This Week in Marketing: Your Dashboard Lied, the Platforms Blinked, and Reddit Became Your SEO Team

If you only had time for one marketing headline this week, make it this: AI didn’t just change how customers find us. It changed whether we can even see them arriving.

Three stories from the week of October 12 point to the same lesson. Our measurement is more fragile than our dashboards admit. The ad platforms are handing back some controls while quietly taking others. And the content that wins in AI search increasingly lives on platforms we don’t own.

Here’s my read on each and what I’d actually do about it this week.

Read more: This Week in Marketing: Your Dashboard Lied, the Platforms Blinked, and Reddit Became Your SEO Team

1. Your ChatGPT traffic didn’t disappear. Your tracking did.

Starting October 9, right after OpenAI’s GPT-6 rollout, marketers watched ChatGPT-referred traffic fall off a cliff. Depending on whose data you looked at, the drop ran anywhere from 22% to 57% (Search Engine Journal).

The twist: total traffic barely moved. Adobe’s analysis of roughly 500 pages showed ChatGPT visits down about 31,000 a day while Direct traffic rose about 26,000 a day. The visits were still showing up. They had just lost their referrer data and UTM tags, so GA4 filed them under Direct. OpenAI reportedly confirmed the issue on October 14 and began rolling out a fix the next day (Ignite Visibility roundup).

My take: this is the scariest kind of bug, because it looks like a strategy problem. Picture the meeting where someone sees AI referrals down 40% and proposes pulling back on the content investment that’s actually working. A week of bad attribution can kill a good program.

As AI assistants become a real acquisition channel, we’re stacking our decisions on a measurement layer we don’t control. That means two things:

• Never react to a single channel’s swing without checking where the traffic went. Look at totals, Direct, and conversions together.
• Annotate October 8 to 15 in every report now, before it shows up in a Q4 review with no context.

2. Google gave back a control. Meta took one away.

Google Ads is rolling out Channel Prioritization and Budget Allocation controls globally across Search, Shopping, Performance Max and Demand Gen (Search Engine Land). You can rank channels like Search, Search Partners and the Display Network, and on tCPA or tROAS the system funds high-priority channels first. You can also cap how much of the daily budget low-priority channels can touch.

That’s a long-requested win. But read the fine print: every channel defaults to High Priority at rollout, so doing nothing can shift your spend mix. Google’s own suggestion to preserve current behavior is setting Display to low priority with a 50% cap, and the old spend allocation setup starts going away in February 2027.

The same week, Meta went the other direction. More campaign types have lost ad-set-level placement exclusions, so you can no longer block placements like Explore or in-stream video at that level (Social Media Today). That follows Microsoft removing Max CPC from new automated campaigns as of October 1.

My take: the pattern isn’t “automation good” or “automation bad.” It’s that every platform is redrawing the line between what you decide and what its AI decides, and the defaults are rarely written in your favor. The teams that win aren’t the ones fighting automation. They’re the ones who know exactly which levers still exist and check them every time a platform ships an update.

3. AI search is citing Reddit and YouTube, not just your website

An SE Ranking study of more than 100,000 keywords across 20 niches found social or community platforms in about 68% of Google AI Overviews and 32% of AI Mode answers, versus about 8% of ChatGPT answers (Social Media Today). YouTube led in AI Overviews; Reddit led in AI Mode and ChatGPT.

Two other updates this week push the same direction. YouTube expanded Ask YouTube, its Gemini-powered conversational search, to all US users on mobile, web and TV (Social Media Today). And Bing Webmaster Tools added a Topics Portfolio that groups the queries where you’re cited in AI answers and lists high-volume queries where competitors are cited and you aren’t (Search Engine Land).

My take: for years we treated SEO and social as separate budgets with separate owners. AI search is merging them. If the answer engine trusts a Reddit thread or a YouTube explainer more than your service page, your “SEO strategy” now includes your video library, your community presence and what customers say about you in public.

That’s good news for brands with real expertise and real customers. It’s bad news for brands that have been publishing thin content and hoping to rank. Bing’s new report is effectively a free content-gap list, so use it.

What I’d do this week

1. Audit AI referral traffic. Compare ChatGPT and Direct in GA4 for October 8 to 15, confirm tagged clicks have recovered, and annotate the gap.
2. Check your Google Ads channel priorities. Don’t let the High Priority default reshuffle your spend. Decide on Display deliberately before February 2027 decides for you.
3. Review Meta account-level placement controls. With ad-set exclusions gone, account-level restrictions and third-party verification are what’s left for brand safety.
4. Run a social-in-AI-search check. Search your top 10 money queries in AI Overviews, AI Mode and ChatGPT. Note which Reddit threads, YouTube videos and community posts show up, and whether you’re in them.
5. Tighten your YouTube basics. Clear titles, chapters and videos that directly answer customer questions now feed Ask YouTube, not just YouTube search.
6. Pull Bing’s Topics Portfolio. Turn its “competitors cited, you’re not” list into next quarter’s content plan.

The bigger picture

Every story this week comes back to one question for marketing leaders: how much of your growth engine do you actually understand?

AI is now sitting between us and our customers at every layer: in how they search, in how platforms spend our budgets, and in how our traffic gets counted. That’s not a reason to slow down on AI. It’s a reason to get sharper about measurement, defaults and where your brand really shows up.

The CMOs who thrive in this next stretch won’t be the ones with the most AI tools. They’ll be the ones who can still explain, in plain English, why the numbers moved.

What did you see in your own data this week? I’d love to compare notes. Connect with me on LinkedIn or reach out through richbernich.com.

Sources

• ChatGPT traffic drop appears to be a tracking problem, Search Engine Journal
• Google Ads channel prioritization and budget allocation controls, Search Engine Land
• Meta updated ad placement control guidance, Social Media Today
• AI favors brands with an active social media presence, Social Media Today
• YouTube rolls out AI search chatbot to all US users, Social Media Today
• Bing Webmaster Tools Topics Portfolio, Search Engine Land
• Digital Marketing News 10/10 to 10/16/2026, Ignite Visibility