ChatGPT Ads: RampID Custom Audiences in 2026

By 2026, brands putting budget toward ChatGPT ads will be able to build incredibly granular custom audience segments using persistent identifiers like RampID. This isn’t just another version of cookie-based targeting. It’s a total replacement that uses your own consented, first-party data to find people in a more precise and privacy-forward way. This guide will walk you through exactly how to get these powerful new segments operational inside the ChatGPT advertising interface.

Step 1: Gaining Access and Initial Setup for RampID Integration

Before you can even think about building custom audiences with RampID for your ChatGPT campaigns, you need to confirm your ad account is actually enabled for it. It’s almost never on by default, so don’t assume you can just jump in. For advanced identity tools like this, most platforms, including ChatGPT’s ad system, make you request access or flip a specific switch.

1.1 Verify Platform Integration Status

Log into your ChatGPT Ad Manager and look at the left-hand navigation panel for “Settings”. Inside that menu, find a section called “Integrations” or maybe “Data Partnerships”. You should see a list of data providers. You’re looking for LiveRamp, and you need its status to say “Active.” If it’s not active or isn’t listed, you’ll have to get in touch with your account rep or the ChatGPT Ads support team to get it turned on. This usually means signing a data processing addendum that specifies how your data will be handled between the platforms.

1.2 Establish Data Sharing Permissions

Just because LiveRamp is integrated doesn’t mean you can share data yet. You have to give explicit permission. Go to “Account Settings” inside the Ad Manager, then click on “Data Privacy & Permissions.” Look for a toggle labeled “Enable LiveRamp Data Sharing” or something similar. Make sure it’s switched “On.” Skipping this step is a classic mistake. If you don’t enable it, your uploaded data will not be processed for RampID matching, and your efforts will be completely wasted. It’s a frustratingly common oversight that can stall a launch for weeks.

Step 2: Preparing Your First-Party Data for Upload

Your RampID match rate lives and dies by the quality and format of your first-party data. You have to be obsessive in this step if you want to achieve high match rates and build audiences that actually perform.

2.1 Data Cleansing and Normalization

Before you upload anything, you have to clean your CRM data. This means standardizing fields like email, phone, and address. For emails, convert them all to lowercase and remove any spaces. For phone numbers, strip out parentheses and dashes, and ensure you have consistent country codes. For physical addresses, parse them into separate columns (street, city, state, zip) and use a standard format like “ST” instead of “Street.” If your data is messy, with inconsistent capitalization or mixed phone number formats, the matching algorithm will simply fail to find connections. A recent Nielsen study on this topic showed that normalized data sets can increase match rates by as much as 25% because the algorithms aren’t tripping over simple formatting errors.

2.2 Hashing Your Data

For security and privacy, you must hash all personally identifiable information (PII) before it leaves your system. ChatGPT’s Ad Manager and LiveRamp use the SHA256 hashing standard. You’ll need to apply this to your columns of email addresses and phone numbers. Most good marketing automation platforms have this function built-in. If yours doesn’t, there are plenty of open-source libraries like Python’s hashlib that can do the job. Never upload raw PII. The platform will reject it, and you’d be risking a major compliance headache.

2.3 Formatting Your Upload File

Once your data is clean and hashed, put it into a CSV file. The file needs specific column headers that the ChatGPT Ad Manager expects for a RampID match. You’ll almost certainly need columns for hashed_email and hashed_phone, but you should also include first_name, last_name, address_line_1, city, state, and zip_code. While a hashed email is often enough for a match, providing more fields like name and address gives LiveRamp’s identity graph more signals to work with, which can seriously improve your match rate. Also, make sure your final CSV is saved with UTF-8 encoding so you don’t corrupt any characters on upload.

Step 3: Uploading Data and Creating Custom Audiences

With your data prepped and formatted, you can finally upload it to the ChatGPT Ad Manager to build your segments.

3.1 Working through to Audience Creation

Inside the ChatGPT Ad Manager, click on “Assets” from the main left-hand menu, and then choose “Audiences.” This is your audience library. Look for a big button that says “Create Audience” or “New Audience” and click it. This starts the process.

3.2 Selecting Audience Type and Uploading Your File

A popup will ask you what kind of audience you want to build. Choose “Custom Audience,” and from the options that appear, select “Customer List (RampID Enabled).” This selection flags your audience for LiveRamp’s identity resolution process. You’ll then get a prompt to upload your CSV. Click “Upload File,” find the CSV you just prepared, and select it. The system will do a quick scan to make sure the file structure and headers seem correct.

3.3 Mapping Data Fields

After the upload, you’ll see a data mapping screen. Here, you have to confirm that your CSV columns are matched to the right fields in the system. For example, you need to check that your hashed_email column is mapped to the platform’s “Hashed Email” field. Get this mapping wrong and your matches will fail, so double-check everything. The interface will usually flag any required fields you’ve missed with a red outline or an error message, blocking you from continuing until it’s fixed.

3.4 Configuring Match Settings and Privacy Controls

Now you tell LiveRamp how to process your data. You’ll see an option for “Match Precision vs. Reach.” Higher precision will demand more confident matches from LiveRamp, giving you a smaller but cleaner audience. Higher reach will be more lenient, creating a bigger audience but with more potential false positives. For a new campaign, a balanced approach is usually smart. If your budget is tight, though, lean toward precision to make sure every impression has the best chance of counting. You should also review the privacy controls to make sure they align with your company’s policies, even if they’re pre-set by your service agreement.

Step 4: Monitoring Audience Match Rates and Activation

Uploading the data is just the beginning. Now you have to watch the match process and actually activate the segments in your campaigns.

4.1 Tracking Match Progress and Reviewing Match Rates

Once you submit your data, LiveRamp starts its work, matching your hashed data against its identity graph to find RampIDs. The process usually takes between a few hours and a full day, depending on your list size and system load. Back in the “Audiences” section of the Ad Manager, your new audience will have a status like “Processing.” When it’s done, it will switch to “Ready,” and you’ll see two key numbers: the “Match Rate” percentage and the final “Audience Size.” You’re looking for a match rate between 40% and 70%, though the final number depends on your data quality and customer demographics. If you see a match rate below 30%, you need to go back to step 2 and re-evaluate your data cleansing or add more data points to your upload file.

4.2 Activating Your Custom Audience in Campaigns

To put this new audience to work, go to the “Campaigns” section and either create a new campaign or edit an existing one. When you get to the “Targeting” module in the campaign setup, find the “Audiences” subsection and choose “Custom Audiences.” Your new RampID segment should be listed there by the name you gave it. Select it. You can then layer on other targeting like demographics or locations, but be careful. Don’t dilute the precision you just worked so hard to get from the RampID segment by adding overly broad additional targeting.

4.3 Performance Monitoring and Iteration

Once the campaign is live, you need to watch its performance closely. Focus on metrics like click-through rate (CTR), conversion rate, and ROAS for the campaigns using these RampID segments. Compare them to your other campaigns. If a segment isn’t performing, refine it. This could mean creating more niche lists from your first-party data (like high-value customers vs. low-value), adjusting the match precision settings on upload, or segmenting your source list by customer lifetime value (CLTV) from the start. A 2025 IAB report found that advertisers who constantly refine their custom audiences saw an average 15% improvement in conversion rates within six months of activation. This isn’t set-and-forget. For more on improving efficiency, see how AI Ad Optimization can drop CPA by 15%.

Using RampID for ChatGPT ads gives you a serious advantage in targeting precision and privacy compliance. It requires careful data prep, a clear understanding of the platform’s settings, and constant performance monitoring. The real wins come from continuously refining your audiences based on live campaign data. For example, understanding how tools use contextual memory for ad personalization can give you new ideas for segmentation, and for anyone still running on Meta, these Facebook Direct Response strategies are still essential knowledge.