View-Through Attribution for ChatGPT Ads: What It Measures and When to Trust It
Someone sees your ChatGPT ad, does not click, and buys later. Should you credit the ad? Here is how view-through attribution works on ChatGPT and when to trust it.
View-through attribution for ChatGPT ads measures conversions from people who saw your ad but did not click it, then later visited your site and purchased. It is one of the trickiest numbers in any ad platform because it answers a hard question: did the ad cause the purchase, or would the person have bought anyway?
This guide covers how view-through attribution works on ChatGPT ads, what the OpenAI pixel actually tracks, when view-through conversions are meaningful, and how to read them alongside click-through data without overcounting.
What view-through attribution is
View-through attribution (VTA) credits an ad for a conversion when:
- A user saw the ad (an impression was served)
- The user did not click the ad
- The user later visited your site and completed a conversion event (usually a purchase)
- The conversion happened within a defined window (often 1 day for view-through)
The assumption is that seeing the ad influenced the purchase, even without a click. That assumption is sometimes true and sometimes generous.
Click-through attribution (CTA) is simpler: the user clicked the ad and converted within a window (often 7 days). The click is a clear signal of engagement. View-through is weaker because no action was taken on the ad itself.
For a broader look at how attribution works across channels, see attribution for AI channels and attribution windows explained.
How the OpenAI pixel tracks impressions and conversions
When you install the OpenAI Pixel or Conversions API on your site, OpenAI matches conversions back to ad impressions and clicks. The process works roughly like this:
- A logged-in ChatGPT user sees your ad. OpenAI records the impression tied to that user.
- The user does not click. They finish their chat and close the tab.
- Later (within the view-through window), the user visits your site and makes a purchase.
- Your site fires a purchase event to the OpenAI pixel or Conversions API.
- OpenAI matches the conversion to the earlier impression via user identity.
If the user clicked the ad before converting, it counts as a click-through conversion. If the user did not click but saw the ad, it counts as view-through.
The match relies on OpenAI knowing who the user is across the ad impression and the site visit. If the user is logged out, in a different browser, or blocking cookies, the match may not happen, and the conversion may appear unattributed.
For setup details, see ChatGPT ads conversion tracking.
When view-through conversions are meaningful
View-through attribution can be useful in a few situations:
Short consideration cycles
If someone sees an ad for a $15 product and buys it an hour later without clicking, the ad plausibly reminded them or introduced the product. The shorter the window between impression and purchase, the more credible the credit.
High impression frequency
If someone sees your ad multiple times and then buys, the combined impressions may have built familiarity even without a click. This is more common on platforms with high frequency, like social feeds. ChatGPT ad frequency is usually lower because people do not scroll a feed, but repeated queries in your category can add up.
Products that prompt research
Someone asks ChatGPT "best enzyme cleaner for pet stains," sees your ad, does not click, but Googles your brand name later and buys. The ad introduced the brand; the purchase happened elsewhere. View-through captures this path, though so would a post-purchase survey.
Complementary channel measurement
If you are running multiple channels and want to see whether ChatGPT ads influence buyers who convert through search or email, view-through gives a rough signal. It is not proof, but a spike in view-through conversions after adding ChatGPT ads can suggest the channel is doing awareness work.
When view-through conversions are noisy
View-through attribution overestimates impact in several common situations:
Long windows
A 7-day view-through window credits the ad for purchases that may have nothing to do with it. Someone who saw an ad on Monday and bought on Sunday for a completely different reason still counts. Shorter windows (1 day or less) are more defensible.
High baseline purchase rates
If many of your buyers would have purchased anyway (returning customers, email subscribers, brand searchers), view-through credits the ad for sales it did not cause. This is the overlap problem: the ad reached someone already in your funnel.
Low-intent impressions
If the ad appeared but the user barely noticed it (scrolled past, did not read the answer fully), crediting a later purchase is generous. ChatGPT ads are more visible than some feed placements, but attention still varies.
Multi-channel buyers
Someone who saw a ChatGPT ad, then saw a Meta ad, then clicked a Google ad and bought will often be counted by all three platforms. View-through makes this worse because it requires no click to claim credit.
How to read view-through and click-through together
Most ad platforms, including OpenAI's Ads Manager, report view-through and click-through conversions separately. Here is how to use both without double-counting:
Default to click-through for primary decisions
Click-through conversions are the cleaner signal. Use them for campaign optimization, CPA calculations, and budget allocation. If an ad has strong click-through performance, it is working.
Use view-through as a secondary indicator
View-through conversions can help explain why a campaign with low clicks still seems to correlate with revenue. If you added ChatGPT ads and blended MER improved but click-through conversions are modest, view-through may show where the impressions are contributing.
Never add them together blindly
Reporting "100 conversions" when 60 are click-through and 40 are view-through overstates the known impact. Report them separately: "60 click-through conversions and 40 view-through conversions." Let stakeholders see which is which.
Compare to blended MER
Blended MER (total revenue divided by total ad spend) does not care who claims credit. If view-through conversions say your ChatGPT ads drove $10,000 in revenue, but MER did not improve when you added the channel, the view-through credit may be inflated. MER is the reality check. Use the MER calculator to track it.
Ask buyers directly
Add "ChatGPT" or "AI assistant" as an option in your post-purchase survey. If buyers select it, you have direct evidence that is not tied to pixel matching. Our guide to post-purchase survey attribution explains how to set this up.
Setting the right attribution window
OpenAI allows you to adjust attribution windows in your Ads Manager settings. Common options include:
- 1-day view, 7-day click: Credits view-through conversions within 1 day of impression, click-through within 7 days of click.
- 1-day view, 1-day click: More conservative; only counts conversions very close to the ad event.
Shorter windows reduce overcounting but may undercount delayed purchases in high-consideration categories. Most DTC brands selling products under $100 do well with 1-day view and 7-day click.
If you are comparing ChatGPT ads to Meta, make sure both platforms use the same windows. Meta's default has changed over the years; check your settings. Apples-to-apples comparisons require apples-to-apples windows.
For more on how windows affect reported performance, see attribution windows explained.
A simple framework for trusting view-through
Ask these questions before crediting a view-through conversion:
- How long ago was the impression? Same-day is credible. Seven days ago is noisy.
- Would this buyer have purchased anyway? Returning customers, email clickers, and brand searchers often would. New buyers are more credibly influenced.
- Did blended MER improve? If total revenue grew without a matching hole, the channel is doing something. If only platform-reported ROAS improved, you may be moving credit around.
- Do post-purchase surveys confirm it? Buyers who say they heard about you on ChatGPT are direct evidence.
If the answer to most of these is favorable, view-through conversions are probably telling part of the story. If not, discount them.
View-through in a multi-channel world
Every ad platform claims view-through conversions. If someone saw a Meta ad, a Google display ad, and a ChatGPT ad, then bought, all three may report a view-through conversion for the same purchase. Your store made one sale, but your dashboards might sum to three.
The fix is the same across channels: use blended MER as the source of truth, add incrementality tests when budget allows, and never trust platform-reported ROAS in isolation. Incrementality testing for paid ads explains how to run holdout tests that cut through the noise.
ChatGPT ads are no different from Meta or Google here. The platform wants to show results, so it tracks generously. Your job is to cross-check.
How SecondWin helps
SecondWin installs conversion tracking correctly and reports both click-through and view-through conversions in your OpenAI account. It does not inflate numbers or combine them into a single total. When you review performance, you see what the pixel actually recorded.
Beyond tracking, SecondWin focuses on the ads themselves: writing chat cards from proven Meta Ad Library angles, testing context hints, and refreshing creative so your campaigns stay sharp. You can run a free analysis of your site to see the questions and messages it would target. Plans are on the pricing page. SecondWin is independent and not affiliated with OpenAI. ChatGPT is a trademark of OpenAI.
FAQ
Should I include view-through conversions in my CPA calculation?
It depends on how you use CPA. For internal optimization (deciding which ads to keep), stick to click-through CPA so you compare on the cleaner signal. For broader reporting (how much did ChatGPT ads contribute?), you can include view-through but label it separately and discount it when explaining to stakeholders.
What view-through window should I use for ChatGPT ads?
Most DTC brands do well with a 1-day view-through window. This credits conversions that happen shortly after an impression, which is more defensible than 7-day windows. Pair it with a 7-day click-through window to capture purchases from people who clicked but waited to buy.
How do I know if view-through conversions are real?
Compare to blended MER. If adding ChatGPT ads improved total revenue relative to total spend, the channel is doing work. If only platform-reported numbers improved, view-through may be claiming sales that would have happened anyway. Also check post-purchase surveys for direct buyer confirmation.
Can view-through attribution double-count with other platforms?
Yes. If someone saw ads on Meta, Google, and ChatGPT before buying, all three platforms may report a view-through conversion for the same purchase. This is why blended MER and incrementality tests matter: they show real results regardless of which platform claims credit.
Is view-through attribution more or less reliable on ChatGPT than on Meta?
The mechanism is similar. Both rely on matching logged-in users across ad impressions and site conversions. ChatGPT may have lower impression volume per user than Meta's feed, which can make individual view-through events more noticeable. But the same skepticism applies: view-through is weaker than click-through everywhere.