Pinterest scores spend, Threads opens ads: Sept. 22

Ad platforms want fewer hands on the controls, while the stores receiving their traffic want more. Pinterest now grades campaign setups partly on how much control advertisers surrender. Threads has lowered the cost of entry. Amazon, meanwhile, shut a Meta shopping agent out. Today’s money question is who gets to steer the sale.

Pinterest turns automation into a score

Pinterest gave advertisers a campaign optimization score on September 21. The dated product report says the score grades how closely a campaign follows the platform’s preferred setup, including automated targeting and bidding, varied ads, broad audiences, and steady budgets.

That score is guidance, not profit. Pinterest says higher scores correlate with better campaign results, but correlation only means the two tend to move together. It does not prove that chasing the score caused cheaper customers. The system also rewards adoption of Performance+, Pinterest’s automated campaign suite. Pinterest’s own Performance+ documentation says it can run consideration, conversion, and catalog-sales campaigns with fewer advertiser inputs.

The margin risk is simple: an easy score can become the target while customer acquisition cost, or what you spend to win one new customer, gets ignored. Retailers with clean sales tracking can test the recommendations. Everyone else should fix the measurement first. Our paid-ad budget controls guide explains why a platform’s preferred setup and an owner’s profit target are different jobs.

Threads removes an ad-access tollbooth

Meta will let businesses run Threads ads without an Instagram profile starting in September 2026. Its advertiser help page also removes the requirement for matching Threads and Instagram usernames. The rollout is gradual, so some accounts will not see the option yet.

This saves setup time and lets a business treat Threads as its own paid channel. It does not make the audience interchangeable with Instagram. A text-led conversation feed invites different behavior from a visual feed, so copying the same creative can buy reach without buying intent.

Run Threads as a separate placement test. A placement is simply the part of an app where an ad appears. Keep its spend, message, and completed sales visible on their own before rolling it into a wider Meta campaign. That is the same control principle behind our Meta Advantage+ budget analysis: automation earns more budget only after the business result survives inspection.

Amazon closes Meta’s route to the cart

Amazon blocked Meta’s Muse shopping agent on September 21 after users saw a notice saying the agent was unauthorized. Amazon told GeekWire, which first reported the block, that Meta had not sought permission and that Muse appeared to access customer credentials and account data without identifying itself. Meta had not commented in the report.

For sellers, the immediate effect is limited: a shopper using Muse cannot complete an Amazon purchase through that route. The larger acquisition lesson is more expensive. A recommendation made inside one company’s assistant does not guarantee access to another company’s checkout. Whoever controls the cart can close the door.

Businesses testing sales from assistants should measure completed orders, not product mentions or shopping clicks. Keep product feeds accurate, but preserve direct demand too. Our guide to branded search economics shows why customers who deliberately look for your name are less exposed to a fight between platforms.

What to watch

• Watch whether Pinterest shows how each recommendation changes sales, rather than only the score. Better explanations would make budget decisions safer.
• Check when standalone Threads advertising reaches your account, then look for placement-level sales reporting before assigning meaningful spend.
• Follow the Amazon and Meta talks. A formal integration could reopen the route, while a lasting block keeps assistant-driven purchases fragmented.