ChatGPT Ads Campaign Structure: A Complete 2026 Guide

Henry Purchase

Co-Founder

Most ChatGPT Ads accounts underperform because of structure, not creative.

Wrong campaign type for the goal. Wrong bidding model for the funnel stage. One big ad group instead of three small ones. Budget split evenly instead of by margin. The platform can't optimise what isn't organised, and an unstructured account burns budget on the wrong placements before you ever know it.

This guide covers how to structure ChatGPT Ads campaigns properly in 2026. The two campaign types you can run, the three bidding objectives available, the four-level hierarchy that ties it all together, and concrete examples for DTC, B2B SaaS, and service businesses.

If you're new to ChatGPT Ads entirely, start with our pillar guide on what ChatGPT Ads are first. This post assumes you know the basics and want to set up campaigns the right way.

Want a specialist team to structure your ChatGPT Ads account from scratch? Book a free 30-minute discovery call with Focal. We've structured accounts for DTC eCommerce, B2B SaaS, and service businesses since the channel opened.

TL;DR: What You'll Decide

Five decisions. They all stack into a clean structure that the platform can optimise. Get them right at the start and the campaign improves week over week. Get them wrong and you'll be re-doing the structure in three months when the data shows what's broken.

The ChatGPT Ads Campaign Hierarchy

ChatGPT Ads use the same four-level hierarchy as Google Ads and Meta Ads. If you've structured campaigns on either of those platforms, the mental model is identical.

Account. One per business. Houses billing, users, and all campaigns.

Campaign. The strategic container. The campaign decides three things: the type (standard or product feed), the bidding objective (CPM, CPC, or cost per conversion), and the total budget. Most accounts run 3 to 8 campaigns concurrently, each for a different funnel stage or product line.

Ad group. The targeting container inside a campaign. Each ad group defines who or what the ads inside it run against (audience segment, product filter, conversational topic). Most campaigns have 2 to 5 ad groups.

Ad. The creative unit. Each ad has its own headline, description, image, and landing page URL. Most ad groups run 2 to 6 ad variants for testing.

Decisions made at higher levels cascade down. The bidding objective set at the campaign level applies to every ad group and ad inside that campaign. The targeting set at the ad group level applies to every ad inside that ad group. Structure is hierarchical, not flat.

The Two Campaign Types

There are two campaign types live on ChatGPT Ads in 2026.

Standard campaigns

Standard campaigns are the default format. You write each ad manually: headline, description, image, landing page URL. The campaign serves the ads you build against the conversational topics, geographies, and audiences you target.

Best for:

• B2B SaaS, where each offer needs its own positioning
• Professional services (agencies, consultants, accountants, coaches)
• Service businesses with a small set of high-value offerings
• DTC brands with fewer than 1,000 SKUs
• Brand awareness campaigns where you control the creative narrative

Setup time: Quick. A single standard campaign with one ad group and three ads can launch in under an hour.

Optimisation pattern: Manual. You're testing specific creative against specific audience segments and iterating weekly.

Product feed campaigns

Product feed campaigns generate ads automatically from an uploaded product catalogue. You upload a structured feed (titles, descriptions, images, prices, availability), set filters that define which products are eligible per ad group, and the platform creates and serves ads from that data.

Best for:

• Retailers with 1,000 or more SKUs (this is the platform minimum)
• DTC eCommerce brands with broad or seasonal catalogues
• Multi-brand retailers
• Retailers already running Google Shopping (existing feed mapping transfers)

Setup time: Longer initial setup (3 to 5 hours for SFTP configuration and feed validation). Faster ongoing iteration because individual product ads don't need manual creative.

Optimisation pattern: Catalogue-driven. You optimise the feed (titles, descriptions, images, attributes) and the ad groups (which products are eligible), not individual ads.

For the full walkthrough of how product feed campaigns work, see our ChatGPT product feed ads guide.

When to use which (and when to use both)

Most accounts don't need to choose. They run both side by side.

• A retailer with 5,000 SKUs runs product feed campaigns for the catalogue, and standard campaigns for brand awareness and seasonal promotions.
• A B2B SaaS company with 8 product modules runs standard campaigns for each module, plus a brand awareness standard campaign.
• A service business with 4 packages runs standard campaigns for each package.

The decision is not "which type" but "which campaign type per funnel stage." Most multi-product accounts end up with both types in the same account, each doing what it's best at.

The Three Bidding Objectives

ChatGPT Ads offers three bidding objectives in 2026. Each one optimises for a different outcome, which means each one suits a different funnel stage.

CPM (cost per thousand impressions)

You pay for impressions, not clicks. CPM bidding tells the platform to maximise reach within your budget.

What it optimises for: Impressions. The platform serves your ad to as many qualified users as your bid and budget allow.

Best for: Top-of-funnel brand awareness campaigns where the goal is reach, not direct response. New product launches, category awareness, and brand campaigns that complement a separate direct-response engine.

Pricing in 2026: Typically $5 to $25 per 1,000 impressions, varying by category and conversational context. Premium B2B and finance categories land at the higher end.

The honest caveat: CPM is the easiest objective to misuse. Most ChatGPT Ads accounts don't need a CPM campaign in their first 90 days. The channel rewards intent-driven optimisation, and CPM doesn't give the platform a conversion signal to optimise against. Run CPM campaigns only after you've built conversion data with CPC and cost-per-conversion campaigns first.

CPC (cost per click)

You pay each time someone clicks your ad. CPC bidding tells the platform to maximise clicks within your budget.

What it optimises for: Clicks. The platform serves ads to users likely to click through to your site.

Best for: Mid-funnel campaigns where the goal is driving qualified traffic to a landing page. This is the default objective for most new ChatGPT Ads accounts and the right starting point for almost every advertiser.

Pricing in 2026: Typically $0.50 to $5 per click. B2B SaaS, finance, and high-AOV ecommerce land at the higher end.

Why it's the right starting point: CPC bidding gives you a clean cost signal (dollars per qualified click) without requiring conversion data to optimise. New accounts launch with CPC, gather enough conversion data over the first 30 to 60 days, and then graduate specific campaigns to cost-per-conversion bidding once the platform has signal to work with.

Cost per conversion (CPA)

You bid on conversions, not clicks or impressions. The platform optimises spend toward the users most likely to complete the conversion event you've defined.

What it optimises for: Conversions, as defined by the conversion events you've set up in your pixel and tracking setup.

Best for: Bottom-of-funnel campaigns with established conversion data. Retargeting, cart recovery, repeat-purchase campaigns, lead-form-focused B2B SaaS campaigns once you have at least 30 to 50 conversions across your account.

Pricing in 2026: Variable. You set a target cost per conversion. The platform tries to deliver conversions at or below that target, and adjusts bids in real time per conversation.

The prerequisite: Cost-per-conversion bidding doesn't work without conversion data. The platform needs at least 30 to 50 conversions in the recent past to predict which users to bid for. New accounts launching with no conversion history should run CPC for the first 30 to 60 days, build the data, and then graduate specific campaigns to cost per conversion as the conversion volume justifies it.

Which objective to use when

The right structure for most accounts in 2026 is: start everything on CPC. Build conversion data. Graduate specific campaigns to cost per conversion once they hit 30+ conversions in the last 30 days. Add CPM campaigns last, only if reach is a strategic priority alongside the direct-response engine.

Real Campaign Structure Examples

Abstract structure advice only goes so far. Here's what working campaign structure actually looks like for three common business types.

Example 1: DTC eCommerce retailer (5,000 SKUs)

Account structure:

• Campaign 1: Brand awareness (Standard, CPM)
  • One ad group, top-of-funnel conversations
  • 3 hero ads featuring brand story
• Campaign 2: Best sellers (Product Feed, CPC)
  • Ad Group A: High-margin products (filter: margin > 50%)
  • Ad Group B: Mid-margin products (filter: 25-50% margin)
  • Ad Group C: Sale and clearance (filter: sale_status = true)
• Campaign 3: New arrivals (Product Feed, CPC)
  • Ad Group A: Last 30 days launches
  • Ad Group B: Last 90 days launches
• Campaign 4: Cart recovery (Standard, Cost per Conversion)
  • Ad Group: Retargeting cart abandoners
  • 2 ads with discount creative variants

Total: 4 campaigns, 8 ad groups, ~16 ads. Budget split: 50% feed campaigns, 30% cart recovery, 20% brand awareness.

Example 2: B2B SaaS company (5 product modules)

Account structure:

• Campaign 1: Module A - Demo (Standard, CPC)
  • Ad Group: Conversations about Module A's use case
  • 4 ads testing different value propositions
• Campaign 2: Module B - Demo (Standard, CPC)
  • Same structure as Module A
• Campaign 3: Free trial signup (Standard, Cost per Conversion)
  • Ad Group: Lower-funnel research conversations
  • 3 ads focused on signup creative
• Campaign 4: Pricing page traffic (Standard, CPC)
  • Ad Group: Comparison and pricing conversations
  • 2 ads emphasising specific pricing differentiators

Total: 4-6 campaigns, 4-6 ad groups, ~12-18 ads. Budget split: 60% module demos (CPC), 30% free trial (CPA), 10% pricing page.

Example 3: Service business (agency or consultancy with 4 packages)

Account structure:

• Campaign 1: Discovery call (Standard, Cost per Conversion)
  • Ad Group: All packages, broad conversational targeting
  • 3 ads emphasising the free 30-min call
• Campaign 2: Package-specific lead gen (Standard, CPC)
  • Ad Group A: Package 1 conversations
  • Ad Group B: Package 2 conversations
  • Ad Group C: Package 3 conversations
  • Ad Group D: Package 4 conversations

Total: 2 campaigns, 5 ad groups, ~12-15 ads. Budget split: 70% discovery call CPA, 30% package CPC.

Service businesses typically don't need a brand awareness campaign in the first 12 months. The discovery call is the conversion, and the package-specific campaigns drive qualified traffic that converts into that call.

Best Practices for Structuring ChatGPT Ads Campaigns

Six things to get right.

1. One objective per campaign

Don't mix bidding objectives inside a single campaign. Each campaign sets one objective. If you want to test CPC against cost per conversion for the same offer, run two parallel campaigns, not one.

2. Split by margin, not by category

The instinct on a new product feed campaign is to split ad groups by product category (e.g. Men's, Women's, Kids'). The smarter split is by margin. High-margin products get a high bid in their own ad group. Low-margin or sale stock gets a lower bid in a separate ad group. Same total budget, very different ROAS.

3. Keep ad groups focused

An ad group with 50 products and 2 ads can't optimise. An ad group with 200 products and 8 ads can. The platform needs enough variation within an ad group to identify which combinations work, but not so much variation that the signal gets noisy. The sweet spot for most accounts is 20 to 200 products (or 1 audience segment) per ad group, with 3 to 8 ad variants.

4. Stagger your campaign launches

If you're launching 4 campaigns from a new account, don't launch all 4 on day 1. Launch the highest-confidence campaign first (usually CPC against your best-known audience), let it gather data for 7 to 14 days, then launch the next campaign. The platform learns better when it can attribute early signals cleanly.

5. Use the right naming convention from day one

The Ads Manager doesn't enforce naming conventions, but you'll lose your mind in three months without one. The convention we use across Focal's client accounts:

[Campaign Type] - [Objective] - [Funnel Stage] - [Geo/Audience]

Example: PFA - CPC - Mid Funnel - US-Best Sellers

This compresses the most important campaign attributes into the name, so the Ads Manager dashboard becomes scannable instead of confusing.

6. Set up conversion tracking before structure

This is upstream of structure but worth repeating. Without working conversion events, none of the bidding objectives above (except CPM) can optimise. See our pixel and conversion event setup guide before you structure campaigns.

Common Campaign Structure Mistakes

1. Running cost-per-conversion bidding with no conversion history

Cost-per-conversion bidding needs at least 30 to 50 conversions in the last 30 days before it works. Accounts that launch with cost per conversion as the default bidding model burn budget while the platform tries to optimise toward something it can't reliably predict. Always start with CPC. Graduate specific campaigns to cost per conversion once data justifies it.

2. One mega-campaign with everything inside it

The most common structural mistake. The account has one campaign containing every product, every audience, every offer. The platform can't optimise. Budget flows toward whatever happens to convert first, regardless of whether it's the highest-margin product. The right structure is multiple smaller campaigns split by funnel stage and product line.

3. Mixing standard and product feed in the same campaign

You can't. Each campaign is one type. But operators sometimes try to put standard creative inside a product feed campaign by uploading static product feeds with custom creative overlays. This breaks the optimisation pattern of feed campaigns. Keep them separate.

4. Equal budget splits across all campaigns

Splitting $10,000 evenly across 4 campaigns means each campaign gets $2,500. That's often below the data threshold needed for the platform to optimise. Better: concentrate budget on your 2 highest-confidence campaigns and let lower-confidence campaigns run on smaller budgets until they prove themselves.

5. Pausing campaigns within the first 7 days

The platform needs 7 to 14 days to find which conversations match your offer. Pausing a campaign on day 3 because it hasn't converted yet aborts the learning phase. Give every campaign the full learning window before you decide whether it works.

6. Never re-structuring

Account structure that worked at $5,000/month doesn't work at $50,000/month. As budget scales, the structure should evolve. New campaigns, finer ad group splits, more targeting layers. Plan to revisit structure quarterly.

How Focal Structures Accounts

Focal is the done-for-you ChatGPT Ads agency that builds account structure from scratch for retailers, B2B SaaS companies, and service businesses. The structuring work breaks into four parts:

Funnel mapping. Before any campaigns get built, we map the customer journey end to end. What does prospect awareness look like in your category? What's the qualified-lead definition? What's the revenue event? The funnel decides the campaigns, not the other way around.

Campaign architecture. We design the campaign hierarchy (which campaigns, which objectives, which ad groups, which creative variants) based on the funnel map and your existing conversion data. For DTC retailers with product catalogues, we use the Bulk Ad Launcher to build feed-based ad groups at scale without manual setup.

Cross-channel integration. Most clients also run Google Ads and Meta Ads. Focal's Google Ads and Meta Ads connectors pipe existing conversion data and audience signals into ChatGPT Ads campaigns, so your account starts with signal instead of from zero.

Ongoing structure reviews. Every quarter we audit campaign performance, split or merge ad groups, graduate campaigns to cost-per-conversion bidding as data justifies, and update the structure as budget scales.

If you want a specialist team running ChatGPT Ads end to end with proper structure from day one, book a free 30-minute discovery call. The call is free and we'll tell you straight whether your account is ready for the channel.

Three other ways to get started without booking a call:

• Try the free ChatGPT Ads tools. No signup, no email gate.
• Download the free ChatGPT Pixel Helper Chrome extension. Verify pixel setup before launching any campaign structure.
• Join the FutureProof community. 500+ operators sharing what's working on ChatGPT Ads in real time, including structure questions.

Frequently Asked Questions

What is the campaign structure for ChatGPT Ads?

ChatGPT Ads use a four-level hierarchy: Account → Campaign → Ad Group → Ad. Each campaign sets the campaign type (standard or product feed), the bidding objective (CPM, CPC, or cost per conversion), and the budget. Each ad group sets the targeting filters within the campaign. Each ad is one creative variant.

What are the two ChatGPT Ads campaign types?

Standard campaigns and product feed campaigns. Standard campaigns use manually written ad creative (headline, description, image, URL). Product feed campaigns generate ads automatically from an uploaded product catalogue and are built for retailers with 1,000 or more SKUs.

What are the three ChatGPT Ads bidding objectives?

CPM (cost per thousand impressions), CPC (cost per click), and cost per conversion (also called CPA). CPM optimises for reach, CPC optimises for traffic, and cost per conversion optimises for completed conversion events. The right choice depends on funnel stage and conversion data history.

Should I start with CPC or cost per conversion bidding?

Start with CPC for at least the first 30 to 60 days. Cost-per-conversion bidding requires conversion history (typically 30 to 50 conversions in the recent past) for the platform to optimise effectively. New accounts that launch on cost per conversion without that history tend to burn budget. CPC is the right default while the platform learns.

How many campaigns should a ChatGPT Ads account have?

Most accounts run 3 to 8 campaigns concurrently, split by funnel stage and product line. DTC eCommerce accounts with broad catalogues often run more (4 to 10). B2B SaaS and service businesses typically run fewer (2 to 5). The right number depends on budget. Each campaign needs enough budget to gather signal, which usually means $1,500 to $5,000 minimum per campaign per month.

Can I run both standard and product feed campaigns in the same account?

Yes, and most retailers should. The two types serve different purposes. Product feed campaigns handle catalogue-driven acquisition. Standard campaigns handle brand awareness, seasonal promotions, retargeting, and any offers that need custom creative. The two run side by side in the same account.

What is CPM bidding on ChatGPT Ads?

CPM (cost per thousand impressions) bidding charges advertisers per 1,000 ad impressions. The platform optimises for reach, serving the ad to as many qualified users as the bid allows. Typical CPM ranges from $5 to $25 in 2026, varying by category. CPM is best for top-of-funnel brand awareness campaigns where conversions aren't the primary goal.

When should I switch a campaign from CPC to cost per conversion?

When the campaign has at least 30 conversions in the last 30 days. Below that threshold the cost-per-conversion algorithm can't predict reliably. Once you cross 30 conversions, you can graduate the campaign to cost-per-conversion bidding and let the algorithm optimise for the conversion event directly. Most accounts have one or two campaigns ready for this graduation within the first 60 to 90 days.

How does Focal structure ChatGPT Ads accounts for new clients?

Focal builds account structure based on funnel mapping, existing conversion data, and budget. For DTC retailers, we run feed-based campaigns for the catalogue plus standard campaigns for brand and promotions. For B2B SaaS, we run standard campaigns per module plus a cost-per-conversion campaign for free trial signups. For service businesses, we run a discovery-call CPA campaign plus package-specific CPC campaigns. Book a discovery call to see how we'd structure your account.