---
title: "10 Things to Check in Your ChatGPT Ads Account by Day 14"
description: "Paying someone to run your ChatGPT ads? Check these 10 account, tracking and landing-page artefacts before you pay for another month."
url: "https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14"
image: "https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/a446098b-ac69-43fe-9bde-9c807d2c6618.png"
published: "2026-10-31T06:23:38.264Z"
modified: "2026-10-31T06:23:38.316Z"
---

October 31, 2026 · 11 min read

# 10 Things to Check in Your ChatGPT Ads Account by Day 14

[Alexander PerelmanHead Of Product @ groas](https://groas.com/author/alexander-perelman)[LinkedIn](https://www.linkedin.com/in/alexander-433793253/)

![An open car hood reveals an empty engine bay holding only a glossy spiral-bound strategy deck, suggesting you paid for advice instead of working machinery.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/a446098b-ac69-43fe-9bde-9c807d2c6618.png)

In this article

1. [1. Put the ad account in your name and on your billing](#1-put-the-ad-account-in-your-name-and-on-your-billing)
2. [2. Fire a test conversion and find it in Ads Manager](#2-fire-a-test-conversion-and-find-it-in-ads-manager)
3. [3. Build ad groups around distinct conversation intents](#3-build-ad-groups-around-distinct-conversation-intents)
4. [4. Send each intent to a page that answers it](#4-send-each-intent-to-a-page-that-answers-it)
5. [5. Cap the test budget and require approval to raise it](#5-cap-the-test-budget-and-require-approval-to-raise-it)
6. [6. Record your Google and Bing baseline before judging ChatGPT](#6-record-your-google-and-bing-baseline-before-judging-chatgpt)
7. [7. Keep a change log that explains the work](#7-keep-a-change-log-that-explains-the-work)
8. [8. Launch ad variations instead of promising them](#8-launch-ad-variations-instead-of-promising-them)
9. [9. Agree on the number that kills the test](#9-agree-on-the-number-that-kills-the-test)
10. [10. Name the person who can answer for the account](#10-name-the-person-who-can-answer-for-the-account)

A strategy deck cannot fire a test conversion. If you hired someone to run your ChatGPT ads and all you can open after two weeks is a PDF, you bought advice, not management. This list is for the business owner who wants the work done, not a course on how to do it. **By day 14, these 10 artefacts should exist in your account, tracking or landing pages.** Most take minutes to check. Start with access: without it, you have to take the provider’s word for the other nine.

## 1. Put the ad account in your name and on your billing

You should own the advertiser account in ChatGPT Ads Manager under your company email, with your payment method on file. The provider should appear as an invited user under members and permissions. OpenAI requires an owner to create the account and invite others; [agencies should not create accounts on a client’s behalf](https://searchengineland.com/run-chatgpt-ads-484513).

This is not paperwork for its own sake. If a provider owns the account, leaving them can also mean leaving your account history and pixel data behind. There is no Google MCC equivalent here to make an agency-owned setup a harmless shortcut. The convenient arrangement is the one you can still use after the relationship ends.

**Check this first:** log in yourself, confirm that you are the owner and the provider is a user, then confirm that media spend bills to you separately from the management fee. While you are there, check the country, currency and timezone. Those choices are permanent at account creation; getting one wrong means starting a new account. If you cannot log in and see your own billing, stop asking about the reporting deck. You cannot independently verify the work.

## 2. Fire a test conversion and find it in Ads Manager

Tracking is not finished when someone says the pixel is installed. For this test, you want the OpenAI Pixel on your site, Conversions API using the same event ID for deduplication, and at least one test conversion fired before spend starts. The proof is an event you can identify in Ads Manager, not a green tick in a setup document.

Here is the mechanism. ChatGPT appends an `oppref` click ID to your landing URL. That ID has to survive redirects if a click is going to connect to a form fill or purchase. Capture the ID, pass the conversion back cleanly, and you can begin reading cost per acquisition. Lose it along the way and you are paying for traffic you cannot reliably attribute.

**Run the check yourself:** submit a test lead or make a test purchase, then ask the provider to find that event by name and timestamp. Have them show you what happens to the click ID across the final URL and any redirects. Do not mistake ordinary reporting quirks for broken tracking: spend can lag 7 to 8 hours, CTR refreshes about every 15 minutes, and [modeled conversions can make Ads Manager totals differ from GA4](https://www.airops.com/blog/how-to-advertise-on-chatgpt). But if nobody can show you the test event, do not raise the budget. No test conversion, no budget increase.

## 3. Build ad groups around distinct conversation intents

The account structure should read Campaign > Ad group > Ad. At ad-group level, a freeform _context hint_ describes the conversations where an ad fits. It guides matching; it does not guarantee delivery. That is a different job from pasting a Google keyword list into a new box and calling it a build.

I would look for intent groups that answer different buyer questions. ChatGPT matches ads against the context of a conversation, so the provider needs to say which conversation each group is meant to enter. A list of 200 old search terms does not explain that choice. Neither does splitting the same list into two tabs.

**Open two ad groups and read their hints aloud.** Can you tell which question each is trying to meet? By day 14, I would expect at least two distinct intent groups live, not one catch-all with a new platform name on it. Do not turn that preference into a claim that question-style hints always win: [one early test found a keyword-style group drew more volume, slightly higher CTR and lower CPC than a question-style group](https://searchengineland.com/run-chatgpt-ads-484513). That is a reason to test the structure, not a reason to skip one. The artefact you need now is a deliberate set of live groups whose differences you can explain.

## 4. Send each intent to a page that answers it

A dedicated landing page should connect the ad-group hint, the ad and the offer. If the conversation is about dental implants, the page should answer that question rather than deposit the visitor on a generic homepage and hope they keep looking. The headline, proof and offer need to belong to the same conversation.

There is a review reason as well as a conversion reason. [OpenAI reviews an ad together with its landing page, and they need to match](https://www.airops.com/blog/how-to-advertise-on-chatgpt). A relevant page also gives the click somewhere useful to go while preserving the ID needed for attribution. The sequence matters: match the question, keep the tracking intact, then judge whether the traffic converts.

**Click every live ad’s final URL.** Check that it leads to the intended page rather than the homepage, that the page answers the question in the hint, and that it allows both OAI-AdsBot and OAI-SearchBot. Ask the provider to show where the click ID lands after any redirect. If every ad goes home, they built ads, not a funnel.

## 5. Cap the test budget and require approval to raise it

A live campaign needs a daily cap you agreed to in writing. The [starting figures in this setup are a $25-per-day minimum per campaign and a $3 to $5 max CPC](https://www.wordstream.com/blog/chatgpt-ads). A bid below $3 can draw a may-not-deliver warning; treat that as a delivery concern, not an excuse to spend more without permission.

A cap keeps the first test at a known cost while you learn which conversations produce conversions. It also makes the decision to scale separate from the decision to launch. Since self-serve opened, a provider cannot hide behind a $200,000 pilot or a $50,000 minimum as the reason you have nothing live. You still need enough budget to learn, and you need a number at which you will keep going or stop.

Say you spend $20,000 a month on Google. A $750 ChatGPT test over 30 days is a bounded way to learn, not a promise of a result. **Check the cap in Ads Manager billing, then find the email or document saying spend rises only with your sign-off.** Treat the first four weeks as learning cost you are comfortable losing. If the budget goes up without your written yes, the provider has confused managing spend with spending it.

![A paper receipt for ChatGPT ad spend stamped with a hard cap](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/fe2da94e-dfcf-4601-8668-b7f359ad3ffa.png)

## 6. Record your Google and Bing baseline before judging ChatGPT

You need a one-page snapshot of your current cost per acquired customer on Google and Bing, plus a consistent way to tag the new traffic. That means UTMs you can distinguish, such as `utm_source=chatgpt` and `utm_source=google`, before the first performance chart arrives. Without a baseline, a good-looking CTR can become the entire argument for continuing to spend.

The traffic may not behave like your existing search traffic. In one early account of about 1,500 users, [GA4 showed 17 seconds of engagement for ChatGPT traffic versus 41 seconds for Google, and conversion rates of 0.21% versus 3.71%](https://searchengineland.com/run-chatgpt-ads-484513). That is one account, not your forecast. It illustrates why clicks alone are a poor verdict on whether a channel is producing customers.

**Ask for the baseline document and UTM naming before you look at a CTR chart.** Check that the provider is comparing the outcome you care about, not quietly switching between clicks, leads and acquired customers as the numbers become convenient. Consistent tags and a pre-test CPA give you something to judge against. No baseline, no verdict.

## 7. Keep a change log that explains the work

Ads Manager has a [native change log for members, permissions, API keys, billing and account changes](https://ppc.land/openai-opens-chatgpt-ads-manager-to-all-us-businesses-with-cpc-bidding/). Your provider should pair the record of what changed with a plain-language account of why. A timestamp is useful; a timestamp plus the decision behind it lets you tell management from motion.

I use the same principle I used with Google Change History. If someone says they adjusted a bid, tested an ad or fixed a tracking issue, I want to find the corresponding action and hear what prompted it. Not every healthy test needs a busy log. A quiet account can be the right call for a period. The problem is an account that is quiet while the monthly report describes a lot of work nobody can locate.

**Request view access and compare the last two weeks of changes with what the provider says they did.** If the only visible entries are billing and user invites, ask where the campaign work happened. The same accountability test runs through [what a fully managed ads service should actually do](https://groas.com/post/fully-managed-google-ads-service-checklist-what-agency-should-do). If nobody can connect actions to reasons, you are paying for reporting, not management.

## 8. Launch ad variations instead of promising them

By day 14, I would expect at least three live ads per intent group, each with its own headline, description and image. Count ads that are approved and serving. A roadmap full of future creative is not a variation test. [Done-for-you packages commonly promise 3 to 15 creative variations, alongside tracking and optimization](https://www.rankpage.com.my/chatgpt-ads-services/); the useful part of that promise is what has actually made it into the account.

The unit is compact: advertiser name and logo, a [headline around 50 characters, description around 100 characters, landing page and square 256 × 256 image](https://blog.hubspot.com/marketing/chatgpt-ads-vs-google-search-ads), with newer carousel and product-feed cards also available. Different angles give the provider something to compare within an intent group. They can see what earns attention, cut a weak version and keep testing the connection between the ad and its page. One ad cannot tell you much about that choice.

**Filter Ads Manager by ad group and count what is live.** Then read the variations. Are they distinct reasons to click, or the same line with a word swapped? Ask which page each one uses and what the provider plans to change when results arrive. Live and rotating beats planned every time.

## 9. Agree on the number that kills the test

Before the first month becomes the second, put the stop condition in one sentence. For example: stop if cost per acquired customer remains above $X after four weeks at $Y spend. Fill in X and Y together before the results are in. Otherwise every disappointing week becomes another request to wait for the next report.

Fit can also kill a test. Ads show on Free and Go plans for logged-in adults and in some logged-out sessions, not on Plus, Pro, Business, Enterprise or Edu, and not near under-18, health or politics content. If the buyers you need are concentrated in inventory you cannot reach, more creative will not fix the mismatch. Nor does paid placement solve a different goal: [buying ads does not buy organic mentions](https://blog.hubspot.com/marketing/chatgpt-ads-vs-google-search-ads). I would rather agree to stop than use one channel to pretend to deliver another.

**Find the written kill criterion and check that it names an outcome, a spend limit and a time limit.** This is also the point to skip outsourcing if you have no site that can take a pixel, no offer that can take a form fill, or no four-week learning budget you are comfortable losing. Fix the site or wait. A capped test still costs money before it teaches you anything. If a provider will not name the number that ends the test, they can keep billing past it.

![Stopwatch on a ChatGPT ads report beside a marked kill line](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/133a7bb4-dabf-4cf1-a4cd-2a8ee49e8349.png)

## 10. Name the person who can answer for the account

You need one human who can open the account on a call, explain what changed, say why CPA moved and tell you what happens next. Not a success manager who forwards your question to a queue. Not a course community. The first nine artefacts are checkable in minutes; this person has to make sense of them when the signals disagree.

**Get their name, a direct Slack or email contact, and their cadence for working in the account.** Then test the arrangement in week two. Ask a specific question about an intent hint or the test conversion. Does the answer begin in Ads Manager, with the relevant setting or event on screen? Or does another deck arrive? That is a fairer test of hands-on management than how polished the kickoff felt.

This is what I would ask of [groas’s done-for-you Search Ads service across Google, Bing and ChatGPT](https://groas.com/pricing): its flat $1,499 monthly fee, work live in your accounts within 24 hours of onboarding, around-the-clock work on bids, keywords and budgets, and a landing page for every ad group should leave things you can inspect. You keep your accounts, pages and data. The same standard applies to anyone you hire. Ask for the login, the test conversion and the kill number before paying for month two.

If I had to begin with one check, though, I would go straight back to **No. 1: your account, your billing, your access**. It is the condition that lets you verify every other promise on this list. Without it, even a convincing answer about the other nine is still just a slide.

## Frequently Asked Questions

### How should an advertiser structure ownership and billing for a ChatGPT Ads account?

The advertiser must own the account under their company email, and the provider should be listed as an invited user. The payment method should be on file for the advertiser, ensuring media spend is billed separately from the management fee.

### What setup is required to verify that ChatGPT ad tracking is working before running an ad campaign?

Tracking requires the OpenAI Pixel and Conversions API to use the same event ID for deduplication. A critical step is firing a test lead or purchase and confirming that the ad click ID, or oppref, survives redirects and appears in Ads Manager.

### How should ad groups be structured in a ChatGPT Ads campaign to ensure relevant matching?

Ad groups should be organized around distinct conversation intents, meaning a freeform context hint describes the specific buyer question the ad meets. This structure is achieved by creating focused groups rather than simply pasting a large list of old search terms.

### What are the requirements for a landing page linked to a ChatGPT ad?

A dedicated landing page must match the conversation intent defined in the ad group's hint and must lead to a page that answers that specific question. The page must also preserve the click ID and be accessible to both OAI-AdsBot and OAI-SearchBot.

### What are the recommended starting budget and bidding caps for testing ChatGPT ads?

A live campaign requires a daily cap agreed upon in writing, with starting figures suggesting a $25-per-day minimum per campaign. The maximum CPC should be between $3 and $5, as bids below $3 can trigger a may-not-deliver warning.

### Why is it necessary to establish a performance baseline on other platforms before judging ChatGPT ad results?

Establishing a baseline of CPA and consistent tracking on Google and Bing is necessary because traffic behavior on this channel may differ significantly from existing search traffic. This allows for a fair judgment against the metrics you care about, such as acquired customers, not just clicks or CTR.

### What elements must be included in a written test 'kill criterion' for a ChatGPT ads campaign?

A clear termination condition must be agreed upon before the first month begins. This condition needs to specify an outcome, a measurable spend limit, and a defined time limit to prevent indefinite spending on a failing test.

## Paid Ads & SEO, Run For You

Businesses pay one flat monthly fee per service, agencies resell it under their own brand, and groas does the work either way.

[See Pricing](https://groas.com/pricing)

## Structured data

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://graph.groas.com/entity/groas","name":"groas","url":"https://groas.com/","description":"groas is a done-for-you marketing service, not software: elite people working with proprietary technology systems run search ads, Meta ads, SEO and AI search for businesses, and white label for agencies, with a dedicated senior strategist owning every account.","logo":{"@type":"ImageObject","url":"https://groas.com/icon-512.png","width":512,"height":512},"sameAs":["https://www.linkedin.com/company/groas/"]},{"@type":"WebSite","@id":"https://groas.com/#website","name":"groas","url":"https://groas.com/","inLanguage":"en-US","publisher":{"@id":"https://graph.groas.com/entity/groas"}},{"@type":"WebPage","@id":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14#webpage","url":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14","name":"10 Things to Check in Your ChatGPT Ads Account by Day 14","description":"Paying someone to run your ChatGPT ads? Check these 10 account, tracking and landing-page artefacts before you pay for another month.","inLanguage":"en-US","isPartOf":{"@id":"https://groas.com/#website"},"about":{"@id":"https://graph.groas.com/entity/groas"},"primaryImageOfPage":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/a446098b-ac69-43fe-9bde-9c807d2c6618.png"},"breadcrumb":{"@id":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14#breadcrumb"},"dateModified":"2026-10-31T06:23:38.316Z","mainEntity":{"@id":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://groas.com/"},{"@type":"ListItem","position":2,"name":"Blog","item":"https://groas.com/blog"},{"@type":"ListItem","position":3,"name":"10 Things to Check in Your ChatGPT Ads Account by Day 14","item":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14#article","url":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14","mainEntityOfPage":{"@id":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"10 Things to Check in Your ChatGPT Ads Account by Day 14","description":"Paying someone to run your ChatGPT ads? Check these 10 account, tracking and landing-page artefacts before you pay for another month.","datePublished":"2026-10-31T06:23:38.264Z","dateModified":"2026-10-31T06:23:38.316Z","image":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/a446098b-ac69-43fe-9bde-9c807d2c6618.png"},"author":{"@id":"https://groas.com/author/alexander-perelman#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"keywords":"account, ads, test, check, chatgpt, chatgpt ads, provider, day","wordCount":2247,"inLanguage":"en-US"},{"@type":"Person","@id":"https://groas.com/author/alexander-perelman#person","name":"Alexander Perelman","jobTitle":"Head Of Product @ groas","description":"My background includes Goldman Sachs and Stanford Computer Science.","image":"https://groas.com/media/blog/4f17cac3a81acc224cc1b2cfab8f94bbc295a3089d13e8421125771ef8d4064f.jpg","worksFor":{"@id":"https://graph.groas.com/entity/groas"},"sameAs":["https://www.linkedin.com/in/alexander-433793253/"]},{"@type":"FAQPage","@id":"https://groas.com/post/10-things-should-exist-in-your-chatgpt-ads-account-by-day-14#faq","mainEntity":[{"@type":"Question","name":"How should an advertiser structure ownership and billing for a ChatGPT Ads account?","acceptedAnswer":{"@type":"Answer","text":"The advertiser must own the account under their company email, and the provider should be listed as an invited user. The payment method should be on file for the advertiser, ensuring media spend is billed separately from the management fee."}},{"@type":"Question","name":"What setup is required to verify that ChatGPT ad tracking is working before running an ad campaign?","acceptedAnswer":{"@type":"Answer","text":"Tracking requires the OpenAI Pixel and Conversions API to use the same event ID for deduplication. A critical step is firing a test lead or purchase and confirming that the ad click ID, or oppref, survives redirects and appears in Ads Manager."}},{"@type":"Question","name":"How should ad groups be structured in a ChatGPT Ads campaign to ensure relevant matching?","acceptedAnswer":{"@type":"Answer","text":"Ad groups should be organized around distinct conversation intents, meaning a freeform context hint describes the specific buyer question the ad meets. This structure is achieved by creating focused groups rather than simply pasting a large list of old search terms."}},{"@type":"Question","name":"What are the requirements for a landing page linked to a ChatGPT ad?","acceptedAnswer":{"@type":"Answer","text":"A dedicated landing page must match the conversation intent defined in the ad group's hint and must lead to a page that answers that specific question. The page must also preserve the click ID and be accessible to both OAI-AdsBot and OAI-SearchBot."}},{"@type":"Question","name":"What are the recommended starting budget and bidding caps for testing ChatGPT ads?","acceptedAnswer":{"@type":"Answer","text":"A live campaign requires a daily cap agreed upon in writing, with starting figures suggesting a $25-per-day minimum per campaign. The maximum CPC should be between $3 and $5, as bids below $3 can trigger a may-not-deliver warning."}},{"@type":"Question","name":"Why is it necessary to establish a performance baseline on other platforms before judging ChatGPT ad results?","acceptedAnswer":{"@type":"Answer","text":"Establishing a baseline of CPA and consistent tracking on Google and Bing is necessary because traffic behavior on this channel may differ significantly from existing search traffic. This allows for a fair judgment against the metrics you care about, such as acquired customers, not just clicks or CTR."}},{"@type":"Question","name":"What elements must be included in a written test 'kill criterion' for a ChatGPT ads campaign?","acceptedAnswer":{"@type":"Answer","text":"A clear termination condition must be agreed upon before the first month begins. This condition needs to specify an outcome, a measurable spend limit, and a defined time limit to prevent indefinite spending on a failing test."}}]}]}
```