What This Chart Shows
This chart shows you what would have happened if we had removed a specific channel. It shows the incremental value each channel contributes to your outcomes. The chart has two side-by-side panels: the left panel shows total attributed contribution per channel (the lift you would lose if that channel were removed), and the right panel shows efficiency (lift generated per dollar invested). Channels are grouped by event category, so you see aggregated performance across all tactics within each category. This helps you understand which channels are truly driving results and which are most cost-effective.
Key Questions This Chart Helps Answer
- Which channels contribute the most incremental value to our outcomes?
- What would we lose if we turned off a specific channel entirely?
- Which channels deliver the best return per dollar spent?
- Are any channels producing negative incremental value (cannibalization or saturation)?
- How do channel contributions compare when accounting for investment levels?
Axes, Metrics, and Units
Element | Description |
|---|---|
| X-axis (both panels) | Channel / Event Category. Each bar represents one event category, aggregating all series (tactics, campaigns) within that category. |
| Y-axis (left panel) | Contribution: Total attributed lift (modeled incremental value) for that channel. Units match your KPI (e.g., revenue, conversions). |
| Y-axis (right panel) | Lift per $ Invested: Efficiency metric = contribution divided by total cost for that channel. Higher = more efficient. |
| Bar color | Green = positive contribution (channel adds value). Red = negative contribution (channel may cannibalize or saturate). |
| Text labels | Contribution values shown above/below each bar. |
| Hover | Shows channel name, contribution or efficiency value, and number of series in that category. |
All values are modeled. This is an ablation proxy, not a live simulation.
Control Options Reference
Control | Meaning |
|---|---|
| KPI classification / hierarchy / drilldown | Which outcome (KPI / target series) to attribute to. |
| Event categories | Filter by specific event categories (include or exclude). |
| Funnel stages | Filter by specific funnel stages (include or exclude). |
How to Interpret the Results
- Left panel (Contribution): Taller bars = more incremental value. These are the channels that matter most to your outcomes. Negative (red) bars indicate channels that may be hurting net performance (e.g., overcrowding or saturation).
- Right panel (Efficiency): Taller bars = better ROI per dollar. A channel can have high contribution but low efficiency (high spend) or low contribution but high efficiency (low spend, good returns).
- Compare panels: A channel with high contribution but low efficiency may benefit from optimization rather than scaling. A channel with high efficiency but low contribution may have room to scale.
- Negative values: Red bars (negative) suggest the channel's net effect is negative—either due to cannibalization, over-saturation, or misattribution. Investigate further before cutting budget.
- Click for details: Clicking a bar can show the individual series (tactics) within that category and their contributions.
Practical Applications for Marketers
Application | How to use this chart |
|---|---|
| Budget reallocation | Shift spend from low-contribution or negative channels to high-contribution, high-efficiency channels. |
| Channel justification | Use contribution values to justify continued investment in a channel to stakeholders. |
| Cut or scale decisions | Channels with large positive contributions are candidates for scaling; channels with negative contributions are candidates for review or cuts. |
| Efficiency optimization | Focus on improving efficiency in high-contribution but low-efficiency channels (better targeting, creative, etc.). |
| Experiment prioritization | Use ablation estimates to prioritize which channels to test with holdout experiments for validation. |
Common Mistakes & Misinterpretations
Mistake | Why it is a problem | How to avoid |
|---|---|---|
| Treating this as a live experiment | This is a modeled proxy for ablation, not a true holdout test. Real ablation requires turning off a channel and measuring. | Use as directional guidance; validate major decisions with actual experiments. |
| Ignoring efficiency when scaling | A channel with high contribution but low efficiency may not scale profitably. | Always compare both panels before increasing spend. |
| Assuming negative = broken | Negative contributions can arise from legitimate saturation or cross-channel effects, not just data errors. | Investigate negative channels before cutting; check for over-frequency or audience overlap. |
| Comparing different time periods | Contribution values depend on the time window and data coverage. | Keep time periods consistent when comparing channels. |
| Ignoring series composition | A category's contribution is the sum of its series. One dominant series can skew the category total. | Click into the category to see series-level breakdown. |
Caveats & Considerations
- Ablation proxy: True ablation (removing a channel and observing the outcome) requires live simulation or experiments. The model estimates what would happen, but real-world effects may differ.
- Top 12 categories: Only the top 12 categories by absolute contribution are displayed. Smaller categories are not shown.
- Aggregation: Contributions and costs are aggregated by event category. Individual series (tactics) within a category are available via click-through.
- Efficiency denominator: Efficiency = contribution / cost. If cost is near zero, efficiency can be very high or undefined; a small floor is used to avoid division by zero.