ChatGPT Ads now selects an automated bid strategy for you by default, and it does not promise a cost target. In the same wave, OpenAI shipped view-through conversion reporting on a window fixed at one day. For B2B advertisers, those two changes pull against each other: the platform is optimizing for volume, on a measurement window too short to tell you whether that volume was worth buying.
- Default-on. Maximize results is selected by default on eligible new ad groups
- No cost line. It explicitly does not guarantee a CPA, CPC, or ROAS target
- One day. The view-through window is fixed and cannot be configured
- Separate column. View-through appears as VTA (1d) and is excluded from Conversions
- The B2B problem. Thin conversion volume plus unequal lead values is the worst case for volume optimization
What changed, and what OpenAI does not promise
Maximize results is a bid strategy that hands bid decisions to ChatGPT Ads. OpenAI describes it as a strategy that lets the platform automatically set and adjust bids on your behalf, optimizing toward the campaign goal within your budget.
Two sentences in that documentation matter more than the rest. The first: for eligible new ad groups, Maximize results is selected by default. You do not opt in. Any ad group you have built since it shipped may already be running it.
The second is the one to read twice. OpenAI writes that Maximize results prioritizes result volume and does not guarantee delivery against a specific CPA, CPC, ROAS, or other cost-efficiency target at this time. That is not marketing hedging. It is the platform telling you plainly that this strategy has no cost ceiling.
A default that optimizes for volume and declines to promise a cost target is a reasonable product decision. It is a poor fit for an account that measures success in pipeline rather than clicks.
Why a volume-maximizing default is riskier for B2B
Automated bidding needs dense, comparable conversions to work well. Most B2B accounts have neither. The two conditions that make volume optimization safe are the two conditions B2B accounts most reliably violate.
The first condition is data density. Bid automation learns from conversion events. An ecommerce account producing hundreds of purchases a week gives the system a rich signal within days. A B2B account running ChatGPT Ads may produce a handful of demo requests in a month. There is very little for a volume optimizer to learn from, and a great deal of budget it can spend while trying.
The second is value parity. Volume optimization treats every result as interchangeable. In B2B they are not. A free-tier signup and an enterprise trial request both register as one conversion, and they can differ in eventual value by an order of magnitude. The bidder cannot see that difference, so it will happily buy more of whichever is cheaper to acquire.
This is why cost-per-action targets exist as a concept. A cost target is how you tell a bidder that not all volume is worth the same money. Maximize results is documented as not honoring one.
The measurement window that cannot verify it
ChatGPT Ads added view-through conversion reporting on a window fixed at one day. OpenAI defines a view-through conversion as one counted when someone converts within one day after seeing an eligible ad impression, where no qualifying click receives credit for the same conversion.
The window is not adjustable. OpenAI states that the view-through window is fixed at one day and is independent of your configured click-through window. You can change the click window. You cannot change this one.
Set that against how B2B actually buys. Our own early-results breakdown puts B2B cycles in the 30 to 180 day range, which matches most operators' experience of a committee purchase with a security review in it. A buyer who sees your ad on a Tuesday and books a demo three weeks later is invisible to a one-day view window.
So the two changes arrive pointing in opposite directions. The bidder is now free to chase volume without a cost ceiling, and the new measurement column that might have shown you the influenced conversions it produced only looks one day back.
You have been given more automation and a shorter lens to inspect it through. For a channel with a long consideration window, that is the wrong pairing.
One further detail worth getting right in your reporting. The Conversions column counts click-through conversions only, and view-through appears separately as an optional column named VTA (1d). OpenAI is explicit that view-through conversions are supplemental reporting and do not affect CPA, post-click conversion rate, bidding, billing, or conversion optimization. They are not inputs to anything. Do not sum the two columns and report the total as performance. If the pixel is not firing cleanly to begin with, none of these columns mean much; the pixel install walkthrough covers that setup.
What “eligible new ad groups” does not tell you
OpenAI says the default applies to eligible new ad groups and never defines eligibility. The documentation names no campaign type, no objective, no account status, and no date from which the behavior applies.
That matters operationally. You cannot reason from the docs to a list of which of your ad groups flipped to the default and which did not. There is no published rule to apply, which means the only reliable method is inspection: open each ad group in the Ads Manager and read the bid strategy field.
If you run a handful of campaigns this is ten minutes of work. If you manage ChatGPT Ads across a client roster it is a standing audit item, and worth adding to whatever checklist governs your campaign reporting routine.
The three bid strategies, side by side
Only one documented strategy holds a cost line you set. ChatGPT Ads currently offers CPM, CPC, and conversion-optimized objectives, and the bid strategy sits underneath whichever objective you pick.
The oCPC row carries a constraint that catches people out. A conversion-optimized campaign uses one selected conversion event, and that event cannot be changed after the campaign is created. Existing CPM or CPC campaigns also cannot be converted into oCPC campaigns. Choose the event deliberately, because the decision is durable.
For most B2B accounts starting out, Manual: Max bid is the honest first setting. It produces cost data you own rather than cost data the system chose for you, and it gives you a baseline to test automation against later. The campaign setup walkthrough covers where this sits in the build flow.
What this costs at B2B unit economics
The risk is not that automated bidding overspends. It is that it optimizes toward the wrong result and you find out a quarter later.
Work it through with your own numbers rather than a benchmark. Take your monthly conversion count on the channel, your average contract value, and the share of conversions that are genuinely qualified. If your qualified share varies a lot between conversion types, a volume optimizer will drift toward the cheap ones, because cheap and plentiful is exactly what it was asked to find.
Two accounts can run identical spend at an identical cost per conversion and produce completely different pipeline, purely because one bought more of the expensive-but-qualified type. Volume optimization has no way to prefer the second account's mix. Neither does a one-day view window help you notice the drift, since most of the qualified conversions land outside it.
This is the same reasoning behind measuring incrementality rather than raw conversion counts, and it is why our ROI measurement framework starts from pipeline rather than platform-reported conversions. For the cost side of the picture, the ChatGPT Ads cost reference covers what budgets buy at current rates.
What to change in your account this week
Five things, in priority order. None of them takes long, and the first one is the one that matters.
- Audit every ad group. Open each one and read the bid strategy field. Do not assume a campaign-level setting covers it; the default is applied per ad group.
- Switch to Manual: Max bid anywhere you have not deliberately chosen automation, then set the ceiling from your own cost data.
- Split your reporting columns. Keep Conversions and VTA (1d) visibly separate, and make sure nobody downstream is summing them into a single performance number.
- Check your conversion events. If a conversion-optimized campaign is pointed at a low-value event, you cannot change it later. That decision is durable.
- Write down your baseline before you test automation. Cost per qualified conversion for the last full month, from your own analytics and consistent UTM parameters, not the platform's numbers.
OpenAI has not published what makes an ad group eligible for the default. Until it does, the only way to know which of yours flipped is to open them and look.
If your ChatGPT Ads conversions are not reaching your CRM yet, the measurement side of this is the part to fix first. The AI-Advisors ChatGPT Ads integration installs the conversion pixel and overlays paid performance against your organic AI citations, so you can see which conversions the channel actually influenced. Free on all plans.
What we do not know yet
Three open questions, stated as open. The documentation does not answer them, and neither will we by guessing.
- Eligibility. OpenAI has not published what makes an ad group eligible for the default.
- Performance. There is no public data comparing Maximize results against manual bidding on B2B accounts. Anyone claiming a lift figure today is extrapolating.
- Where the window goes. OpenAI has signaled a conversion model that optimizes on views, which would make view-through an input rather than a footnote. It has not shipped, and no configurable window has been announced.
We will update this post when any of the three resolves. The broader platform picture, including the 52 markets now live and the objectives available in each, sits in the complete ChatGPT Ads guide.
Frequently Asked Questions
#Is Maximize results turned on by default in ChatGPT Ads?
Yes, for new ad groups. OpenAI's Help Center states that for eligible new ad groups, Maximize results is selected by default. Existing ad groups keep whatever strategy they were built with. OpenAI does not publish what makes an ad group eligible, so the only reliable way to know which of your ad groups are on it is to open each one and check the bid strategy field.
#Does Maximize results hit a target CPA?
No. OpenAI states that Maximize results prioritizes result volume and does not guarantee delivery against a specific CPA, CPC, ROAS, or other cost-efficiency target at this time. It spends the budget to produce as many results as it can. If you need a cost ceiling, Manual: Max bid is the documented option that holds one.
#Should B2B advertisers use automated bidding on ChatGPT Ads?
Not as a default, and not yet. Automated bidding works best with dense conversion data and roughly equal-value conversions. Most B2B accounts have neither: a handful of conversions a month, and lead values that differ by an order of magnitude. Start on Manual: Max bid, gather cost data you control, then test automation deliberately against it.
#What is a view-through conversion in ChatGPT Ads?
It is counted when someone converts within one day after seeing an eligible ad impression, and no qualifying ad click receives credit for the same conversion. It appears as an optional column called VTA (1d). OpenAI states that view-through conversions are supplemental reporting and do not affect CPA, post-click conversion rate, bidding, billing, or conversion optimization.
#Can I change the ChatGPT Ads view-through window?
No. OpenAI states the view-through window is fixed at one day and is independent of your configured click-through window. You can adjust the click window, but the one-day view window is not configurable. For businesses with buying cycles measured in weeks or months, that window will miss most influenced conversions.
#Do view-through conversions get added to my Conversions total?
No, and this is the detail most likely to cause a reporting error. The Conversions column counts click-through conversions only. View-through conversions are reported separately. Adding the two columns together double counts nothing but overstates performance, because the two metrics answer different questions and are not designed to be summed.
#How do I switch a ChatGPT Ads ad group off automated bidding?
Open the ad group, find the bid strategy section, and change it to Manual: Max bid. Do this per ad group rather than assuming a campaign-level setting covers it, since the default is applied at the ad group level. Audit every ad group created since the strategy shipped, because any of them may have launched on the default without an explicit choice.
Related Reading
- ChatGPT Ads Conversion Tracking Is Real: What OpenAI's April 30 Privacy Policy Update Tells You About the Roadmap
- ChatGPT Search Ads for B2B: Early Results, What's Working, and What Isn't
- How to Set Up a ChatGPT Ad Campaign in 8 Steps (September 2026)
- How Much Do ChatGPT Ads Cost? CPM Range, Bid Floors (September 2026 Update)
