Every once in a while, Google likes to remind the SEO industry who's actually in charge. Over the past six weeks, that reminder came in the form of three updates: a spam update, a Discover update, and perhaps most consequentially, a core update.
The industry’s instinct with this last announcement, predictably, was to look for the tactical explanation. Which pages dropped? Was it the title tag? The word count?
In my view, these are the wrong questions. Core updates recalibrate what Google rewards at a structural level, and chasing page-level explanations almost always misses the point.
This one accelerated a direction Google has been heading for years, and we can sum it up in three simple words: “Search rewards substance.”
Here are the 10 biggest takeaways for marketers looking to embrace this principle.
1. Topical authority now beats keyword coverage.
The old SEO playbook was to publish a page for every keyword variation your audience might search. This worked for a while, because Google couldn’t reliably tell the difference between a site that genuinely understood a subject (true “topical authority”) and one that had just written about it 47 times.
In recent years, Google’s algorithms have become more skilled at making that distinction. Now, sites with real depth in two or three topics are consistently outranking sites with thin coverage across twenty. Google is no longer counting pages—rather, it’s evaluating whether you actually own a subject. If your content strategy looks like a keyword spreadsheet, this update was aimed directly at you.
Action to take today: Audit which subjects you can genuinely own, and consolidate your content investment around those.
2. You need to obsess over “helpful content.”
When Google originally rolled out the helpful content system, it was positioned as an additional layer running alongside core ranking—a kind of "bonus penalty” for low-effort content. Many sites treated it as a separate problem to manage.
The latest update makes a structural change to this dynamic: The helpful content system has now been absorbed into core ranking. There’s no scenario anymore where a page fails the basic test of genuine human usefulness but survives on the strength of clean on-page optimization. The algorithmic bar and the editorial benchmark are now one and the same.
Action to take today: Moving forward, your content needs to address at least one of these three questions:
Does it provide original information, reporting, research, or analysis?
Does it provide substantially more value than other pages on the same topic?
Was it written by someone with direct knowledge of the subject?
If it can't check any of these boxes, you need to make some adjustments.
3. Low-quality SEO fluff is dragging down your entire domain.
Experience, Expertise, Authoritativeness, and Trustworthiness (E-E-A-T) has been Google’s standard of content quality for years. But in the past, it applied at the page level. Google would find the good content and rank it, even if your site also contained some filler.
Moving forward, Google plans to evaluate publishers the way a careful reader evaluates a publication—not by its best piece, but by its average one.
In other words, the low-quality SEO fluff your team was churning out in 2018 might be causing cumulative drag on your whole domain. If you haven’t audited your full content inventory lately, now’s the time.
Action to take today: Take a close look at your site as a whole and try to be objective. Delete genuinely unhelpful content—like articles that were produced for search engines rather than readers.
4. The problem with AI content isn’t actually the use of AI.
A persistent misread of recent Google updates is that AI-generated content is being penalized. In reality, Google is penalizing content where you’re not bringing any genuine human expertise to the table.
Using AI to improve efficiency and extend reach is not the problem. But if you’re using AI to replace thinking and human expertise—e.g., a model generates articles whole-cloth and an agent publishes them without any human expert ever taking a look or adding context—that content is more likely to get filtered.
Action to take today: Ditch a strategy that prioritizes maximum volume over human judgment. If no one on your team would stake their name on it, it probably shouldn't go out.
5. Ranking for a query and actually answering it are two different things.
For years, showing up for a query was the goal. Rank, get the click, done. Google has since built a more unforgiving feedback loop.
Today, Google tracks what happens after the click: how long someone stays, whether they bounce straight back to search, and whether they immediately run a follow-up query. A page that partially answers a question is losing ground to one that fully resolves it. If a reader leaves your page and goes back to Google, that’s a problem.
Action to take today: Read your top-traffic pages as if you're the person who searched for them. If you'd still have questions, so does Google.
6. Your brand needs to exist somewhere other than your own website.
The traditional SEO model was essentially self-contained: optimize your pages, build some links, publish content, watch the rankings move. Now, Google is also looking for corroboration—evidence that your brand has real presence and authority beyond what you’ve said about yourself.
Third-party mentions, citations in credible publications, community discussion… even modest external presence carries a meaningfully different authority signal than a well-optimized site that only talks about itself.
Action to take today: Assess where your brand is mentioned outside your own domain. If the list is short, that's your next content marketing priority.
7. Niche depth > aggregate breadth.
The aggregator model, where you create an enormous content library and broad topic coverage, was a defensible SEO strategy for a long time. Scale was protection.
But Google’s latest update continues a pattern that’s been building across multiple cycles: smaller, credible publishers with a defined subject area and a real audience are gaining ground against mega-sites that cover everything at surface depth.
Action to take today: If your content strategy covers 15 topics adequately, consider whether you'd be better served covering 5 topics well. Specificity is now a competitive advantage.
8. The spam update and core update were coordinated.
Google’s March spam update cleared out the most obviously manipulative signals; the core update then reevaluated everything that made it through.
In other words, surviving the spam update was not a clean bill of health. It just meant you cleared the first door. The core update is a stricter second filter to clear.
Action to take today: Pull your Search Console data and check your traffic timeline. A drop during the spam window points to a different problem than a dip between March 27 and April 8—knowing which filter tripped you up will help you prioritize the fix.
9. There’s no shortcut to recovery; expect it to take some time.
Google’s recovery guidance has always been consistent: Improve content quality, and the rankings will reflect it over time. That reads like a deflection, but it’s an accurate description of how the mechanism actually works.
Google runs smaller, unannounced updates continuously. A site that makes real, sustained quality improvements can see ranking recovery before the next major named update. The operative word here is “sustained”—Google’s systems need enough signal, over enough time, to confirm that a site has genuinely changed. Its own guidance states that the process will take several months.
Action to take today: Start the work now. Google's systems need months of steady signals to confirm quality improvement and adjust rankings or citations accordingly.
10. GEO and SEO are now pulling from the same infrastructure.
A lot of marketing teams still treat GEO as a separate workstream—something to figure out alongside traditional search, once the core SEO strategy is stable.
This update makes it harder to justify that framing.
Google’s AI features, like AI Overviews and AI Mode, use the same quality signals as core ranking. Clear entity definitions, structured factual claims, authoritative positioning, and answer-formatted content all contribute to both AI citation visibility and traditional search performance simultaneously.
Action to take today: Building for AI visibility and building for search ranking are the same investment—start treating them that way.
The playbook that worked in 2018 isn't just less effective; in many cases, it's now actively working against you. The good news is there’s a clear path forward and a new golden rule your marketing team should live by. Say it with us now: Search rewards substance.
Build accordingly.
Kishan Panpalia is a founding team member of Pepper and the company’s resident master of GEO.
AI news is as overwhelming as a Coachella influencer's content calendar. Here’s what you need to know this week.
OpenAI is going all-in on automated research. The company told MIT Technology Review it’s treating a fully automated AI researcher—an agent that can independently tackle large, complex problems—as its “North Star” for the next few years.
OpenAI launched a new $100/month ChatGPT Pro tier as a middle tier between its $20 Plus and $200 Pro plans, built for heavy Codex users who need longer, high-effort coding sessions. It’s priced identically to Anthropic’s Claude Max plan. Total coincidence, we’re sure.
OpenAI’s leaked memo is the corporate version of the “this is fine” meme. The company’s chief new revenue officer sent an internal memo meant to outline the company's strategy, but it came off sounding more like a petty venting session about rival Anthropic. It’s not helping with the company’s ongoing vibes problem.
Coachella’s hottest act? AI influencers. Generative AI has now progressed to the point where virtual creators are posting photorealistic content from Coachella without being there at all. Faking festival attendance is nothing new, but this is a decidedly more dystopian version of it.
Meta is building an AI clone of Mark Zuckerberg. The company is training a photorealistic, 3D-animated AI character on Zuckerberg’s mannerisms, tone, and publicly available statements, with the idea that it could offer advice to employees when he’s unavailable. Cue the soulless cyborg jokes.
Microsoft is building its own version of OpenClaw for enterprise. The company is testing OpenClaw-inspired features inside Microsoft 365 Copilot, aimed at enterprise customers and designed with stronger security controls than the famously risky open-source tool. As always, imitation is the sincerest form of enterprise software.
Chief Marketing Officer @ Fashion Institute of Technology (NYC, $360,000–$390,000 USD annual salary)
Head of Brand @ Apptegy (US Remote, no salary listed)
Sr. Brand Marketing Director @ Acorns (US Remote, $215,000–$230,000 annual salary)
