Why Self-Reported Conversions From Ad Platforms Are Unreliable

Ad Platforms Are Grading Their Own Homework

Here's a scenario that plays out in thousands of ecommerce businesses every month:

Meta Ads Manager says you drove 500 purchases last month. Google Ads says 400. TikTok claims 150. Your Shopify dashboard shows 600 total orders.

500 + 400 + 150 = 1,050 platform-reported conversions for 600 actual orders. That's 75% over-counting.

This isn't a bug. It's how self-reported conversion systems are designed to work. Each platform independently tracks user interactions and claims credit for conversions that happen within its attribution window. When a customer sees a Meta ad, clicks a Google ad, and then converts, both platforms count the same conversion.

For ecommerce founders making budget decisions based on these numbers, the implications are significant. You're not just seeing inflated totals -- you're seeing inflated performance metrics for each platform, which distorts your understanding of which channels actually drive profitable growth.

How Double-Counting Happens

The mechanics are straightforward. Every major ad platform uses its own tracking pixel and attribution window to determine which conversions to claim.

Scenario: A Single Customer, Multiple Claims

1. Monday: Customer sees a Meta ad in their Instagram feed (impression logged by Meta)
2. Wednesday: Customer searches your brand on Google, clicks a Google ad (click logged by Google)
3. Thursday: Customer sees a TikTok ad (impression logged by TikTok)
4. Friday: Customer visits your site directly and purchases

What each platform reports:

• Meta: 1 conversion (view-through, within 1-day view window)
• Google: 1 conversion (click-through, within 30-day click window)
• TikTok: 1 conversion (view-through, within 1-day view window)

Reality: 1 conversion occurred. Three platforms claimed it.

This isn't fraud. Each platform genuinely had an interaction with the customer within its attribution window. The problem is that each platform measures independently and has no incentive to share credit with competitors.

The Scale of Over-Reporting

The over-reporting gap varies by industry, channel mix, and audience overlap, but the pattern is consistent.

Research findings:

• A 2024 analysis by Measured found that Meta over-reports conversions by an average of 20-40% compared to incrementality-tested baselines
• Google Ads over-reports by 15-30%, with the highest inflation in branded search (which captures existing intent rather than creating it)
• TikTok over-reports by 30-60%, partly because its attribution relies heavily on view-through conversions for its predominantly video ad format

The overlap effect: The more platforms you advertise on, the worse the over-counting gets. If a customer is in Meta's audience, Google's remarketing pool, and TikTok's interest-based targeting simultaneously, every platform will claim their conversions.

For a brand spending across 3-4 platforms, total platform-reported conversions typically exceed actual conversions by 40-80%.

Why Platforms Over-Report (It's Not Malicious)

Platform over-reporting stems from three structural factors, not deliberate deception:

1. Misaligned Incentives

Ad platforms make money when you spend more. Reporting higher conversion numbers encourages higher spend. There's no regulatory requirement for platforms to report conservatively, and no penalty for over-claiming.

This doesn't mean platform employees are manipulating data. It means the system design decisions -- default attribution windows, view-through inclusion, modeled conversions -- consistently err on the side of claiming more credit rather than less.

2. Privacy-Driven Modeling

iOS 14.5 and browser privacy changes reduced the data platforms can directly observe. In response, Meta, Google, and TikTok all introduced statistical modeling to estimate conversions they can no longer track directly.

Meta's Aggregated Event Measurement uses modeling to fill gaps in conversion data. Google's Enhanced Conversions uses modeled data to supplement observed conversions. These models are tuned to approximate the pre-privacy conversion volumes, which means they tend to add conversions rather than subtract them.

3. View-Through Attribution Defaults

Meta includes 1-day view-through conversions by default. If you're reaching hundreds of thousands of people with impressions, some percentage will convert regardless of whether your ad influenced them. The platform counts those as ad-driven conversions.

For brands with strong organic traffic, view-through attribution can be particularly misleading. A customer who was going to buy anyway sees an ad, doesn't click it, then converts through direct traffic. Meta counts this as a Meta conversion.

How to Find Your Real Numbers

Method 1: Cross-Reference Against Your Source of Truth

Your ecommerce platform (Shopify, WooCommerce, etc.) or CRM is your source of truth for actual transactions. Compare total platform-reported conversions against actual orders.

If platforms collectively report 1,000 conversions and you had 700 orders, you have a 43% over-count. This gap is your starting point for understanding which platform's numbers are most inflated.

Method 2: UTM-Based Attribution

Tag every ad URL with UTM parameters and track conversions through your analytics tool using those UTMs. This gives you click-based attribution that's independent of platform pixels.

Limitations: UTM tracking doesn't capture view-through conversions (which may have some real value) and depends on cookies for multi-session tracking (which degrade over time).

Method 3: Server-Side Attribution

Implement server-side tracking that logs every touchpoint in your own database and applies your chosen attribution model. This gives you:

• A single source of truth that doesn't rely on any platform's self-reporting
• Consistent attribution rules across all channels
• Data that isn't affected by ad blockers or cookie restrictions

Method 4: Incrementality Testing

Run holdout tests where you pause specific campaigns or platforms for segments of your audience, then measure the actual impact on conversions. This is the gold standard for understanding true incremental value, though it requires significant traffic volume and willingness to temporarily sacrifice some performance.

What to Do With This Information

Stop optimizing to platform-reported ROAS

If Meta says your ROAS is 4x but over-reports by 30%, your actual ROAS is closer to 3x. That might still be profitable, but the optimization decisions you'd make at 4x vs 3x are different.

Set realistic internal targets

If you know platforms over-report by 30-50%, adjust your internal ROAS targets accordingly. A breakeven ROAS of 2x on platform-reported numbers might mean a real breakeven of 1.3-1.5x.

Invest in independent measurement

The cost of implementing server-side tracking and independent attribution is typically a fraction of the budget waste caused by optimizing on inflated platform data. A $100K/month advertiser making a 10% budget reallocation based on more accurate data recovers the measurement investment many times over.

Don't throw out platform data entirely

Platform-reported conversions are inflated, but they're not useless. Trends within a platform are directionally accurate -- if a campaign's ROAS is declining on Meta, it's probably actually declining. The issue is comparing across platforms and using absolute numbers for budget allocation.

Frequently Asked Questions

How much does Meta over-report conversions compared to actual sales?

Based on aggregate data from incrementality studies and cross-platform analyses, Meta typically over-reports conversions by 20-40%. The variance depends on your audience overlap with other platforms, how much of your traffic comes from iOS users (where Meta's tracking is most limited and modeling is most aggressive), and whether you're including view-through conversions. Prospecting campaigns tend to have lower over-reporting than retargeting campaigns, since retargeting claims credit for customers who were already close to converting.

Why does Google Ads report different conversion numbers than Google Analytics?

Google Ads and Google Analytics use different attribution models, different attribution windows, and count conversions differently. Google Ads uses its own click attribution and includes modeled conversions. Google Analytics uses session-based or user-based attribution and only counts observed events. The discrepancy can be 15-30% even within Google's own ecosystem. Neither is necessarily "right" -- they're measuring slightly different things.

Should I turn off view-through conversion reporting in Meta?

Don't turn it off entirely, but separate it from your primary performance reports. View-through conversions have some signal value -- ad impressions do influence purchase behavior -- but they also include significant noise from customers who would have converted anyway. Report click-through and view-through conversions separately, apply different conversion values to each (many advertisers discount view-through conversions by 50-75%), and make budget decisions primarily on click-through data.

---

Go Funnel uses server-side tracking and multi-touch attribution to show you which ads actually drive revenue. Book a call to see your real numbers.