Marketing Channel Attribution Models and Conversion Effectiveness Evaluation

Marketing activities aim to combine brand awareness with direct performance. To allocate budget rationally, the contribution of each channel to conversion goals must be measured through channel attribution.

Basic Concepts

Channel evaluation requires four elements: dimensions, metrics, segmentation, and attribution.

Dimensions

Dimensions are attributes used to classify data for comparative analysis. Common dimensions in channel effectiveness analysis include channel type, device, region, etc.

Metrics

Metrics measure data performance and can be basic or composite. They cover volume, quality, cost, and revenue on both traffic and conversion sides.

Segmentation

Segmentation defines subsets of users, sessions, or events. Typical user segmentation rules (e.g., gender = female, age = 18‑35, city tier = first‑tier) create target audiences for filtering, comparison, or remarketing.

Attribution

An attribution model assigns conversion credit to touchpoints along the conversion path. Models can be rule‑based (last interaction, first interaction, linear, time decay) or algorithmic (Shapley value, etc.). Attribution can be single‑touch or multi‑touch, combined with an attribution window.

Attribution Analysis

What is Attribution?

Attribution distributes the value of a conversion to the various marketing touchpoints (exposures or clicks) that a user experienced before the conversion.

Attribution Models

Choosing a model influences how channel contributions are allocated. Traditional last‑click assigns all credit to the final touchpoint, often under‑valuing assisted channels. Multi‑Touch Attribution (MTA) distributes credit across all touchpoints within a defined window, revealing a fuller picture of channel synergy.

Common models include last interaction, first interaction, linear, time decay, and data‑driven models based on Shapley values.

Data‑Driven Attribution (Shapley Value)

The Shapley‑based model evaluates each channel’s marginal contribution across all possible subsets of channels, weighting by the probability of each coalition. The resulting Shapley value represents the fair contribution of a channel within the whole set.

Example analysis shows that linear allocation over‑estimates the contribution of a video ad channel, while the Shapley‑based approach assigns lower credit consistent with its true marginal impact.

Attribution Window

The attribution window defines the time span after a touchpoint during which conversions are credited to that touchpoint (e.g., 7 days).

Building an Attribution Model

The model construction follows four generic steps: data collection, touchpoint definition, model selection, and contribution calculation.

Case Studies

Conversion Time Analysis

Purpose: understand decision‑making cycles, evaluate appropriate attribution windows, and identify remarketing moments.

Method: select conversion type, channels, and attribution model; the report sorts conversions by elapsed days, showing counts, values, and proportions.

Insights: a 30‑day window can be sliced into 1, 7, 15, 30‑day views to assess speed of conversion and optimal remarketing timing.

Attribution Comparison

Goal: compare multi‑touch models against last‑click to detect undervalued channels and adjust budgets.

Example: a skincare brand’s 30‑day attribution window showed that linear attribution better reflected the value of brand‑building display ads compared with last‑click.

Channel Conversion Effectiveness Evaluation

Key indicators: reach, conversion volume, conversion rate, ROI, conversion cost, CPUV.

Flow Conversion Report presents traffic, conversions, cost, and ROI per channel, enabling quality analysis and ROI calculation.

Conversion Quality Matrix visualizes channels on a bubble chart (e.g., ROI vs. interaction vs. traffic) to classify channels into four quadrants: high‑quality, efficient, low‑quality, and high‑volume‑low‑efficiency.

Conclusion

By selecting appropriate attribution models and windows, marketers can accurately assess channel contributions, optimize budget allocation, and improve ROI. Data‑driven Shapley attribution offers a fair, mathematically grounded method for multi‑channel credit assignment.