Vegamour | Cookie Consent Incident & Recovery Report

Executive Summary

What Happened and How We Fixed It

On August 10, 2026, a cookie consent policy issue on vegamour.com broke Google Ads conversion tracking. Conversions stopped attributing properly, making it appear that ROAS had dropped to near-zero levels. The Melleka team identified the root cause, recommended implementing Google Consent Mode v2, and the SEO team executed the fix on August 16. Within days, conversions began re-attributing, and by August 20, reporting had fully normalized. Critically, the issue was a tracking/attribution problem only, not a true performance decline.

Reported ROAS (Aug 12)

0.83x

Before fix, as reported that day

➔

Actual ROAS (Aug 12)

2.17x

After conversions re-attributed

ⓘ

Key Takeaway: The real-time ROAS shown during August 10-16 was misleading because the cookie consent issue prevented conversion tracking tags from firing. Once Google Consent Mode v2 was implemented, Google retroactively attributed the "lost" conversions back to their correct dates. The actual performance never dropped; only the reporting did.

Chronology

Incident Timeline

August 1 - 9 (Pre-Incident)

Normal Performance Baseline

ROAS averaging 3.40x across all Google Ads campaigns. Conversion tracking functioning correctly. Daily revenue averaging $25,000+ attributed properly.

August 10 (Incident Start)

Cookie Consent Policy Breaks Conversion Tracking

A cookie consent/policy change on vegamour.com prevented Google Ads tracking tags from firing for users who had not accepted cookies. Real-time dashboards began showing a sharp ROAS decline, dropping from ~2.5x daily average to approximately 0.3x.

August 12 (Peak Visible Impact)

ROAS Brief Shows 0.83x

The daily ROAS Brief email reported 0.83x ROAS for the previous day, graded "F - Critical, Immediate Review Required." This was the clearest evidence of the tracking disruption. Even after full re-attribution, Aug 12 remains the lowest ROAS day of the month at 2.17x.

August 13 (Diagnosis)

Root Cause Identified

Melleka team analyzed the sudden drop, ruled out campaign-level issues (spend and traffic remained stable), and identified the cookie consent policy as the cause. Recommended implementing Google Consent Mode v2 to the SEO team.

August 16 (Fix Deployed)

Google Consent Mode v2 Implemented

The SEO team implemented Google Consent Mode v2 on vegamour.com, following Melleka's recommendation. This allows Google to model conversions for users who decline cookies, while still respecting privacy preferences. Conversion signals began flowing again immediately.

August 17 - 20 (Recovery)

Conversions Begin Re-Attributing

Over the following days, Google retroactively attributed previously "lost" conversions back to their original click dates. ROAS reporting gradually normalized. By August 20, the NCAC Performance Brief showed $232.55 NCAC with 201% ROAS, confirming the attribution pipeline was healthy again.

August 20+ (Fully Resolved)

Full Recovery Confirmed

All conversion data fully re-attributed. Historical data corrected across the incident window. Monthly performance fully intact. No actual revenue was lost.

Evidence

Internal Alert Emails During the Incident

These screenshots show the automated performance briefs sent during the incident window. The first shows the 0.83x ROAS "F" grade from August 12, when the cookie issue was at peak impact. The second shows the NCAC brief from August 20, when attribution had recovered and metrics normalized.

ROAS Brief showing 0.83x F grade on August 12

August 12: ROAS Brief (During Incident) Grade F, 0.83x ROAS. Conversion tracking was broken; tags were not firing due to cookie consent blocking.

NCAC Brief showing $232 NCAC and 201% ROAS on August 20

August 20: NCAC Brief (Post-Recovery) $232.55 NCAC, 201% ROAS. Conversions flowing properly after Google Consent Mode v2 fix.

The Reporting Gap

What We Saw in Real-Time vs. What the Data Shows Now

The cookie consent issue caused a massive disconnect between what was actually happening (sales continued as normal) and what Google Ads was reporting (near-zero conversions). After Consent Mode v2 was implemented, Google retroactively attributed the missing conversions back to their original dates.

⚠ Real-Time Reporting (During Incident)

~0.3x - 0.8x

ROAS as reported Aug 10-16

❌ Conversions appeared to drop to near zero
❌ Daily ROAS brief showed "Grade F" performance
❌ Email evidence: Aug 12 showed 0.83x ROAS
❌ Smart bidding strategies receiving incomplete data

✓ Corrected Data (After Re-Attribution)

2.17x - 4.00x

ROAS after conversions backfilled

✔ Aug 10 corrected to 4.00x ROAS (was ~0.3x)
✔ Aug 12 corrected to 2.17x (was 0.83x)
✔ All missing conversions attributed to correct dates
✔ Proves revenue was never lost, only tracking broke

Why the difference? Without Consent Mode v2, users who declined cookies generated zero conversion data. Google could not see their purchases. After the fix, Google modeled those conversions using aggregated signals and backfilled them to the original purchase dates.

Impact Assessment

By the Numbers

Pre-Incident Avg ROAS

~2.5x

Consistent daily performance

Real-Time ROAS During Incident

~0.3x - 0.8x

As reported Aug 10-16

Days to Identify

Aug 10 start, Aug 13 diagnosed

Days to Fix

Aug 10 start, Aug 16 deployed

Revenue Lost

Only tracking was affected

Recovery Visible By

Aug 20

Full attribution restored

Resolution

Google Consent Mode v2: The Fix

Google Consent Mode v2 is a privacy-compliant framework that allows Google tags to adjust their behavior based on a user's cookie consent choices. Here is how it resolved the issue:

Before (Without Consent Mode)

❌ Users who declined cookies generated zero conversion data
❌ Google Ads could not attribute any purchases from those users
❌ ROAS appeared artificially low due to missing conversions
❌ Bid strategies received incomplete signals, potentially degrading optimization

After (With Consent Mode v2)

✔ Tags adjust behavior based on consent status (no privacy violation)
✔ Google models conversions for non-consented users using aggregated data
✔ ROAS reporting reflects true campaign performance
✔ Smart bidding receives complete signals for proper optimization

Key Insight

No Revenue Was Lost. Only Reporting Was Affected.

The most important takeaway: actual sales and revenue were never impacted. Customers continued purchasing throughout the incident. The cookie consent issue only prevented Google from attributing those purchases back to the ads that drove them.

Once Google Consent Mode v2 was implemented, the conversion data backfilled to its correct dates. The proof is visible in the corrected historical data, where days that initially showed near-zero performance now reflect their true ROAS.

The Melleka team identified the issue within 3 days, provided the specific technical recommendation (Consent Mode v2), and the SEO team executed the fix within 6 days of the incident start. The account has been tracking accurately since.