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

# Data and Setup FAQ

> Answers to common questions about connectors, data layers, composite layers, QA, and common data issues.

## Overview

## What data does Provalytics need?

At a minimum, Provalytics needs reliable inputs for marketing activity and the business outcome being modeled.

Depending on the setup, that can include:

* spend
* impressions
* clicks
* website activity
* conversions
* revenue

## What is a data layer?

A data layer is the logical input layer Provalytics uses to connect source data to reporting and modeling.

## What is a composite layer?

A composite layer combines multiple source layers into one logical website or conversion layer.

This is useful when a client has a historical source plus an ongoing source.

## Why do data naming conventions matter?

Because stable naming improves mapping, interpretation, QA, and reporting continuity.

## What does Connector QA do?

Connector QA helps confirm that the source totals being ingested look correct before teams rely on them in reporting or model work.

## What are the most common data issues?

The most common issues are:

* missing days
* unexpected zeros
* stale files
* inconsistent naming
* hierarchy shifts

For more detail, see:

* [Common Data Issues and Prevention](/integrations/common-data-issues-and-prevention)