Understanding ad attribution for e-commerce: Simplified guide

Customer sees Facebook ad for your product. Doesn’t buy. Three days later searches Google for your brand, clicks ad, browses but doesn’t purchase. Week later opens your email, clicks product link, finally buys. Who gets credit for the sale? Facebook says Facebook (initial discovery). Google says Google (last ad click). Email platform says email (final touchpoint). Total revenue claimed across all three exceeds actual sale value. Welcome to attribution confusion.

Ad attribution attempts to answer “which marketing generated this sale?” But rarely is answer singular. Most purchases involve multiple touchpoints before buying decision. Attribution models distribute credit differently based on assumptions about what mattered most. This creates conflicting reports where platforms over-credit themselves and total attributed revenue exceeds actual revenue.

This guide explains how ad attribution works, why platforms report different numbers, and how to use attribution data for optimization decisions without getting paralyzed by conflicting claims about who deserves credit.

Why ad attribution is complicated for e-commerce

Unlike direct response where person sees ad and immediately buys, e-commerce customers often research across days or weeks before purchasing. They might:

• See Facebook ad, don’t click

• Later search branded term on Google, click ad, browse

• Receive cart abandonment email, ignore

• Return directly to site few days later, purchase

Which touchpoint “caused” the sale? Facebook created awareness. Google captured search intent. Email reminded about cart. Direct return showed purchase decision. All contributed. But platforms using last-click attribution credit final touchpoint exclusively, ignoring everything before.

This creates three problems: overlapping credit (multiple platforms claim same sale), missing credit (none capture direct return influenced by earlier ads), and optimization confusion (platforms push you to optimize wrong things based on incomplete attribution view).

What doesn’t fix attribution complexity

❌ Believing one platform’s attribution as “truth”

Why it doesn’t work: Google Ads says it generated $12,400 revenue. Meta says $8,600. Email platform says $6,200. Added together: $27,200. Actual store revenue from all sources: $18,300. Accepting any single platform’s claim as truth ignores that they’re measuring partial journey, not complete picture. Each platform optimistically credits itself.

❌ Trying to reconcile every discrepancy between platforms

Why it doesn’t work: Attribution windows differ (7-day, 30-day, click versus view), tracking methods vary (pixel versus SDK versus server-side), and user behavior changes across devices making perfect reconciliation impossible. Spending hours trying to make numbers match perfectly wastes time better spent optimizing campaigns.

❌ Only tracking last-click attribution

Why it doesn’t work: Last-click gives full credit to final touchpoint before purchase, ignoring all earlier interactions. This systematically undervalues discovery channels (Facebook, Instagram, display) that introduce customers, while overvaluing capture channels (Google branded search, email) that get final click from already-interested customers. Leads to cutting discovery spend that actually feeds bottom-funnel conversions.

Three practical approaches to ad attribution

Approach 1: Platform-specific optimization with cross-channel reconciliation

What it is: Use each platform’s own attribution data for optimizing that platform’s campaigns, but reconcile cross-channel at business level using website analytics showing all traffic sources.

How it works:

1. Optimize Google Ads using Google Ads conversion data—which keywords, campaigns, ad groups perform best within Google Ads

2. Optimize Meta Ads using Meta conversion data—which audiences, creative, placements perform best within Meta

3. Use Google Analytics or e-commerce platform analytics to see total revenue by channel with multi-touch attribution

4. Accept that platform totals won’t sum to Analytics total due to overlap

5. Make budget allocation decisions (how much to spend on Google versus Meta versus email) based on Analytics cross-channel data, not platform self-reporting

Time investment: Daily 5 minutes per platform for optimization, weekly 15 minutes reconciling in Analytics.

Cost: Free if using Google Analytics, or included with e-commerce platform.

Best for: Most small stores, founders wanting simple approach without attribution software, stores running 2-4 advertising channels.

Limitations: Doesn’t provide perfect multi-touch attribution. Still relies on last-click or platform defaults. Gives directional guidance rather than precise credit allocation.

Approach 2: First-click and last-click comparison

What it is: Compare performance using first-click attribution (credit goes to first touchpoint) versus last-click attribution (credit goes to final touchpoint) to understand different channel roles.

How it works:

1. Configure Google Analytics to show both first-click and last-click attribution models

2. Review revenue by channel under each model

3. Channels with high first-click attribution but low last-click are “discovery” channels (introduce customers)

4. Channels with low first-click but high last-click are “conversion” channels (close deals with already-interested customers)

5. Budget allocation recognizes both roles—need discovery to feed pipeline and conversion to close deals

6. Avoid cutting discovery channels just because they don’t get last clicks

Example insight: Facebook shows $6,400 first-click revenue but $2,100 last-click revenue—indicates Facebook introduces customers but doesn’t close them. Google branded search shows $1,800 first-click but $8,200 last-click—captures existing intent created elsewhere. Both valuable for different reasons.

Time investment: Setup 30 minutes configuring Analytics models, monthly 20 minutes reviewing comparison.

Cost: Free with Google Analytics 4.

Best for: Stores with multiple paid channels, founders wanting to understand channel roles beyond last-click, businesses seeing large discrepancies between platform reporting and Analytics.

Limitations: Still simplified view. Many journeys involve 3-5+ touchpoints. First and last click ignore middle interactions. Better than pure last-click but not comprehensive multi-touch attribution.

Approach 3: Dedicated attribution software with multi-touch modeling

What it is: Third-party attribution tools that track full customer journey across all touchpoints and distribute credit using sophisticated models (time decay, data-driven, position-based).

How it works:

1. Install attribution platform pixel or integrate with existing platforms

2. Tool tracks customer journey across all touchpoints—ads, email, organic, direct, social

3. Choose attribution model (linear, time decay, data-driven, custom)

4. Platform distributes revenue credit across all involved touchpoints based on model

5. Reports show “true” contribution accounting for multi-touch journeys

6. Use for budget allocation and channel mix optimization

Example attribution breakdown: $100 sale distributed as: Facebook ad (first touch) $30, Google search (research) $25, Email (reminder) $20, Direct return (final visit) $25. More realistic than giving full $100 to last touchpoint.

Time investment: Setup 1-2 hours, daily/weekly 5 minutes reviewing dashboards, monthly 30 minutes analyzing cross-channel impact.

Cost: Typically $100-500/month for small business needs (Northbeam, TripleWhale, Rockerbox, Hyros).

Best for: Stores spending $10k+ monthly across multiple channels, businesses with complex multi-touch journeys (long consideration periods), teams wanting data-driven budget allocation across 4+ channels.

Limitations: Monthly cost, setup complexity, tracking limitations in privacy-focused environment (iOS blocking, cookie restrictions), still estimates rather than perfect measurement.

Which attribution approach is right for you?

Choose Approach 1 (platform-specific + Analytics) if:

• Spending under $5k monthly on ads total

• Running 2-3 channels (Google Ads, Meta, maybe email)

• Want simple, free approach without software costs

• Comfortable with directional insights rather than precise attribution

Choose Approach 2 (first-click/last-click comparison) if:

• Seeing large discrepancies between platform reporting and reality

• Need to understand channel roles (discovery versus conversion)

• Want better attribution than last-click without software costs

• Spending $3k-10k monthly across multiple channels

Choose Approach 3 (attribution software) if:

• Spending $10k+ monthly across 4+ channels

• Customer journeys involve 3-5+ touchpoints regularly

• Budget allocation decisions benefit from precise multi-touch data

• Software cost ($100-500/month) is small relative to ad spend

How to use attribution data without paralysis

For tactical optimization: Use platform-specific data. Google Ads data optimizes Google Ads campaigns. Meta data optimizes Meta campaigns. Don’t worry about cross-platform reconciliation for these decisions.

For budget allocation: Use cross-channel attribution (Analytics or attribution software). Decide how much total budget to Google versus Meta versus email based on multi-touch reality, not platform self-reporting.

For profitability assessment: Use conservative attribution (website analytics or last-click). Better to underestimate channel contribution and be pleasantly surprised than overestimate based on platform claims and discover unprofitability.

For understanding channel roles: Compare first-click versus last-click. Reveals which channels introduce customers versus convert them. Prevents cutting valuable discovery channels that don’t get last-click credit.

Common attribution mistakes

Mistake 1: Summing platform-reported revenue

Adding Google Ads $12k + Meta $8k + Email $6k = $26k total when actual revenue is $18k. Overlapping attribution creates double-counting. Use total revenue from website analytics as source of truth, not sum of platform claims.

Mistake 2: Cutting channels that don’t get last clicks

Pausing Facebook because it shows low last-click attribution, then watching Google branded search conversions decline because Facebook was creating awareness driving branded searches. Discovery channels feed bottom-funnel channels. Cutting top kills bottom.

Mistake 3: Over-crediting branded search

Celebrating Google branded search ROAS of 12.0 without recognizing those customers already knew brand from somewhere else. Branded search captures demand created by other channels. Still valuable but different from generating new demand.

Mistake 4: Ignoring direct traffic after ads

Customer clicks ad, browses, returns directly later to purchase. Ad platforms don’t credit direct traffic they influenced. Attribution software can connect these journeys. Without it, ads look less effective than reality.

Frequently asked questions

Why do Google Ads and Google Analytics show different conversion numbers?

Attribution windows differ, cross-device tracking varies, privacy settings affect tracking differently. Normal variance is 10-20%. Use Ads numbers for campaign optimization, Analytics numbers for business reporting. Accept they won’t match perfectly.

Is view-through attribution (Meta) legitimate or inflated?

View-through credits impressions without clicks. Can be legitimate (person saw ad, remembered brand, purchased later) or inflated (person saw ad accidentally, would have purchased anyway). Truth is usually somewhere between. Compare view-through to click-through attribution to gauge impact. If view-through claims 2-3x more revenue than click, probably inflated. If 20-40% more, likely capturing real incremental value.

Should I use 1-day, 7-day, or 28-day attribution window?

For most e-commerce, 7-day click attribution balances recency with consideration time. 1-day is too short for products requiring research. 28-day risks crediting ads for purchases that would happen anyway. Use 7-day for optimization, check 28-day occasionally to see fuller impact.

How do I prove which ads actually drive sales versus just get credit?

Run holdout tests. Pause channel for 2-4 weeks, measure total revenue impact. If pausing Facebook reduces overall revenue 15%, Facebook drives incremental value beyond its attributed revenue. If pausing has no impact, attribution is inflated—credit but no causation. Test one channel at a time with sufficient gap to measure true impact.

Peasy delivers store analytics to your inbox every morning—no dashboard checking needed. Starting at $49/month. Try free for 14 days.