admin 
Advertisers can now bring more of their existing first-party audience data into ChatGPT Ads through an expanded partnership between LiveRamp and OpenAI.
The integration allows marketers to use LiveRamp’s RampID to activate customer and prospect audiences in ChatGPT Ads. That includes data from CRM systems, loyalty programs, websites, apps and other first-party sources.
Advertisers can use those audiences to reach or suppress specific customers and prospects, adding targeting options that have long been available across platforms such as Google and Meta.
For marketers already using LiveRamp elsewhere, the update also makes ChatGPT another destination for data they may already be activating across their media plans.
RampID Extends Existing First-Party Data Into ChatGPT
The new audience integration expands a partnership between LiveRamp and OpenAI that initially focused on measurement.
In June, LiveRamp became a measurement partner for ChatGPT Ads, giving advertisers another way to connect ad exposure with downstream conversions. RampID now extends that relationship into audience activation.
LiveRamp describes RampID as its identifier for connecting customer data across advertising platforms and other destinations. For existing customers, ChatGPT can now be added to the places where that data is activated.
The integration is initially available in 11 markets. LiveRamp says it plans to expand into additional markets as ChatGPT Ads becomes available there.
However, LiveRamp isn’t required to use first-party audiences in ChatGPT Ads.
OpenAI also supports Custom Audiences directly through Ads Manager.
Advertisers can also create Custom Audiences directly in OpenAI Ads Manager. They can upload a CSV or TXT file containing email addresses, phone numbers, hashed versions of either, or Google Advertising IDs.
Those audiences can be included or excluded at the campaign level. Advertisers can also apply bid multipliers at the ad group level.
LiveRamp gives its existing customers another way to get that data into ChatGPT. Instead of preparing and maintaining separate audience files, advertisers already using RampID can add ChatGPT as another place to activate those audiences.
Those capabilities make first-party audience targeting in ChatGPT look more familiar to advertisers already using Google and Meta. The similarities, however, don’t necessarily mean the platforms offer the same targeting opportunity.
How Close Does This Put ChatGPT To Google And Meta Targeting?
For advertisers already using Customer Match in Google Ads or Custom Audiences in Meta, much of this will look pretty recognizable.
Advertisers can use their own customer data to decide who they want to reach, who they want to exclude and how much they’re willing to bid for certain audiences. Those are controls marketers have had on Google and Meta for years.
But having the same targeting options doesn’t mean advertisers should expect them to work the same way in ChatGPT.
Google and Meta have years of advertiser and conversion data behind their ad platforms. ChatGPT Ads is still relatively new, and conversation context also plays a role in determining which ads users see.
There is also very little performance data available for Custom Audiences in ChatGPT so far.
That may make suppression a reasonable place to start. A retailer could avoid spending money to acquire someone who recently purchased, for example. A subscription business could exclude current customers from an introductory offer.
Using first-party data to find and bid on specific audiences comes with more unknowns. OpenAI’s audience-size requirements limit how narrowly advertisers can segment their data, and there aren’t public benchmarks showing how these audiences perform against ChatGPT’s other targeting options.
So while the feature list is starting to look more like what advertisers already have elsewhere, there is still a lot to learn about how those audiences perform in ChatGPT.
For advertisers already using the same customer data across several platforms, LiveRamp could make testing that much easier.
Should Advertisers With First-Party Data Test This Now?
For advertisers that already have their first-party data organized and available through LiveRamp, testing ChatGPT Ads may not require much additional setup.
LiveRamp is already pointing to one early example. Allegiance Group & Pursuant (AGP) is implementing the integration for a nonprofit client.
Megan Morris, Director of Integrated Media Solutions at AGP, mentioned in the LiveRamp announcement:
We’re continually optimizing the channels we use to reach our customers, and with ChatGPT Ads becoming an increasingly popular engagement surface for donors, it’s important for us to view the omnichannel impact of our ads across all of our investments.
That may be where this integration makes the most sense early on. An advertiser already using LiveRamp across multiple platforms can add ChatGPT without creating a separate process for its customer data.
But easy access alone isn’t a strong reason to shift budget.
Advertisers still need a reason to believe ChatGPT can add something they aren’t already getting from Google, Meta or other channels. That could be reaching the same customers at a different point in their research, testing whether conversation context changes response rates or finding additional conversions from an audience they already know.
There isn’t much public performance data yet to help answer those questions. That makes a smaller test easier to defend than a meaningful budget shift, particularly when the money would otherwise come from campaigns with established performance.
Advertisers that would need to do significant work just to prepare their first-party data have even less reason to rush. The integration will still be there as more performance data becomes available.
For advertisers ready to test, though, another question comes quickly: what will it cost to reach those first-party audiences?
Will First-Party Audiences In ChatGPT Cost More?
One thing advertisers don’t know yet is whether reaching their first-party audiences in ChatGPT will come at a higher cost.
OpenAI’s documentation doesn’t identify separate pricing for Custom Audiences. ChatGPT Ads supports CPM and CPC buying, along with conversion-optimized campaigns that can be billed on clicks or impressions.
Advertisers can, however, change their bids when someone matches a Custom Audience.
OpenAI allows bid multipliers ranging from 0.1x to 10x at the ad group level. A 2x multiplier, for example, tells the system to increase the maximum bid when someone belongs to that audience.
That gives advertisers control over how much more, or less, they’re willing to bid for people they already know something about. A retailer could bid differently for previous purchasers, while another advertiser might place a higher value on qualified prospects from its CRM.
What we don’t have yet are benchmarks showing how Custom Audiences perform compared with ChatGPT’s other targeting options. OpenAI reports average CPC and CPM in Ads Manager, but it hasn’t published comparative performance data for first-party audiences.
If those audiences do carry higher CPCs or CPMs, cost alone won’t tell advertisers whether they’re paying too much. They’ll also need to look at conversion rates, customer value and whether those audiences produce better results than the targeting they’re already using.
First-Party Data Gives Advertisers More To Test In ChatGPT
Advertisers don’t need LiveRamp to start using their own customer data in ChatGPT Ads. OpenAI already allows them to upload Custom Audiences directly through Ads Manager.
The LiveRamp partnership gives larger advertisers another option. Brands already using RampID can bring those audiences into ChatGPT without managing the platform separately from the rest of their first-party data program.
That makes the announcement more relevant for existing LiveRamp customers, but it also shows how quickly OpenAI is filling out the targeting options available in Ads Manager.
There is still very little performance history behind those options.
LiveRamp says it plans to share results from marketers using the integration. Those results could help answer some of the questions advertisers have now, including how first-party audiences compare with ChatGPT’s native targeting and whether higher bids translate into better business results.
Until then, advertisers have two ways to bring first-party audiences into ChatGPT Ads. What they don’t have yet is enough data to know how much those audiences should change their investment in the platform.
Featured image: Sergey Nivens, Shutterstock
