What actually changed in the May 2026 update
Google doesn't publish a changelog when it rolls out a core update. What we know comes from observed ranking changes, Google's own guidance documents, and patterns in what moved up versus what moved down.
The May 2026 update has two primary focuses:
1. Separating expert content from mass-produced content
Google has been working on this problem for years — the flood of content that looks helpful on the surface but was written by someone with no direct experience of the subject. The May 2026 update is their most aggressive attempt yet to surface content written by people who actually know what they're talking about, and demote content that could have been written by anyone (or anything).
The signals Google uses for this are the same E-E-A-T factors they've been developing since 2022 — but the weighting has shifted meaningfully. First-hand experience is now a primary signal, not a secondary one. Content that demonstrates real experience — specific examples, personal outcomes, genuine opinions based on doing the actual work — is outperforming generic informational content that covers the same topic at surface level.
2. Holistic Core Web Vitals scoring
This is the change most people haven't noticed yet. Previous Core Web Vitals scoring was somewhat forgiving — you could underperform on one metric if you were strong on the others. The May 2026 update introduced holistic scoring, where all three metrics need to pass:
If your site fails any one of these, you're at a disadvantage — even if you're perfect on the other two. Check yours right now at pagespeed.web.dev.
Who got hit and who didn't
Based on what's been observed across the industry since the update rolled out, the pattern is clear:
Sites that dropped tend to share a few characteristics: content that covers topics broadly without specific expertise or first-hand experience, high volume of pages with similar thin content, heavy reliance on AI-generated text without meaningful human editing or perspective added, and one or more failing Core Web Vitals scores.
Sites that held or improved have content written by people with demonstrable expertise in their subject, include specific examples and outcomes rather than generic advice, have strong technical fundamentals (fast load, stable layout, responsive), and have genuine authority signals — links from relevant sources, consistent brand presence, verifiable credentials.
For a Boston small business, the honest assessment is that most local service business sites are not at risk from this update. Your plumbing company's service pages aren't competing with mass-produced AI content farms. Where it matters more is if you're publishing a blog, trying to rank for informational queries, or competing in markets where larger players have been producing high volumes of AI content.
The honest answer on AI content
I'm going to give you the straight answer here because most of what you'll read about this is either alarmist or dismissive.
Google is not penalizing AI content because it was written by AI. They've said this explicitly and the data supports it. AI-generated content that is accurate, helpful, and reflects genuine expertise is not losing rankings.
What Google is penalizing is lazy content. The kind that gets produced when someone gives an AI a topic, hits generate, copies the output into a CMS, and publishes it without adding anything. That content is generic by definition — it's the average of everything the AI was trained on, not a perspective that comes from actual experience. Google has gotten very good at identifying it.
If you've been publishing AI-generated content and haven't been adding substantial expert perspective, real examples, and genuine editorial judgment — you need to audit those pages. Not because Google will penalize you for using AI, but because the content probably isn't as good as it needs to be to compete. That's the real issue.
The test I use: could this content have been written by someone with no direct experience of the subject? If yes, it needs work — regardless of how it was produced. If it reflects specific knowledge, real outcomes, and genuine perspective that only comes from doing the actual work, it's fine.
I use AI in my own content workflow. This post was researched and outlined with AI assistance. But every word of the analysis — the framing, the opinions, the specific advice — comes from 20 years of doing this work. That distinction is what Google is trying to measure, and it's a legitimate one.
What Boston small businesses should do right now
Here's the prioritized action list based on what this update actually targeted:
The bigger picture
The May 2026 update isn't a one-off event. It's the continuation of a direction Google has been moving for several years — and one that's accelerating as AI-generated content becomes cheaper and more widespread.
The businesses that will win in search over the next few years are the ones that are genuinely good at what they do and can demonstrate it clearly online. Not the ones with the most content, or the most backlinks, or the most optimized title tags — though those things still matter. The ones with the most genuine expertise, the clearest communication of it, and the technical fundamentals to let Google deliver it efficiently.
For a small business in Boston, that's actually good news. You have something that mass-content operations don't — real experience in a specific place, doing specific work, for specific customers. The May 2026 update is making that more valuable, not less.
Don't wait for the next core update to find out if your fixes worked. Google continuously re-evaluates pages outside of core updates — meaningful improvements to content quality and Core Web Vitals can show ranking movement within weeks. Start with the highest-traffic affected pages and work down the list.
If you want a second opinion on where your Boston business site stands after this update — what's at risk, what's fine, and what's worth fixing — I offer a free strategy call. I'll look at your actual data, not just give you a generic checklist.