71 months of Google Ads data. January 2021 through November 2026. Every conversion action change, every budget shift, every ROAS inflection point documented and visualized.
Vegamour's Google Ads account has had multiple conversion actions active simultaneously throughout its history. Understanding which actions were PRIMARY at any given time is essential for interpreting ROAS. The "conversions_value" metric (used for ROAS calculation) only counts PRIMARY conversion actions.
Era 1: Clean Purchase Tracking
Primary actions: "Website Purchases" (Pixel), "Shopify Order" (Littledata), "Subscription First Order" (Littledata), "New subscriptions". Also tracking "Shopping Direct Purchase - Google" (Google Shopping native, $0 value). ROAS in this era reflects actual purchase revenue accurately.
Era 2: Multiple Overlapping Purchase Actions
Multiple conversion actions counting the SAME purchase differently: "Website Purchases", "Shopping Direct Purchase", "New subscriptions", "vegaLASH - GA4 (web) purchase", "Transactions (Littledata)", "vega Sales", "Subscription". The all_conversions count ballooned (e.g., May 2023: 807K all_conversions vs 17K primary conversions) due to add-to-cart events and micro-conversions being tracked. ROAS from conversions_value remained relatively clean as it only counts PRIMARY actions, but conversion COUNT was massively inflated.
Era 2b: Add-to-Cart Event Explosion
The all_conversions metric spiked dramatically. May 2023 reported 807K all_conversions and Jun 2023 reported 885K. This was caused by add-to-cart, page view, and micro-conversion events being counted as conversion actions. The PRIMARY conversions (17K and 15K respectively) remained stable, confirming the spike was non-purchase events. If anyone used "all conversions" for ROAS during this period, numbers would be meaningless.
Era 3: Melleka Takes Over, Proper Purchase Tracking
Active PRIMARY conversion actions: "Website Purchases", "vega Sales", "Subscription", "Purchase - Straight Sale No Subs". ROAS calculated from conversions_value is the correct methodology. The correct ROAS uses conversions_value directly from the API (confirmed May 2026). Budget scaled down from ~$1.4M/mo to ~$250K-$310K/mo. Focus shifted to efficiency over volume.
Era 3b: Corrected Methodology Confirmed
After discovering the ROAS calculation error (May 5, 2026), all reports now use conversions_value directly. "Purchase - Straight Sale No Subs" was confirmed as a PRIMARY action that was previously excluded from manual calculations. All three purchase actions (vega Sales + Subscription + Straight Sale) are now properly counted.
PRIMARY conversions only. Hover for monthly detail. Shaded regions indicate conversion methodology eras.
Area chart showing the gap between ad spend (cost) and attributed revenue (conversions_value). Green = profitable month, Red = unprofitable.
Cost per conversion (Spend / Primary Conversions). Lower is better. Note: conversion count includes all PRIMARY actions, not just new customers.
Compare ROAS performance across quarters. Each bar group represents a quarter. Includes year-over-year comparison.
| Year | Total Spend | Conv. Revenue | ROAS | Conversions | All Conversions | Clicks | Impressions | Avg CPC | NCAC | Conversion Era |
|---|
| Month | Spend | Revenue | ROAS | Conv. | All Conv. | Clicks | Impr. | CTR | Avg CPC | NCAC | Flags |
|---|
All conversion actions found in the account, their type, category, status, and whether they were ever PRIMARY (counted in ROAS).
| Action Name | Type | Category | Status | Primary | Impact on ROAS |
|---|---|---|---|---|---|
| Website Purchases | Pixel | Purchase | ENABLED | Yes | Core purchase action, always primary. Main ROAS driver. |
| vega Sales | Pixel | Purchase | ENABLED | Yes | Primary purchase action. Combined with Subscription for total revenue. |
| Subscription | Pixel | Purchase | ENABLED | Yes | Subscription purchases. Paired with vega Sales. |
| Purchase - Straight Sale No Subs | Pixel | Purchase | ENABLED | Yes | Was MISSED in early reports. Confirmed PRIMARY May 2026. Excluding this caused ROAS undercount. |
| Shopping Direct Purchase - Google | Native | Purchase | HIDDEN | Yes | Google Shopping native tracking. Reports $0 value but counts conversions. Inflates conv count without affecting revenue. |
| New subscriptions | Pixel | Purchase | HIDDEN | Yes | Subscription-specific. Carries revenue value. Active 2021-2024. |
| Shopify Order (Littledata) | UA Transaction | Purchase | REMOVED | Yes | Early tracking via Littledata integration. Removed when GA4 took over. |
| Subscription First Order (Littledata) | UA Transaction | Purchase | REMOVED | Yes | Littledata subscription tracking. Removed with UA sunset. |
| vegaLASH - GA4 (web) purchase | GA4 Custom | Purchase | ENABLED | Yes | GA4 purchase event. Potential double-counting with pixel-based purchases. |
| vegaLASH - GA4 (web) addToCart | GA4 Custom | Add to Cart | REMOVED | No | NOT primary. Would inflate conv count massively if ever set to primary. |
| vegaLASH - GA4 (web) add_shipping_info | GA4 Custom | Micro-event | ENABLED | No | Not primary. Observation only. |
| vegaLASH - GA4 (web) add_payment_info | GA4 Custom | Micro-event | ENABLED | No | Not primary. Observation only. |
| Calls from ads | Click-to-Call | Default | ENABLED | Yes | Phone calls, minimal count. Reports $0 value. |
| Begin checkout | Pixel | Default | ENABLED | No | Not primary. Checkout initiation tracking only. |
Meta's reported purchase ROAS of ~0.33x appears extremely low. This is likely because Meta's attribution model (7-day click, 1-day view by default) captures a smaller window of conversions compared to the actual customer journey. Many Vegamour customers research extensively before purchasing. The true blended ROAS including post-view conversions, brand halo effects, and multi-touch attribution is almost certainly higher. Meta's primary role in Vegamour's funnel is top-of-funnel awareness and demand generation, not direct last-click conversion.
Monthly spend, purchase revenue, and ROAS from Meta Ads. Revenue = purchase action_values only.
Microsoft Advertising performance since account activation in 2024.
How Vegamour allocates budget across advertising platforms over time.
Since Melleka took over (early 2025), Google Ads ROAS has stabilized around 1.9x-2.4x on dramatically lower budgets (~$250-310K/mo vs $1.4M/mo). Efficiency is significantly higher even though absolute revenue is lower. The account went from a bloated, multi-million dollar machine with questionable conversion tracking to a lean, purchase-focused operation.
The all_conversions metric was deeply contaminated throughout 2022-2024. In May-June 2023, the account reported 800K+ "all conversions" per month when actual purchases were ~15-17K. Any ROAS calculations using "all conversions" during this period would be completely unreliable. The PRIMARY conversions metric and conversions_value remained relatively clean, but anyone reporting on "total conversions" would have massively overstated performance.
Google Ads spend dropped from a peak of ~$1.9M/mo (Nov 2022) to ~$250-310K/mo in 2025-2026. This 75-85% budget cut coincided with ROAS improvement, suggesting the previous high spend had severe diminishing returns. The account was likely over-spending into low-value impressions and non-converting audiences at scale.
Meta's 0.33x purchase ROAS looks alarming in isolation, but Meta is driving $11.7M+/year in top-of-funnel awareness spending. The true value is in brand searches, organic traffic, and Google brand search conversions that Meta influences but doesn't get credit for. The purchase ROAS should not be compared 1:1 with Google's because their funnel roles are fundamentally different.
Google Ads NCAC has risen from ~$25 in 2021 to $55-75 in 2025-2026. This reflects both market-wide CPC inflation and reduced budget (which concentrates spend on higher-intent, more competitive keywords). The shift from volume to efficiency naturally raises per-conversion cost while improving ROAS.
As of May 2026, the correct ROAS calculation uses conversions_value directly from the API, which includes all PRIMARY purchase actions (vega Sales + Subscription + Purchase Straight Sale No Subs). Previous manual calculations that excluded "Purchase - Straight Sale No Subs" were understating ROAS by 30-40%. This is now fixed in all daily briefs.
Generated by Melleka Marketing AI Strategist. Data sourced from Google Ads API, Meta Graph API, and Windsor.ai.
All ROAS figures use conversions_value (PRIMARY actions only) / cost_micros from the Google Ads API.
Meta purchase ROAS uses purchase action_values / spend from Meta Insights API.