> ## Documentation Index
> Fetch the complete documentation index at: https://docs.playbackrewards.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Reporting

Reporting API returns a partner-level report as a flat array of rows. Each row represents one combination of your requested grouping dimensions (an offer, a day, or both), with the metric columns requested attached to it. For technical details you can refer to the [API reference](/api-reference/partners/generate-a-partner-report).

## **Metrics reference**

Columns fall into three families.

* Engagement metrics: plain counts and totals.
* Retention: Cohorted data (Resets per window).
* ROAS: Cohorted data (Cumulative).

### **Engagement**

| Column               | Description                               |
| -------------------- | ----------------------------------------- |
| `clicks`             | Total clicks recorded.                    |
| `installs`           | Total installs attributed.                |
| `purchases`          | Total in-app purchase events.             |
| `revenue`            | Total revenue generated.                  |
| `advertiser_revenue` | Total revenue reported by the advertiser. |

### **Retention (Resets per window)**

| **Column**         | **Description**                                  |
| :----------------- | :----------------------------------------------- |
| `day_1_retention`  | Share of installed users who returned on day 1.  |
| `day_3_retention`  | Share of installed users who returned on day 3.  |
| `day_7_retention`  | Share of installed users who returned on day 7.  |
| `day_14_retention` | Share of installed users who returned on day 14. |
| `day_30_retention` | Share of installed users who returned on day 30. |

### **ROAS (Cumulative)**

| **Column**    | **Description**                                               |
| :------------ | :------------------------------------------------------------ |
| `day_1_roas`  | Return on ad spend, revenue earned within 1 day of install.   |
| `day_3_roas`  | Return on ad spend, revenue earned within 3 days of install.  |
| `day_7_roas`  | Return on ad spend, revenue earned within 7 days of install.  |
| `day_14_roas` | Return on ad spend, revenue earned within 14 days of install. |
| `day_30_roas` | Return on ad spend, revenue earned within 30 days of install. |

## **Cumulative vs. resetting**

The two windowed families look alike — both come in day 1/3/7/14/30 flavors — but read them differently when reporting or charting.

**ROAS is cumulative**

Each window includes every prior one. `day_30_roas` already contains the revenue counted in `day_1_roas`, `day_7_roas`, and so on — it's running total ÷ spend, measured further and further out. Later windows are always greater than or equal to earlier ones for the same cohort.

day\_1: 0.12 → day\_7: 0.31 → day\_30: 0.50

**Retention is not cumulative**

Each window is an independent snapshot: the share of the cohort that came back on *that specific day*, not the share that came back at any point up to it. `day_30_retention` does not include the users counted in `day_1_retention` — it can be lower, higher, or unrelated.

day\_1: 0.42 → day\_7: 0.18 → day\_30: 0.09