---
title: "Audience effects"
url: "https://helpdesk.telmarhelixa.com/en/article/audience-effects"
language: "en"
last_updated: "2026-10-27T15:25:10.511Z"
---

# Audience effects

Identifies high-potential target audiences by combining an initial dual-screening scatter evaluation with **CHAID** decision tree modeling (AE).

## **Use cases**

1.  **Quick segment identification:** a fast way to find the most promising segments for advertising. The dual screening on reach and index, tedious by hand across thousands of variables, takes minutes.
2.  **Advanced variable analysis:** works across thousands of demographics, product usage, attitudinal data and digital personas for digital planning.
3.  **Targeted planning:** the resulting segments (a short list of variables) build a **Gain table** of mutually exclusive audience segments.
4.  **Integration into media planning:** segments can move into planning tools such as the optional Audience Planner, or be passed to media and channel planning tools and actuators, such as online publishers and Demand Side Platforms (DSPs).

## **Mechanism and setup**

-   Analyzes a broad battery of candidate variables in **rows** against a defined target **column**.
-   **Only binary variables** (yes/no, agree/disagree) are permitted. Explore automatically prunes and excludes incompatible volumetric and calculated items from CHAID processing.
-   Initial candidates are plotted on the **Pareto Chart** (Index vs. Reach); top performers feed the CHAID engine.

## **Graphs and tables**

### **Pareto Chart**

Scatter plot of candidate variables (including negative NOT conditions) by **Index** (X axis) and audience **Reach** (Y axis). High-performing variables are marked as **Recommended Variables** (green indicators) and supplied to the CHAID segmentation model.

### **Gain Analysis**

-   **Gain Table:** lists the resulting CHAID audience paths in descending order of Index (Lift), with cumulative segment sizing and target concentration metrics.
-   -   Index = 100 × (target percentage in segment ÷ base percentage in segment).
    -   Example from the source: a segment that is 3.3% of the total population but holds 15.1% of target users has an Index of **458**.
-   **Gain Plot:** evaluates model lift against the baseline diagonal. Segments well above the baseline show strong target affinity. A flat trajectory means the analyst should add more discriminating variables.

### **The analogy**

Traditional crosstab analysis is like manually sifting through data. Audience Effects is a magnetic detector: the Pareto Chart does the first high-yield scan, the Gain Table isolates the high-performing segments, and the Gain Plot visually confirms the target segment is statistically rich.