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

# Reporting and optimization

> Understand what the results mean and decide what to do next.

## Agree the measurement plan first

At kickoff, define the primary conversion, its data source, the reporting period and time zone, and the attribution approach. Confirm who validates qualified leads or sales. The measurement work included in your engagement is set out in your scope.

Where suitable, campaign links can use consistent UTM parameters to identify traffic in your analytics. Conversion tracking requires separate setup and validation where supported; a tagged link alone does not prove that a sale came from an ad.

For platform capabilities, see [OpenAI’s measurement documentation](https://developers.openai.com/ads). Available metrics and tracking options depend on the account and setup.

## Read metrics together

| Metric | What it tells you | What to check |
| - | - | - |
| Impressions | Recorded ad displays | Reach does not establish purchase intent |
| Clicks and click-through rate | Traffic and the share of impressions that led to clicks | Whether visits lead to useful actions |
| Media spend | Recorded advertising cost | Reporting delays and the approved limit |
| Cost per click | Media spend divided by recorded clicks | Traffic quality alongside cost |
| Conversions and cost per acquisition | Recorded actions and media spend per action | Event accuracy, duplicates, and attribution |
| Qualified leads or sales | Business outcomes validated by your team | Lead quality, sales delays, and missing records |
| Return on ad spend | Attributed revenue divided by media spend | Revenue coverage; this is not profit or return after agency fees |

When a denominator is zero, report the metric as unavailable rather than inventing a value. Do not present missing conversion data as zero confirmed business results.

## What a review should answer

We agree the reporting cadence with you. A useful review explains the work completed, spend against plan, results against the objective, data limitations, what we learned, and the next recommended action with an owner.

For example, more clicks with fewer qualified inquiries may call for an offer or landing page review before increasing spend.

## Understand attribution limits

Platform reports, website analytics, and your sales records may differ because of attribution windows, consent choices, reporting delays, and the time between a visit and a purchase. Compare like periods and definitions before drawing conclusions.

Share lead quality and sales feedback through the agreed channel using only the necessary information. We use it to improve the next test. Paid campaign results and organic AI visibility should be reported separately.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.