Channel-Level Attribution & Unattributed Metrics
Source Medium’s MTA system provides a complete view of marketing performance by combining ad-level attribution with channel-level metrics. This approach ensures that all marketing spend is accounted for, even when specific ads cannot be directly attributed.Ad-Level vs. Channel-Level Data
The MTA system organizes marketing data at two levels:Ad-Level Data
- Granularity: Individual ad creatives identified by
ad_id
- Content: Complete ad metadata, performance metrics, and attribution metrics
- Attribution: Full attribution across first-touch, last-touch, and linear models
- Identification: Records where
ad_id
is present andsm_marketing_channel
is NULL
Channel-Level Data
- Granularity: Marketing channels (Facebook, Google, etc.) identified by
sm_marketing_channel
- Content: Only unattributed spend, impressions, and clicks not already counted at the ad level
- Purpose: Account for marketing activity that cannot be tied to specific ads
- Identification: Records where
sm_marketing_channel
is present andad_id
is NULL
Unattributed Channel Metrics
To prevent double-counting while ensuring complete marketing visibility, channel-level rows in therpt_ad_attribution_performance_daily
model contain only “unattributed” metrics:
What Are Unattributed Metrics?
Unattributed metrics represent marketing activities that:- Cannot be tied to a specific ad ID
- Are not already counted in ad-level metrics
- Still contribute to overall marketing spend and performance
Key Unattributed Metrics
unattributed_channel_spend
: Marketing spend not associated with specific adsunattributed_channel_impressions
: Impressions not associated with specific adsunattributed_channel_clicks
: Clicks not associated with specific ads
How Unattributed Metrics Are Calculated
The calculation of unattributed metrics follows this process:- Total Channel Metrics: Calculate total spend, impressions, and clicks for each channel
- Ad-Level Metrics: Calculate the sum of these metrics for all ads in the channel
- Unattributed Metrics: Subtract ad-level metrics from total channel metrics
- No metric is counted twice
- All marketing spend is accounted for
- Channel-level analysis is complete and accurate
Analyzing Channel and Ad-Level Data Together
When analyzing marketing performance, it’s important to consider both ad-level and channel-level data:Complete Channel Analysis
To get a complete view of a channel’s performance:Ad-Level Detail
For granular analysis of the specific ads driving performance:Unattributed Analysis
To specifically analyze unattributed spend within channels:Business Value of Unattributed Metrics
Including unattributed metrics in the MTA system provides several business benefits:1. Complete Marketing Picture
- Total Cost Visibility: Account for all marketing spend, not just ad-level spend
- Performance Context: Understand channel performance holistically
- Budget Validation: Reconcile platform-reported spend with actual spend
2. Accurate Attribution Metrics
- No Double-Counting: Attribution metrics remain accurate by avoiding duplication
- Clean Data Separation: Ad-level and channel-level data remain distinct
- Proper Denominator: ROI/ROAS calculations use correct total spend figures
3. Marketing Mix Insights
- Channel Efficiency: Compare attributed and unattributed spend by channel
- Data Quality Gaps: Identify channels with high unattributed percentages
- Optimization Opportunities: Focus on reducing unattributed spend through better tracking
Frequently Asked Questions
Why do some channels have high unattributed spend?
Channels may have high unattributed spend due to:- Poor UTM parameter implementation
- API limitations not providing ad-level data
- Manual or offline marketing activities
- Technical tracking limitations
How can I reduce unattributed spend?
To reduce unattributed spend:- Implement consistent UTM parameters across all campaigns
- Use dedicated tracking links for offline campaigns
- Ensure proper API connections for all platforms
- Work with your SourceMedium account manager to improve tracking