM parameters and GA4 fill a specific gap: understanding the full customer journey across sessions and channels, which neither Google Ads nor Meta Ads reporting shows on its own.
Consistent UTM tagging lets GA4 correctly attribute revenue to specific campaigns even when a customer’s path involves multiple sessions or devices — something platform-level reporting alone often misses or attributes inconsistently.
GA4’s data-driven attribution model distributes credit across multiple touchpoints rather than giving 100% credit to the last click. For businesses running both Google Ads and Meta Ads simultaneously, this is often the only place a genuinely cross-channel view of what’s driving revenue actually exists.
Example: if 1 in 5 leads closes (20% close rate) and the average deal is worth ₹50,000, each lead is worth approximately ₹10,000 — not ₹50,000. Using the deal value instead of the calculated lead value is one of the most common ROI-inflating mistakes in lead-gen accounts.
CAC should always be compared against customer lifetime value, not just the value of the first transaction — a channel with a high CAC can still be genuinely profitable if customer lifetime value is high enough.
A business spends ₹50,000 across Google Ads and Meta Ads in a month and generates 40 leads.
Both numbers are genuinely useful here — the ROAS of 3.0 confirms the ads themselves are efficient, and the 67% ROI confirms the campaign is actually profitable once real delivery costs are included.
A reliable ROI report, reviewed monthly, should separate platform-reported numbers from CRM-verified outcomes:
Reviewing both sets of numbers side by side, rather than trusting either one in isolation, is what actually lets a business tell the difference between an ads problem and a tracking problem.
Neither alone is sufficient. ROAS is useful for quick campaign-level comparisons; ROI is what actually reflects business profitability. Track both.
It depends entirely on gross margin. A business with high margins can be profitable at a ROAS of 2; a low-margin business may need a ROAS of 6+ to be profitable. There is no universal benchmark.
For lead-generation businesses, yes — without it, ad-platform conversion data stops at “lead,” not revenue. E-commerce businesses can often track revenue directly through the platform or GA4.
Daily or weekly is typical. Longer gaps mean the ad platforms’ automated bidding is optimizing against stale outcome data.
