You’re probably thinking about search wrong

Today’s Social Media advertising rates are pricemaxxing and CPM mogging:

↗️ META: $7.45 CPM | ↘️ TIKTOK: $4.15 CPM | ↗️ SNAPCHAT: $14.05 CPM

In this week’s edition: 

• 🛒 TikTok’s biggest shoppers? Not Gen-Z!

•  Meta is rolling out ANOTHER app?! 🤳

• 📺 Report: CTV targeting misses the mark 80% of the time

But first…

🔎 What marketers are getting wrong about search

It’s not just AI: Google Searching behavior works differently now than it used to. And many marketers have yet to catch up. Here’s what you’re missing:

💡 Search now behaves more like a conversation than a lookup. For advertisers, this means rethinking how campaigns are built, optimized, and measured. As SearchEngineLand put it recently: “Keywords aren’t dead, but they’re no longer the blueprint.”

Here’s what’s actually happening when someone searches now: 

• Google’s AI uses a technique called “query fan out,” splitting a complex question into subtopics and running multiple concurrent searches to build a comprehensive response. During this process, the AI infers commercial intent from purely informational queries. For example, someone asks, “Why is my pool green?” They’re not shopping. But Google’s reasoning layer detects a problem that products can solve and serves ads for pool-cleaning supplies alongside the explanation.

• If you want to show up inside these AI responses, you need broad match keywords, Performance Max, or the newer AI Max for Search campaigns.

• Google rewards landing pages that explain why and how someone should use your product, not just its features.

👉 AI placements often capture higher-funnel traffic and should not be compared apples-to-apples with branded search terms.

The Inside Scoop🍦: Now trending in performance media

🔎 Google’s Performance Max testing is getting easier: the platform is rolling out a beta feature that lets advertisers A/B test creative assets within one Performance Max asset group. Early testing suggests that flights should be longer than three weeks to yield stable results.

📺 YouTubeTV is looking to increase paid subscriptions by offering 10 new lower-cost TV packages. These options, such as sports-centered or movie-focused tiers, allow cost-conscious consumers to pay only for the media they plan to consume. So far, only a handful of the 10 packages have been announced; the full details are coming soon.

♾️ Meta has been experimenting with a new image sharing app called Instants. It’s essentially a rebranded version of an Instagram feature called Shots, and it functions like a mix of BeReal and Snapchat. Meta is likely trying to attract and retain younger users with the app.

⏰ TikTok published a new 40-page Creative-Made-Simple Guide to help brands improve content and ad performance on the platform. Among the takeaways:

• The playbook emphasizes that creative is "single-handedly the most important mover when it comes to increasing performance and driving conversions on TikTok." 

• Reminder that agencies can take advantage of TikTok’s free suite of tools to increase content production, including Symphony Assistant, its AI creative companion for brainstorming, script generation, and trend analysis; Symphony Creative Studio, and AI video generator that creates content from simple inputs; Script Generator, which handles automated TikTok-native script writing; and the Creative Center, offering real-time trend data and top ads analysis

• TikTok claims that Smart Creative (when using 5+ creative variations) can lift performance by up to 9% by automatically identifying top performers and pausing fatigued content. 

🛒TikTok’s strongest shoppers aren’t Gen Z, they’re millennials. 54% of users ages 25-44 have researched products after seeing them on the app, with another 47% purchasing a product they first heard about on TikTok. TikTok’s influence expands beyond purchases, with 37% of millennials having recommended a product to others or followed a brand on social media. More than ever, TikTok is reinforcing its position as a powerful tool for product discovery and consideration, particularly among millennials.

🐦 X has rolled out image polls globally after testing in Japan, allowing users to attach photos to poll options. Early examples are showing strong engagement, making this a promising format for visual comparisons.

🐭 Disney Advertising reached out to us recently with a slate of high visibility ad opportunities—these aren’t all new, but a reminder that they’ve got the sauce: 

• Dynamic A position: guaranteed first spot in a commercial break

• Enhanced CTV ads: interactive/immersive streaming ad experiences 

• Blitz: achieve high targeted frequency over 2-4 days

• Hot on Disney/Hulu: adjacency with Top 100 most popular streaming content across Hulu and Disney+ libraries 

• Sequential Ad experience: brand story telling using sequenced ads

• Disney social activation: tapping into Disney's social account & their collective 3 billion + followers

⛰️ Paramount Ads Manager has announced a series of new updates: 

• You can now target certain audience behaviors (e.g. Parents, Online Shoppers, Sports Enthusiasts). Previously you could only target genres. 

• There’s a new Performance Multiplier, which amplifies your reach among high-intent audiences on trusted sites, like Business Insider, Fox Sports, USA Today, Yahoo!, NBC News, MSN, and more. Results are reported in the Paramount Ads Manager dashboard.

• You can now target Media Channels, including Cable, which is in Beta.

  

👽 Reddit is positioning itself as a place where people start and end their searches without ever using Google or other search options. On Reddit’s recent Q4 2025 earnings call, the platform claimed 80 million people now use Reddit search every week.

According to SearchEngineLand, the “increase followed a major change: Reddit merged its core search with its AI-powered Reddit Answers.”

📊  By the Numbers: How accurate is CTV’s targeting?

What’s the likelihood that your CTV campaign is showing ads to the wrong households? According to one company: a lot more than you’d think. 

Data intelligence company Truthset analyzed identity matching from 22 providers — and claim to have found IP-to-email matches fail 84% of the time, while IP-to-address matches fail 87% of the time. 

Even across 25 demographic categories, the company found an average error rate of 55%. 

And the problem's getting worse as ISPs bundle households into collectives: one household that should've had 2 IPs over nine weeks had 11, meaning your "precision targeting" is hitting random neighbors instead.

Thanks for reading! We’ll be back next week to fill your inbox with more tips, tricks, and treats from the addressable universe.

In the meantime, stay up to date with the latest performance marketing intelligence by following us on Linkedin, Instagram, and X.

This week’s edition of the Thread was created by Jess Dashner, Caitlin LaValle, Anoushka Madan, Mariya Samardzhieva, Kiara Quigley, Lara Spijkerman, and Carly Carioli.