ABM on ChatGPT Ads, when your list is big enough

Why most ABM lists can't run on ChatGPT

ABM on LinkedIn or Meta looks like this. Pick 300 target accounts. Pull the buying group at each one. Sync a few small audiences, one per role, and run different ads to each.

ChatGPT Ads breaks that model. A custom audience needs at least 25,000 matched users before you can use it for inclusion targeting or bid adjustments. OpenAI recommends 100,000+. Anything smaller shows "Ready, Exclusion only".

So do the maths before you build anything.

The sizing maths

Multiply accounts by buying-group contacts per account:

→ 300 accounts × 8 contacts = 2,400 people. Exclusion only. → 1,000 accounts × 10 contacts = 10,000 people. Exclusion only. → 3,000 accounts × 10 contacts = 30,000 people. Possible, depending on how many match. → 10,000 accounts × 10 contacts = 100,000 people. OpenAI's recommended size.

And that's people on your list, not matched users. Not every contact matches, so the list has to be well above 25,000 to clear it.

If you run a focused tier-one programme of a few hundred accounts, ChatGPT can't target it. That isn't a ContactLevel limit or a match rate problem. Even a perfect match on 2,400 people is 2,400.

The play for large ABM programmes

If you work a list of thousands of accounts, ChatGPT can carry the buying group. Run it differently from LinkedIn though.

On LinkedIn you split the buying group into per-role audiences. On ChatGPT each role audience would fall under 25,000. So keep one buying-group audience and split by conversation instead. The CFO asks ChatGPT about cost and payback. The engineer asks about integrations and security reviews. The champion asks how to fix the problem. Build an ad group for each, with context hints and copy for that role's questions.

Same audience. Different conversations. Role-specific ads.

The other option is a bid layer. Run a broader persona campaign and add the ABM audience as a bid multiplier, anywhere from 0.1x to 10x at ad-group level. The highest matching multiplier wins, and multipliers don't change who's eligible. But bid adjustments need the same 25,000 matched, so this only helps when the ABM list is already big.

How to set it up with ContactLevel

1. Size it first. Accounts × contacts per account. If the answer is under 25,000, skip to the fallback below.
2. Build the buying groups. Import target accounts from HubSpot or Salesforce, then use Contact Search to add every buying-group role at each account. ContactLevel enriches each contact with the personal identifiers ChatGPT matches on.
3. Sync and check. Connect your ChatGPT Ads account and the audience syncs automatically. Processing takes about 20-30 minutes. If it says Exclusion only, the list is too small for inclusion.
4. Build the campaign. Include the buying-group audience. Exclude customers and open opportunities. The remaining eligible audience must still clear 25,000 matched after the exclusions.
5. One ad group per role. Hints and copy for each role's questions. Max bids sit at ad-group level, so you can pay more for the conversations closest to a decision.
6. Report by account. OpenAI's reporting is aggregated and never names anyone. Add UTM parameters and the ContactLevel pixel shows which contacts at which target accounts visited.

The fallback for everyone else

Run buying-group ABM where small audiences work: LinkedIn and Meta.

Then use ChatGPT for two things:

→ Exclusions. Sync customers and open opportunities so category campaigns don't pay for them. No size floor. → Broad persona audiences. A market-wide list of your buyer titles clears 25,000 more easily, and your target accounts sit inside it anyway.

Only Free and Go users see ChatGPT ads. Enterprise buyers on company Business or Enterprise seats won't see them, whatever list you use. Another reason to keep the core of ABM on other channels.

For how ChatGPT Ads works, start with ChatGPT Ads for B2B.

Frequently asked questions

Can you run ABM on ChatGPT Ads?

Only with a large list. Inclusion and bid adjustments need 25,000 matched users, so it works for programmes with thousands of target accounts, not a few hundred.

How many accounts do I need?

Roughly 3,000 accounts at 10 contacts each gets the list to 30,000 people. Whether that clears 25,000 matched depends on how many contacts match.

Can I build one audience per buying-group role?

Only if each role audience clears 25,000 matched on its own, which almost never happens. Use one buying-group audience and split roles with ad groups and context hints.

Can I use a small ABM list as a bid adjustment?

No. Bid adjustments need the same 25,000 matched minimum as inclusion. A smaller list can only be used to exclude.

Can I see which target accounts engaged?

Not inside ChatGPT Ads. OpenAI only reports aggregates. The ContactLevel pixel with UTM parameters shows visits down to the contact and account.

What should a 300-account ABM team do on ChatGPT?

Exclude customers and open deals from category campaigns, run a broad persona audience if your market is big enough, and keep buying-group targeting on LinkedIn and Meta.

→ Related: LinkedIn ABM Buyer Group Targeting, ChatGPT Customer Exclusion, ChatGPT Job Title Targeting