---
title: "How to Hire a ChatGPT Ads Agency When Your Sales Cycle Runs for Months"
description: "A hiring guide for long-cycle B2B teams: start with the CRM handoff, test for qualified pipeline, and judge ChatGPT Ads before closed-won revenue arrives."
url: "https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months"
image: "https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/65738183-f97f-483b-9571-a0bf104052cb.png"
published: "2026-11-05T05:35:05.262Z"
modified: "2026-11-05T05:35:05.328Z"
---

November 5, 2026 · 10 min read

# How to Hire a ChatGPT Ads Agency When Your Sales Cycle Runs for Months

[](https://groas.com/author/alexander-perelman)

[Alexander Perelman](https://groas.com/author/alexander-perelman)Personally Tested

Head Of Product @ groas

[LinkedIn](https://www.linkedin.com/in/alexander-433793253/)

![Cutaway drawing of a building: a shiny faucet labelled Leads on top, a long pipe down to an empty Revenue tank, broken at the CRM joint where an inspector kneels.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/65738183-f97f-483b-9571-a0bf104052cb.png)

In this article

1. [What sale would your sales team actually accept?](#what-sale-would-your-sales-team-actually-accept)
2. [Can they actually run ChatGPT Ads today?](#can-they-actually-run-chatgpt-ads-today)
3. [How will a response reach the CRM and return as a signal?](#how-will-a-response-reach-the-crm-and-return-as-a-signal)
4. [What should every agency receive before it pitches?](#what-should-every-agency-receive-before-it-pitches)
5. [How do you design a test that can survive three months without revenue?](#how-do-you-design-a-test-that-can-survive-three-months-without-revenue)
6. [How do you decide before any cohort has closed?](#how-do-you-decide-before-any-cohort-has-closed)
7. [Where does groas fit, and what should you fix first?](#where-does-groas-fit-and-what-should-you-fix-first)

A polished lead report is useless if no one can tell you which inquiries became opportunities. Yet that is how most B2B teams hire a ChatGPT Ads agency: compare three creative decks, pick the prettiest mockups and the lowest promised cost per lead, then discover four months later that the clicks have no reliable path into pipeline. I have watched that movie before. I used to tell clients to judge early tests on lead volume. I was wrong.

**Start with the CRM handoff, not the ad creative.** You will probably have to make a keep-or-kill call before you have closed-won revenue; that is normal when your median sales cycle runs 84 days. This guide follows the order that matters: define the sale, check whether an agency can run the channel today, give every finalist the same brief, and set 90-day test rules before spending. If an agency cannot explain how a campaign response becomes a qualified opportunity in your system, it cannot credibly optimize toward pipeline. The lead report can wear a nice jacket. It is still a lead report.

## What sale would your sales team actually accept?

Skip the agency calls until you can write three definitions on one page:

- **Qualified inquiry:** What marketing can filter for, including the right title, company size and business problem.
- **Accepted opportunity:** What sales agrees to work after discovery.
- **Closed-won revenue:** Money in the system, not an encouraging conversation.

Most long-cycle hires go wrong when the brief lumps all three into “leads.” The agency then optimizes toward the cheapest version: form fills. A form fill may be a useful first signal. It is not a sale, and treating it as one makes every subsequent report look more certain than it is.

Put a date next to each definition. The [median B2B SaaS sales cycle is 84 days](https://www.saashero.net/google-ppc/google-ads-management-cac-tracking/), with a mean of 104 days; the range stretches from 14 to 40 days for SMB to 90 to 180-plus days for enterprise. [Google defaults to 30-day attribution](https://www.saashero.net/google-ppc/google-ads-management-cac-tracking/), so a deal that closes after day 30 will not teach Smart Bidding about the sale unless your CRM feeds that outcome back. The machine sees the early form fill. You get paid on what happens later. **Write down both events before anyone proposes a campaign.**

## Can they actually run ChatGPT Ads today?

Ask for a screen share of a live account, not a slide about process. [ChatGPT ads began as a US test on Feb 9, 2026, for Free and Go users only](https://www.campaigncreators.com/blog/how-chatgpt-ads-in-hubspot-turn-ai-search-into-a-measurable-revenue-channel). They appear as a labeled sponsored unit below the answer, never inside it. [Plus, Pro and Business subscribers do not see ads](https://www.campaigncreators.com/blog/how-chatgpt-ads-in-hubspot-turn-ai-search-into-a-measurable-revenue-channel). Part of your B2B buying committee may therefore be invisible to this channel from the start. Ask the agency which part of your audience it expects to reach. If it glosses over the coverage gap, it has not thought hard enough about what you are buying.

Then check buying access. [Self-serve buying runs through Ads Manager at ads.openai.com with a $25-per-day minimum in the US](https://www.webfx.com/blog/ai/chatgpt-ads-cost/), after early six-figure minimums were removed on May 5, 2026. [As of Oct 21, 2026, OpenAI lists 89 countries as available, but access remains billing-entity dependent and in beta](https://www.webfx.com/blog/ai/chatgpt-ads-cost/). Have the finalist explain what it can launch for _your_ billing entity, not what an inventory slide says is available somewhere.

Do not hire around a capability that is still a pilot. [Sponsored Agents, where a click opens a branded AI-agent chat inside ChatGPT, remains a limited US pilot with select advertisers such as Wayfair](https://searchenginewatch.com/openai-tests-chatgpt-ads-that-skip-the-advertisers-website/). **Live search-response management is the work to evaluate; agent execution is not a day-one hiring plan.**

## How will a response reach the CRM and return as a signal?

Creative comes after this conversation. **The measurement handoff is the product you are buying.** Since [Sept 16, 2026, HubSpot has been OpenAI’s first CRM partner](https://www.campaigncreators.com/blog/how-chatgpt-ads-in-hubspot-turn-ai-search-into-a-measurable-revenue-channel), a workable setup is known: HubSpot auto-appends UTMs, creates contacts from lead forms, and can sync lifecycle stages such as SQL, opportunity and customer back to ChatGPT Ads for optimization. If an agency proposes a generic contact form with no lifecycle sync, it is planning to optimize toward form fills and call the result pipeline. Skip that proposal.

![Ad click, inquiry and signed contract connected across a tangled paper trail](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/4ae741cf-6ff7-4121-9a30-4c8d27f365ce.png)

Insist on three owners in writing:

1. **Lead routing:** Who gets the inquiry to sales and owns speed to lead?
2. **Qualification:** Who updates the CRM stage when sales accepts or rejects it?
3. **Signal return:** Who checks that the stage reaches the ad platform daily, rather than at month end?

Do not let “the team” own all three. A named owner makes a broken handoff visible while there is still time to repair it. Timing matters: [conventional GCLID-based uploads more than 90 days after the click are not imported](https://oneppcagency.co.uk/google-ads/offline-conversion-tracking-for-google-ads/), and the [Enhanced Conversions for Leads limit is 63 days](https://oneppcagency.co.uk/google-ads/offline-conversion-tracking-for-google-ads/). If you sell a 120-day deal and wait for closed-won before returning a signal, the platform never learns from it. I import the earliest event sales trusts, usually SQL or accepted opportunity, and keep closed-won as the truth for human decisions. **Pick that trusted event before launch.**

## What should every agency receive before it pitches?

Give two or three finalists the identical one-pager: audience, offer, monthly ceiling, and your definitions of qualified inquiry and accepted opportunity. Add the CRM owners and the time from click to close. If each agency gets a different version of the business, its proposal tells you less about its judgment than about what happened to come up on the call.

Then ask each finalist for its daily operating plan in writing. Who checks what, what can change without approval, and what will appear in the weekly report? [Demand SQLs, pipeline and cost per SQL, not impressions, CTR and CPL alone](https://www.saashero.net/google-ppc/best-no-contract-saas-ppc/). Those traffic metrics can help diagnose a campaign, but they cannot settle whether sales wants the people it produces. If one provider reports on pipeline and another reports on clicks, you are comparing scoreboards, not agencies.

Ask the uncomfortable question directly: **Who owns the problem when ads produce inquiries but sales rejects them?** The honest answer names a process. The agency tightens qualification filters, checks UTMs and lifecycle sync, and pauses spend if follow-up breaks. The evasive answer blames “lead quality” and asks for more budget. I have fired that second type before I hired the first.

That distinction matters when [roughly 40% of attributable conversions happen in the click session and 60% arrive later via direct or branded search](https://www.contexthints.com/guide/chatgpt-ads-roi.html). Platform reporting is also aggregated, without search-term diagnostics. You need someone willing to defend the CRM truth when platform clicks tell a tidier story. Give each finalist the same brief and see who questions the handoff before complimenting the creative.

## How do you design a test that can survive three months without revenue?

Set the budget from auction math, not hope. [Current guidance is a $3-to-$5 max CPC and a $60 default max CPM, with real clearing as low as $25](https://www.webfx.com/blog/ai/chatgpt-ads-cost/). The same guide recommends [$5k to $10k per month for 1,000-plus clicks, $15k to $30k to test CPC and CPM in parallel, and $50k-plus for scaled share of voice](https://www.webfx.com/blog/ai/chatgpt-ads-cost/). Do not buy a large test just because the bigger number looks more committed.

Say you spend $20k per month. At $4 CPC, that buys 5,000 clicks. If 2% become inquiries and 20% of those become accepted opportunities, you will see about 20 opportunities before sales closes a single deal. That is enough to start judging fit. **It is not revenue.** Write down that distinction so nobody turns an early opportunity report into a victory lap.

For a long cycle, judge the test on upstream truth tellers: **cost per qualified lead, opportunity rate, SQL rate, branded search lift and blended MER**. [Optimize toward the deepest event that still fires around 30 to 50 times per month](https://www.contexthints.com/guide/chatgpt-ads-roi.html); use MQL or SQL as a bidding proxy when necessary, and keep closed-won as the human verdict. [Do not compare ChatGPT last-click ROAS side by side with Google data-driven attribution](https://www.contexthints.com/guide/chatgpt-ads-roi.html). They measure different things. A neat comparison would still be a bad decision tool.

Set the sequence before spend starts:

1. **Day one: prove the handoff.** Submit a live test lead and verify that UTMs and click IDs land in HubSpot or Salesforce. Use the [$25-per-day minimum](https://www.webfx.com/blog/ai/chatgpt-ads-cost/) only while proving tracking works, not as the budget for judging audience fit.
2. **During the test: read qualification, not just volume.** Assess fit at the $5k-to-$10k-per-month level, where you have enough clicks to examine qualification rate. Check that sales follows up and updates stages, and that the return signal reaches the platform.
3. **At the review gate: decide with the rules you set.** The practitioner rule I use is a [90-day review gate with a 30-day exit](https://www.saashero.net/google-ppc/best-no-contract-saas-ppc/). A meaningful pipeline signal usually needs 60 to 90 days, but broken tracking or follow-up deserves an earlier exit.

![Bars comparing qualified inquiries and accepted opportunities while revenue remains pending](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/17bec8e6-46d9-4151-b8f6-7494b791aacf.png)

Keep existing Google search running as context, not as a control you did not design. The measurement stack is unglamorous: [pixel plus Conversions API for bidding, UTMs plus CRM for quality, self-reported “How did you hear about us?”, and a geo or time holdout for incrementality](https://www.contexthints.com/guide/chatgpt-ads-roi.html). Google and ChatGPT may both claim credit for the same later branded search. Do not let anyone declare a winner on last-click ROAS. **Decide on qualified-lead cost, opportunity rate, and whether branded search lifts while blended MER holds.**

## How do you decide before any cohort has closed?

There are three decisions, not two. A test can be worth continuing, need a plumbing repair, or deserve a clean stop. Make that call against the rules you wrote before launch, not whichever chart looks best in week eight.

**Continue when the plumbing holds and the math nearly works.** UTMs land on every test lead, lifecycle stages sync back without manual CSVs, and cost per qualified opportunity sits within 30% of what sales can tolerate at your close rate. If you get 30 to 50 SQL-stage fires per month and opportunity rate holds steady through day 60, there is something worth learning even with zero closed deals. I keep spending in that case and let the cohort mature. A structured [pipeline test for a six-month cycle](https://groas.com/post/chatgpt-ads-for-b2b-with-a-six-month-sales-cycle-judge-it-on) helps keep those weekly judgments consistent instead of rewriting the rules when patience runs thin.

**Repair when the handoff breaks, not when the ads look boring.** The signs are familiar: contacts with no source, stages stuck at MQL for weeks, sales working leads three days late, or no import back to the platform. Stop debating creative and pause half the budget until routing, qualification and daily imports work. One SaaS team I worked with fixed exactly this by wiring ad clicks to CRM-qualified pipeline instead of trial starts, as shown in [how that tracking fix generated real pipeline](https://groas.com/post/saas-google-ads-conversion-tracking-fix-pipeline-case-study). Unsexy work. It beat three months of ad testing.

**Stop when audience fit or economics fail.** If your buyers live in Plus, Pro or Business seats that do not see ads, volume will stay thin no matter how good the copy is. If no meaningful stage reaches 30 to 50 fires per month, bidding has little to learn from. If qualified-opportunity cost implies a customer acquisition cost sales cannot repay, kill the test and say so. A clean stop beats a slow bleed into month five.

## Where does groas fit, and what should you fix first?

groas runs Search Ads as one service: [Google plus Bing plus ChatGPT Ads for a flat $1,499 per month](https://groas.com/pricing), media spend separate, month to month, with no cut of your spend. **That structure fits a serious test without charging you a percentage penalty for spending enough to get 1,000-plus clicks.** Bidding, budgets and negatives are worked continuously instead of waiting for a monthly review.

What groas does not do is invent pipeline. You still need HubSpot or Salesforce with lifecycle stages defined, UTMs flowing on day one, and confirmed ChatGPT Ads access for your billing entity. No provider, groas included, can guarantee pipeline or revenue when your cycle runs past 90 days. If your CRM sync is ready and your audience uses Free or Go inventory, coordinated management across Google and ChatGPT keeps the test honest. If your buyers live behind Plus subscriptions or your follow-up is broken, fix that before you pay anyone to run clicks.

**Your first move today takes 30 minutes.** Write the one-pager: your definitions of qualified inquiry and accepted opportunity, your median days from click to close, and who owns routing, qualification and the daily import back to the platform. Send the same page to every finalist and ask for a redline. Hire the agency that rewrites your measurement plan before touching your ad copy. The rest are selling you leads you cannot bank.

## Frequently Asked Questions

### What should a B2B team define before contacting a ChatGPT Ads agency?

Before contacting agencies, write down three definitions on one page: a qualified inquiry that marketing can filter for, an accepted opportunity that sales agrees to work after discovery, and closed-won revenue. Put a date next to each. Lumping all three into 'leads' pushes the agency to optimize toward cheap form fills instead of sales.

### Can ChatGPT Ads reach all B2B buyers?

No. ChatGPT ads began as a US test on Feb 9, 2026, shown to Free and Go users only as a labeled sponsored unit below the answer. Plus, Pro and Business subscribers do not see ads, so part of a B2B buying committee may be invisible to the channel. Ask an agency which part of your audience it expects to reach before hiring.

### How do I check if an agency can actually run ChatGPT Ads for my company?

Ask for a screen share of a live account rather than a process slide, and have the agency explain what it can launch for your specific billing entity. Self-serve buying runs through Ads Manager at ads.openai.com with a $25-per-day minimum in the US, access is billing-entity dependent and still in beta, so inventory slides do not prove anything for your account.

### Why does CRM integration matter more than ad creative when hiring a ChatGPT Ads agency?

The measurement handoff is the product you are buying: if an agency cannot explain how a campaign response becomes a qualified opportunity in your system, it cannot credibly optimize toward pipeline. Since Sept 16, 2026, HubSpot has been OpenAI's first CRM partner, so a workable setup exists with UTMs, contact creation and lifecycle sync back to ChatGPT Ads.

### Which CRM event should be sent back to the ad platform for optimization?

Import the earliest event that sales trusts, usually SQL or accepted opportunity, and keep closed-won as the truth for human decisions. Conventional GCLID-based uploads more than 90 days after the click are not imported and the Enhanced Conversions for Leads limit is 63 days, so a 120-day deal's closed-won would never teach the platform anything.

### What should every agency finalist receive and report on before pitching?

Give two or three finalists an identical one-pager covering audience, offer, monthly ceiling, definitions of qualified inquiry and accepted opportunity, CRM owners, and click-to-close time. Then demand reports on SQLs, pipeline and cost per SQL rather than impressions, CTR and CPL alone, so you compare agencies rather than scoreboards.

### How should a ChatGPT Ads test be judged when the sales cycle runs months?

Judge on upstream truth tellers such as cost per qualified lead, opportunity rate, SQL rate, branded search lift and blended MER. Optimize toward the deepest event that fires around 30 to 50 times per month, and do not compare ChatGPT last-click ROAS side by side with Google data-driven attribution because they measure different things.

### How do I decide whether to continue, fix or stop a ChatGPT Ads test with no closed revenue yet?

Make the call against rules set before launch. Continue when UTMs land, stages sync automatically and cost per qualified opportunity sits within 30% of what sales can tolerate. Repair when the handoff breaks, by pausing half the budget until routing, qualification and daily imports work. Stop when buyers cannot see the ads or the implied acquisition cost is not repayable.

## Run Your Marketing On groas

Plug ChatGPT, Claude or any AI agent into groas and let it do the work of a whole marketing team: free to ask, $49 a month flat with every feature included, and autopilot whenever you’re ready.

[Start Free](https://groas.com/pricing)

[Or have groas run it for you](https://groas.com/done-for-you)

## 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 the marketing infrastructure for your own AI: plug Claude, ChatGPT or any AI agent into groas and run all your marketing from it, for a tiny fraction of the usual cost, with the groas app there whenever you want to see more. It is the layer ChatGPT, Claude and any AI assistant run a business’s online marketing on. Through groas an assistant can do anything you could do natively in Google Ads, Bing Ads, Meta Ads and ChatGPT ads, on your website and in AI search, and, where the owner switches it on, can run connected campaigns by itself, so AI agents do the work a whole marketing team used to (free to read, $49 a month to make changes, delivered as an MCP server). Separately, its done-for-you service has senior people and proprietary quantitative systems run paid ads, SEO and AI search for businesses, and white label for agencies, from $5,000 a month.","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/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months#webpage","url":"https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months","name":"How to Hire a ChatGPT Ads Agency When Your Sales Cycle Runs for Months","description":"A hiring guide for long-cycle B2B teams: start with the CRM handoff, test for qualified pipeline, and judge ChatGPT Ads before closed-won revenue arrives.","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/65738183-f97f-483b-9571-a0bf104052cb.png"},"breadcrumb":{"@id":"https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months#breadcrumb"},"dateModified":"2026-11-05T05:35:05.328Z","mainEntity":{"@id":"https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months#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":"How to Hire a ChatGPT Ads Agency When Your Sales Cycle Runs for Months","item":"https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months#article","url":"https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months","mainEntityOfPage":{"@id":"https://groas.com/post/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"How to Hire a ChatGPT Ads Agency When Your Sales Cycle Runs for Months","description":"A hiring guide for long-cycle B2B teams: start with the CRM handoff, test for qualified pipeline, and judge ChatGPT Ads before closed-won revenue arrives.","datePublished":"2026-11-05T05:35:05.262Z","dateModified":"2026-11-05T05:35:05.328Z","image":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/65738183-f97f-483b-9571-a0bf104052cb.png"},"author":{"@id":"https://groas.com/author/alexander-perelman#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"keywords":"sales, chatgpt ads, ads, chatgpt, agency, before, test, lead","wordCount":2130,"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/how-to-hire-a-chatgpt-ads-agency-when-the-sale-closes-months#faq","mainEntity":[{"@type":"Question","name":"What should a B2B team define before contacting a ChatGPT Ads agency?","acceptedAnswer":{"@type":"Answer","text":"Before contacting agencies, write down three definitions on one page: a qualified inquiry that marketing can filter for, an accepted opportunity that sales agrees to work after discovery, and closed-won revenue. Put a date next to each. Lumping all three into 'leads' pushes the agency to optimize toward cheap form fills instead of sales."}},{"@type":"Question","name":"Can ChatGPT Ads reach all B2B buyers?","acceptedAnswer":{"@type":"Answer","text":"No. ChatGPT ads began as a US test on Feb 9, 2026, shown to Free and Go users only as a labeled sponsored unit below the answer. Plus, Pro and Business subscribers do not see ads, so part of a B2B buying committee may be invisible to the channel. Ask an agency which part of your audience it expects to reach before hiring."}},{"@type":"Question","name":"How do I check if an agency can actually run ChatGPT Ads for my company?","acceptedAnswer":{"@type":"Answer","text":"Ask for a screen share of a live account rather than a process slide, and have the agency explain what it can launch for your specific billing entity. Self-serve buying runs through Ads Manager at ads.openai.com with a $25-per-day minimum in the US, access is billing-entity dependent and still in beta, so inventory slides do not prove anything for your account."}},{"@type":"Question","name":"Why does CRM integration matter more than ad creative when hiring a ChatGPT Ads agency?","acceptedAnswer":{"@type":"Answer","text":"The measurement handoff is the product you are buying: if an agency cannot explain how a campaign response becomes a qualified opportunity in your system, it cannot credibly optimize toward pipeline. Since Sept 16, 2026, HubSpot has been OpenAI's first CRM partner, so a workable setup exists with UTMs, contact creation and lifecycle sync back to ChatGPT Ads."}},{"@type":"Question","name":"Which CRM event should be sent back to the ad platform for optimization?","acceptedAnswer":{"@type":"Answer","text":"Import the earliest event that sales trusts, usually SQL or accepted opportunity, and keep closed-won as the truth for human decisions. Conventional GCLID-based uploads more than 90 days after the click are not imported and the Enhanced Conversions for Leads limit is 63 days, so a 120-day deal's closed-won would never teach the platform anything."}},{"@type":"Question","name":"What should every agency finalist receive and report on before pitching?","acceptedAnswer":{"@type":"Answer","text":"Give two or three finalists an identical one-pager covering audience, offer, monthly ceiling, definitions of qualified inquiry and accepted opportunity, CRM owners, and click-to-close time. Then demand reports on SQLs, pipeline and cost per SQL rather than impressions, CTR and CPL alone, so you compare agencies rather than scoreboards."}},{"@type":"Question","name":"How should a ChatGPT Ads test be judged when the sales cycle runs months?","acceptedAnswer":{"@type":"Answer","text":"Judge on upstream truth tellers such as cost per qualified lead, opportunity rate, SQL rate, branded search lift and blended MER. Optimize toward the deepest event that fires around 30 to 50 times per month, and do not compare ChatGPT last-click ROAS side by side with Google data-driven attribution because they measure different things."}},{"@type":"Question","name":"How do I decide whether to continue, fix or stop a ChatGPT Ads test with no closed revenue yet?","acceptedAnswer":{"@type":"Answer","text":"Make the call against rules set before launch. Continue when UTMs land, stages sync automatically and cost per qualified opportunity sits within 30% of what sales can tolerate. Repair when the handoff breaks, by pausing half the budget until routing, qualification and daily imports work. Stop when buyers cannot see the ads or the implied acquisition cost is not repayable."}}]}]}
```
