Digital marketers constantly evaluate new channels to see where consumer attention is shifting. When OpenAI quietly rolled out its advertising platform beta, our team at Sales Factory decided to test it firsthand rather than relying on speculation or product announcements.
We launched a live campaign between June 12 and June 23, 2026, spending $500 to promote one of our research reports through OpenAI Ads. The campaign was scheduled to run for two weeks, but high demand burned through the full budget three days early.
Our goal wasn’t massive lead generation. We kept the budget small as a learning exercise to understand the platform, set early benchmarks, and find practical insights before recommending it to clients.
Here is what we learned from setting up the platform, running the campaign, and analyzing the results.
Setting Up the Campaign: The Quirks and Capabilities
Because the platform is still in beta, it has a few architectural quirks alongside some strong technical features.
- Agency vs. Client Logistics Right now, OpenAI Ads Manager is built for self-advertisers. Agencies cannot create master accounts to manage multiple clients. To run a campaign for a client, they must create their own Ads Manager account and invite your team as users. Onboarding takes an extra step, but user management itself is simple; adding administrators and managers was straightforward once access was granted.
- Platform Simplicity & Billing The interface is clean and simple. Setup required only basic info like an account name, time zone, and a logo upload. Billing currently requires a credit card; invoice billing was not an option during our test. Our analytics team was glad to find built-in API connectivity, showing OpenAI is already planning for automated reporting and integrations.
- Campaign Building Blocks Launching a campaign is intuitive. The “Create” dropdown lets you build campaigns, ad groups, ads, or bulk-upload assets. We ran a standard campaign optimized for Clicks.
At the time, OpenAI also supported Reach campaigns, while Conversion campaigns were marked “Coming Soon.” OpenAI has actually begun rolling out conversion capabilities since our test ended, showing how fast the platform is moving. Geographic targeting is flexible: we targeted the entire United States, but you can target or exclude specific countries, cities, and ZIP codes. Budgets work via daily or lifetime limits with custom schedules.
- Ad Group & Contextual Targeting OpenAI handles targeting differently than traditional platforms. Instead of building keyword lists, you use contextual targeting. In each ad group, you describe the conversations, topics, or keywords where your product fits. These descriptions guide OpenAI’s matching system rather than acting as strict rules.
The built-in Bid Strength indicator is a helpful feature: after you enter a max CPC bid, it estimates your competitiveness and warns you if a low bid might limit delivery. Since our campaign wrapped up, OpenAI has already added first-party audience uploads via email or phone, audience bid adjustments, and product feed uploads. Several of these features launched less than a month after our test.
- Ad Creative & Dimensions Setting up ads is straightforward. Each ad needs an ad name, landing page URL, image or logo, a headline (max 50 characters), and a description (max 100 characters). We recommend keeping headlines under 25 characters and descriptions under 50 to prevent truncation.
We launched three separate ads to see what messaging resonated. Make sure to add custom UTM parameters to your URLs so you can track traffic accurately in Google Analytics 4 or Adobe Analytics.
The Data: Analyzing Our $500 Test
Because delivery was faster than expected, the $500 budget ran out on June 23 instead of June 26. Here are the final numbers from OpenAI Ads Manager and GA4:
- Impressions: 16,478
- Clicks: 93
- GA4 Sessions: 84
- Click-Through Rate (CTR): 0.56%
- Average Cost-Per-Click (CPC): $5.38
The minor gap between clicks and GA4 sessions is normal, usually caused by users bouncing before the tracking script loads or clicking twice quickly. A 0.56% CTR looks low compared to high-intent Google Search campaigns, but for a small, top-of-funnel awareness test in a brand-new environment, it aligns with early programmatic or display results. Budgets, durations, industries, and conversion optimization will naturally yield different data as the platform matures.
Three Strategic Insights
- Traffic is relevant, but expensive. The campaign generated nearly 16,500 impressions and brought in contextually relevant traffic. However, a $5.38 average CPC means this is currently a premium channel. Brands should focus on conversion probability over raw click volume. High acquisition costs make landing page performance critical.
- Don’t trust the automation blindly. We tested three creative variations. OpenAI’s algorithm allocated over 65% of the budget to Version 3, but our own data showed Version 2 performed better. Version 2 had the highest CTR (0.66%) and the lowest CPC ($5.03). Don’t rely solely on automated optimization while the platform is in beta; manual creative testing is still necessary.
- Optimize for the conversational mindset. Our biggest takeaway wasn’t about the ads themselves, but what happened after the click. Google Analytics showed an Average Engagement Time of just two seconds and a 13.1% Engagement Rate. This might stem from page load speed or the small sample size, but it also indicates AI users have different expectations than traditional searchers.
Users coming from an AI chat want direct answers. A standard marketing landing page with complex navigation creates friction. Future campaigns need dedicated, fast-loading pages that immediately match the context of the user’s prompt.
The Verdict: Is OpenAI Advertising Worth It?
OpenAI Ads still has early limitations. During our test, we couldn’t see the specific prompts that triggered our ads, and the platform lacks the deep audience reporting found in mature networks. Advertisers have to trust OpenAI’s contextual matching algorithm.
The speed of updates is a great sign. Right after our campaign ended, OpenAI added conversion tracking, attributed sales value, ROAS reporting, first-party audience uploads, and bid adjustments. They are clearly listening to advertiser feedback.
Our test proves OpenAI Ads can drive contextually relevant traffic. While we need more data from larger budgets and conversion goals to draw broad conclusions, the platform is a viable new channel for conversational marketing. Testing it now gives brands a clear head start as AI discovery scales.
Ready to test conversational marketing without wasting your budget?
At Sales Factory, we believe advantage in media comes from being early, curious, and willing to test before best practices even exist.
This isn’t the first time we’ve jumped in early. Whether it’s exploring conversational search with OpenAI or becoming the very first company to test Paramount’s new CTV retargeting beta, Sales Factory has a proven track record of moving fast. We test emerging channels firsthand so our clients benefit from the insights before they become mainstream.
