ChatGPT reprices budgets, Google fixes signals: July 28

ChatGPT advertisers can now spend twice their stated daily budget on a busy day, while both OpenAI and Google are giving their ad systems better conversion signals. That combination can improve customer acquisition cost, meaning the amount paid to win a customer. It can also burn cash faster when the tracking underneath it is wrong.

ChatGPT makes the daily budget less daily

OpenAI changed the meaning of a ChatGPT Ads daily budget on July 27. Its new budget rules treat the number as a seven-day average. A campaign can spend up to twice that amount on one day, although billed media spend cannot exceed seven times the daily setting across seven days.

So a $100 daily budget can produce a $200 day, but no more than $700 for the week. That is a cash-flow change, even when the weekly ceiling stays intact. Owners who reconcile spend each morning may see a false emergency. Short campaigns face the sharper risk because a hot delivery day can consume more of the available cash before the sales team knows whether those clicks were useful.

Use a campaign-total budget when the absolute ceiling matters more than smooth delivery. For continuing campaigns, judge spend over the full seven-day window and keep the bank balance ready for the maximum daily charge.

ChatGPT starts bidding toward sales, not just clicks

ChatGPT Ads also added conversion-optimized cost-per-click campaigns. In plain English, the system still charges for clicks but tries to find people more likely to complete a selected action, such as a purchase or lead form.

That moves the buying logic closer to revenue. The catch is decisive: a campaign can optimize toward only one standard conversion event, and neither that event nor the objective can be changed after creation. Existing click or impression campaigns cannot be converted. A weak event, such as a shallow page visit, teaches the system to buy cheap activity rather than customers.

The first job is therefore measurement, not scale. OpenAI says its pixel and server connection can send the same event with one identifier so duplicates are removed. Compare the platform's conversion count with your customer-management system before raising spend. Our landing-page structure guide explains where that handoff should happen.

Google tells advertisers to stop resetting the machine

Google updated its conversion measurement guidance with a 7-to-14-day learning period for automated bidding. It also says linked Google Analytics events can take 24 to 48 hours to appear and stresses the difference between primary conversions, which steer bids, and secondary conversions, which are observed only.

This is a budget discipline problem. Changing targets or budgets repeatedly during learning restarts the period, while marking every small action as primary tells Google that a newsletter signup and a paid order deserve equal money. Both mistakes can make the dashboard busy while revenue stalls.

Freeze material edits during the learning window unless spend is unsafe. Then audit the primary conversion list against actual revenue events. The wider control checklist sits in our Google Ads budget controls breakdown, and the tracking failure mode is covered in the July 14 brief.

Google Ads Editor puts retention into bulk work

Google Ads Editor 2.13 now supports the customer retention goal for Performance Max, Google’s campaign type that places ads across several channels from one setup. The release also brings channel performance reports, bulk controls for generated text, and warnings for video bids and budgets.

The retention setting matters most for businesses that already have a meaningful customer list. Reaching past buyers can shorten payback because trust already exists. But mixing re-engagement with new-customer acquisition muddies the number that matters: what it costs to add a genuinely new buyer. Keep those jobs separate in reporting, even if Editor makes them easier to manage together.

Google draws a harder line around paid reviews

Google added a July 24 rule barring fake or undisclosed incentivized reviews from review markup, the code that can make star ratings appear beside a search result. Its official guideline says reviews exchanged for money, discounts, vouchers, or free products must disclose that incentive clearly.

Stars can lift confidence before a prospect reaches the page. Losing them can therefore lower conversion without changing traffic. Businesses using review rewards should make the disclosure visible to people, not merely buried in page code. Genuine-review collection remains a lead-conversion asset; hidden compensation is now an avoidable search risk.

What to watch

• Watch whether ChatGPT's conversion bidding produces qualified sales after the first full pacing window, not merely a cheaper click.
• Watch Google Ads accounts for Editor 2.13 availability before building retention reporting around it.
• Watch review-rich results after disclosure changes; a missing star treatment can make the same search traffic convert worse.