ChatGPT exclusions, Google hides ads: July 16 brief

Two new ad controls make wasted reach easier to stop, while fresh loyalty data shows why a large member list can still produce thin repeat revenue. ChatGPT advertisers can exclude places and audiences. Google is testing a way for shoppers to hide sponsored products. Both changes put more pressure on paid clicks to be genuinely useful.

ChatGPT gives advertisers two ways to stop wasted reach

ChatGPT Ads Manager added location and audience exclusions on July 15, according to interface screenshots verified by Search Engine Roundtable. An exclusion tells the system where an ad must not appear. Advertisers can now block selected locations and prevent selected audience lists from seeing a campaign.

That sounds like routine account housekeeping. It is really a customer acquisition cost control. Customer acquisition cost, or CAC, is what you spend to win one customer. A business that cannot serve a state wastes every impression and click delivered there. Showing an acquisition ad to existing customers can also make the campaign look productive while buying attention from people already in the database.

The sharper move is to separate two jobs. Acquisition campaigns should reach plausible new buyers. Retention campaigns should speak to existing customers with a different offer and measurement rule. The same discipline applies across the broader paid ads budget and controls lane, even while ChatGPT's ad product is young.

Google lets shoppers collapse sponsored products

Google confirmed on July 15 that it is testing a hide-and-show control for sponsored products in Shopping results, according to reporting that includes confirmation from Google's Ads Liaison. Sponsored products are paid listings shown beside unpaid shopping results. The test extends a similar control already used in standard Search results.

For retailers, the immediate risk is not that every shopper will hide every ad. This is a test, and Google has not published adoption or performance figures. The real pressure lands on weak listings. A shopper who can collapse the paid block has an easy exit when the first screen shows vague titles, poor images, or prices that do not match the landing page.

That makes feed quality part of conversion economics. A product feed is the file that sends Google each item's title, image, price, and availability. Before raising bids, check whether the paid listing earns the click on its own merits. Then use the same waste checks covered in our Google Ads budget controls breakdown: search terms, exclusions, and the difference between reported sales and truly new sales. More bid cannot repair a weak shelf.

Loyalty enrollment keeps overstating retention

Upside reported that 86% of surveyed consumers consider loyalty programs important, yet regular use falls to 59% for grocery programs, 49% for fuel and convenience stores, and 38% for restaurants. The company based the survey on 3,400 consumers, while its broader retention analysis used transactions from 7.7 million customers, according to Upside's research page.

The money leak is measurement. Enrollment counts a customer once. Retention asks whether that person comes back and spends again. When a dashboard celebrates members without tracking repeat visits, reward cost rises while revenue may stand still.

Owners should segment members by last purchase date and purchase frequency, plain English for how recently and how often someone buys. The useful target is the persuadable middle: customers who have bought before but are drifting, not regulars who would return without another discount. Pair that with an email retention system built around customer behavior, and judge each reward on extra visits or margin, not redemptions alone. A redeemed discount is a cost until it changes behavior.

What to watch

• ChatGPT's next reporting update needs to show whether exclusions reduce wasted spend without shrinking qualified pipeline.
• Google's Shopping test matters only if shoppers use the control often enough to change click volume or conversion quality.
• Loyalty teams should watch repeat purchases by cohort, meaning groups that joined in the same period, rather than celebrating the total member count.