Good morning.
Twenty-two years of near-misses, gut-punches, and “maybe next season.” Arsenal FC’s fans have seen it all. Then, finally, they’ve seen their team win the title.
Marketers, take note: the strategy that feels like it’s going nowhere is sometimes the one that’s about to go everywhere. Patience pays off. Let’s move into marketing news now.
How businesses grow on Meta, new creator publishing tools, and brand safety comes to Threads
We know you usually want to skip the recap, but this is one you’ll want to catch.
Here’s the headline: The 2026 Meta Summit shows that the brands growing fastest right now are blending measurement discipline and creator-led chaos.
- On the measurement side: Always-on incrementality testing, media mix models, and geo-holdout experiments running continuously.
- On the creator side: Outbound platforms, Discord communities, and GMV-based incentive tiers that reward top performers with trips and bonuses.
Upfront media spend is losing ground to performance-based creator payouts. Brands only pay when creators deliver.
That shifts risk away from you and onto the creator. But move quickly as the creators who can actually sell are about to get pricey.
And to support that creator push: Meta rolled out a Content Planner and bulk Reels upload tool inside Creator Studio. This cuts the admin work out of running a creator program.
The planner puts scheduling gaps and performance insights in one place. Bulk upload lets you write descriptions and run copyright checks across multiple Reels in one go.
If you’re managing creators at any scale, this is worth diving into.
Bad vibes bouncer: On Threads, you can now use third-party block lists from DoubleVerify, IAS, Scope3, and Zefr to keep your ads away from objectionable content automatically.
Meta claims 99% brand safety, but the company controls what third-party verifiers can actually see. So, fair warning, the number may flatter the reality.
Plan early on Pinterest, X limits free posting, and creator spend hits the big leagues
Trigger warning. Christmas is about to be mentioned.
Santa’s summer bod: Pinterest says brands running shopping campaigns for six months or more see 33% higher ROAS than those who wait. The case for early Christmas…
The longer your campaigns run, the better Pinterest’s algorithm gets at matching your products to the right people. Lower costs and better creative data. Nice.
So what’s the play? Pinterest suggests putting 75% of your platform budget into always-on campaigns and using the quieter months to A/B test audiences and creatives.
Start now, spend smarter, and show up to peak season with your best performers already identified. St Nicholas will thank you for the prep.
X is going on a diet: The platform has capped non-paying users at 50 posts and 200 replies per day, down from 2,400, a move aimed squarely at reducing spam.
This means less junk in the feed means stronger engagement signals, cleaner trend data, and a platform that’s gradually getting easier to run campaigns on.
No longer sitting at the kiddie table: Creator ad spend is projected to hit $44B in 2026, and legacy media is no longer pretending otherwise.
Fox, Warner Bros. Discovery, and Amazon all built creators into their upfront pitches this week, because creators also bring along communities that trust them.
That trust is now worth serious money, and the brands allocating accordingly are treating creator partnerships as a primary channel, not a test-and-learn experiment.
Less marketing theory. More “here’s exactly what we did.”
Every week, Marketing Against the Grain hands you the playbooks, campaign breakdowns, and frameworks that are actually moving the needle right now.
Built by practitioners who are running campaigns today, not consultants rehashing advice from three years ago.
You’ll get HubSpot’s real experiment data, honest analysis of what’s working in SEO, social, and AI, and tactics specific enough to steal and implement this week.
Whether you’re a team of one or leading a full department, they cut straight to what worked and why so you can spend less time researching and more time executing.
How to do content engineering with Claude Code
Building an AI content system is easy. Building one that doesn’t hallucinate your entire editorial reputation into the void? That’s harder.
Ryan Law has documented the content automation system he built using Claude Code. These are the fundamentals:
1) Mimic human workflows by linking editorial skills: The system’s backbone is 23 skill files and each maps to a specific editorial task: keyword research, topic gap analysis, etc.
A master skill called blog-pipeline chains them in sequence taking a keyword to a near-finished draft. Pretty good.
2) Output every step of the process for iteration and troubleshooting: Every stage saves its own output file.
When something breaks, you pinpoint exactly where and restart from that step.
Why this matters: Agentic pipelines can fail without a word. Saving staged outputs is the unattractive engineering decision that keeps it chugging along.
3) Create test cases for recursive self-improvement: Anthropic’s skill-creator skill tests each pipeline stage, generating outputs both with and without custom skill file guidance, then suggesting improvements.
Skill files bloat over time. This self-testing loop keeps them trim and effective, removing guidance that doesn’t actually move the needle on output quality.
4) Give LLMs great data from great sources: Claude pulls real keyword metrics, SERP data, and competitor content gaps directly via the Ahrefs MCP, no hallucinated SEO numbers.
The bottom line: Mandating specific, trusted data sources is what keeps AI output substantive rather than confidently hollow.
5) Front-load human direction: A small amount of human guidance at the start: the angle, key subtopics, specific features to mention shapes the entire draft more efficiently than heavy revisions after the fact.
This gets captured in a context parameter before writing begins, so the AI has a reference point from the first word.
Okay, the surface has been scratched. But Ryan has a few more steps to go through, so make sure to get to the end in his original article here. It’s worth a read. Really.
So. Much. AI news. How do you know what’s worth your attention?
There’s always something new with AI, but not everything deserves your attention.
Good thing The Deep View only focuses on AI developments that could actually impact your work and decisions.
It shares must-know updates into a fun, 5-minute read and has helped 500,000+ professionals stay informed without drowning in the constant flood of updates.
Even readers from Google, Meta, Microsoft, and a16z start their day with these insights.
The first intentional search touchpoint is…
When users want to look up for something, they have to decide where to knock first.
It turns out, Google’s front door is still heavy, but AI is installing a very popular side entrance.
The battle for the “starting line” is tighter than you might think:
- 47% of users still reflexively head to Google or other traditional search engines to begin their journey.
- A massive 37% now start directly with AI tools like ChatGPT, Gemini, or Copilot. That is a significant chunk of high-intent traffic moving away from the ad-heavy SERP.
- Voice assistants (11%) and social media (4%) are still specialized niches, but they aren’t currently the primary “start” buttons for most.
AI is no longer a secondary check. For nearly 4 in 10 users, the search journey starts and potentially ends within a chatbot’s interface, bypassing traditional websites entirely.
Actionable tip: Claim your answer engine territory.
Use schema markup and provide clear, direct answers to common industry questions so AI tools can pull your brand’s “truth” as the first response in these 37% of new-age searches.
AI is restructuring search faster than most marketers can keep up.
Pro subscribers stay ahead with weekly intelligence on what’s changing and exactly what to do about it. $349/year covers your whole team (up to 10 people).
INFLUENCER MARKETING: Why are Martha Stewart and Unilever sending free products to micro-influencers? Because it works: #1 page Amazon ranking, up to 13X monthly revenue, real external traffic. Stack Influence automates the whole play — no upfront creator fees. Save 10% if you sign up this week.*
SOCIAL MEDIA: X’s rebuilt AI targeting platform and audience data shows that users who earn over $90K household income are central for your ad budget. Sports, movies, and politics lead 2025 conversation topics, so campaign alignment is less guesswork than it used to be.
META: “Soft outages” on Meta’s ad platform are causing sudden ROI drops, wrong audience delivery, and bot traffic spikes. However, Ads Manager is showing nothing unusual. Your campaigns are fine, but the platform isn’t. Which is another case for channel diversification.
YOUTUBE: Users watch over 2B hours of vertical video on TVs each month, and Shorts are at the center of that shift. They are even surfacing at search results with comments being added alongside videos, so content for your phone screen now has a legitimate shot at the living room.
WHATSAPP: iOS users are getting a tidier media sharing experience, with a swipeable recent photos strip replacing the full gallery pop-up, chats staying visible while you share, and a long-press shortcut on the + button for quick media access. A small, but useful update.
*This is a sponsored post.
A cowboy rode into town on Friday, stayed three days, and rode out again on Friday. How did he do that?
You can find the answer here.
Interested in advertising in one of our newsletters?
Connect with over 100,000 of the world’s best marketers who read a newsletter by Stacked Marketer.
