Build an SEO Strategy
How to Build an SEO Strategy That Actually Works in 2026
A step-by-step framework that’s honest about what changed, what still works, and why most “SEO guides” were already dead when the March 2026 update landed.
- The March 2026 Core Update hit thin content and mass-AI output hardest. Sites with genuine author expertise and original data mostly survived — or gained.
- AI Overviews are eating informational click-through rates. Being cited inside them is now worth more than ranking position 3 in blue links on the same query.
- E-E-A-T isn’t a checklist. It’s a production discipline — and the March 2026 update made the gap between sites that get this and sites that fake it much wider.
- This guide covers 9 sequential steps. Don’t skip the audit in Step 1. Everything else is wasted effort without a real diagnosis first.
The March 2026 Core Update dropped, and I spent the first two weeks after rollout going through the wreckage. Site after site had lost 40–60% of organic visibility — not because of technical failures, not because of spam links, but because they’d been producing content that looked like SEO content rather than content that actually helped anyone. Google got better at telling the difference. That’s the whole story.
What survived? The same thing that always survives: pages written by someone who actually knows the topic, structured clearly, updated honestly, and built as part of a coherent subject matter library rather than a keyword scatter plot. None of that’s new. But the gap between doing it and faking it just got a lot wider. This guide is a step-by-step framework for the real version.
The March 2026 Core Update rolled out alongside a simultaneous Spam Update — a combination Google used in 2024 as well. The paired rollout amplified volatility, especially in content-heavy verticals. Sites already weakened by the spam filter took compounding losses from the core update. These are two separate systems. You need to figure out which one — or both — hit you before deciding how to respond. Treating a spam penalty like a quality issue (or vice versa) can make things worse.
Start with Diagnosis: Audit First
Every instinct after a core update hit is to do something — fast. Resist it. Google recommends waiting 7–10 days after full rollout for rankings to stabilize before making substantive changes. Rush edits during active volatility and you muddy your diagnostic data. Worse, you might “fix” something that wasn’t the actual problem.
I’ve watched teams spin their wheels for two months because they immediately rewrote their landing pages when the problem was internal linking and orphaned cluster content. Don’t do that. Audit first, act second.
What to audit, in priority order
Probable Priority weighting based on pattern analysis of March 2026 Core Update winners and losers — not an official Google signal weighting.
Fix technical debt first. Not because it outranks content quality in importance — it doesn’t — but because it’s binary and fast to verify. A page that fails Core Web Vitals is failing before a human even reads a word. Eliminate that as a variable before you diagnose anything else.
Go to Performance → Search Results → Compare March 27, 2026 onward vs. the same 30-day period one month prior. Sort by impressions loss, not clicks. AI Overviews can maintain your impressions while completely killing your CTR — and those look like two very different problems with very different fixes. Don’t misread one for the other. Use the ContentEvaluator audit tools to cross-reference GSC data against page-level quality scores.
Align SEO Goals with Revenue
“More organic traffic” is not a goal. It’s a vanity metric dressed up as a strategy — and it fails the moment AI Overviews start absorbing impressions without sending clicks. I’ve had this conversation with clients more times than I’d like: traffic is flat or down, conversions are up. Or traffic is up, revenue is flat. The traffic number tells you almost nothing on its own in 2026.
Organic CTR dropped from 1.76% to 0.61% on queries where AI Overviews appeared — a 61% decline. A site can hold position 3 in blue-link rankings and lose the majority of its clicks to the Overview above it. Position doesn’t equal clicks anymore. The CTR curve has been restructured.
The counterweight: brands cited inside those AI Overviews earn 35% more organic clicks and 91% more paid clicks than uncited brands on the same query. Probable These figures are sourced from Bain Research via secondary sources — I haven’t independently verified the methodology, but the directional pattern is consistent with what I’ve observed across client accounts.
Here’s what actually-useful SMART goals look like in this environment:
- Increase organic traffic from commercial-intent queries by 40% (informational zero-click volume counts for almost nothing in revenue terms)
- Achieve citation in AI Overviews for 25% of your core topic queries within 12 months
- Grow branded search volume by 30% — a signal that indicates real entity recognition, not just keyword ranking
Map every content initiative back to one of these before you greenlight it. If a cluster doesn’t move revenue, qualified traffic, or citation rate — deprioritize it. That sounds obvious. It’s somehow radical in practice.
Research Audience Intent and Build Entity Foundations
Pure keyword research is a 2019 strategy. Still useful — don’t throw it out — but insufficient on its own. What Google’s systems actually evaluate now is entity relationships: who you are, what concepts your site owns, and how clearly your content demonstrates you belong in the conversation on your core topics.
The shift in practice looks like this: instead of asking “what keywords do I want to rank for,” start asking “what entity do I want to be in Google’s knowledge graph for this topic?” That reframe matters because entity recognition is what determines whether your content appears in AI Overviews, knowledge panels, and related searches — not just blue-link position. Probable Entity recognition as a ranking concept is widely accepted in the SEO community; its specific weight relative to other signals in Google’s current systems isn’t publicly documented.
Building an entity hub — the actual process
- Start with Search Console: find the real user questions already driving impressions on your site — not just clicks. Those impression queries are your entity map.
- Group them into intent clusters from informational through commercial and transactional. Each cluster anchors a content section, not a single page.
- Build a central pillar page that defines your core topic with clear language, structured headings, and internal links to every supporting page. See the pillar page checklist for structure requirements.
- Include first-party experience: your own test results, named case studies, original data — not just restated industry summaries. This is what “Experience” in E-E-A-T actually means at the page level.
When Google consistently places your site in AI Overviews for a topic cluster — not just for branded queries — you’ve reached entity-level authority for that cluster. That’s the goal. Not position 1 for one keyword. Recognition as the site that understands this topic. You’ll also see it in knowledge panel appearance for your brand name and in “People also search for” associations that match your topic cluster.
Build Content Clusters for Topical Authority
The March 2026 update reinforced what practitioners had been seeing for months: isolated blog posts can still rank for low-competition queries, but sustainable topical authority — the kind that survives algorithm updates — requires interconnected content clusters where each piece adds something the others don’t.
Evertune’s post-update analysis framed this as an “Information Gain” re-weighting — content gets evaluated not just for accuracy but for how much it adds to what’s already indexed. I’d describe it differently: Google got better at recognizing when a site is covering a topic versus owning it. Coverage is 20 thin posts on adjacent keywords. Ownership is 8 deep, interconnected pieces that collectively leave no major question unanswered. Probable
| Layer | Format | Word count | Primary job |
|---|---|---|---|
| Pillar page | Comprehensive guide | 3,000 – 5,000+ | Define the topic, establish authority, anchor internal links across the cluster |
| Cluster pages (8–15) | Deep-dive per sub-intent | 1,200 – 2,500 | Each targets one specific question with original data, screenshots, or tested examples |
| Comparison pages | Side-by-side analysis | 1,500 – 3,000 | Capture commercial intent; E-E-A-T signals via real testing with named conditions |
| FAQ pages | Structured Q&A | 800 – 1,500 | AI Overview citation bait; conversational and voice query capture |
Write a direct, clear answer to the page’s primary question within the first 200 words. Not after background, not after the history of the topic — right up front. AI systems pull concise, factual answers for Overviews, and they pull from near the top of the page. Burying your best answer in paragraph 6 is effectively hiding it from AI-mediated discovery. You can use the ContentEvaluator scoring tool to check whether your opening paragraph is citation-ready.
Technical SEO and AI Readiness
Technical SEO in 2026 has two audiences: Google’s crawl-and-index systems, and AI parsing systems that extract content for citations. Fortunately, what’s good for one is good for the other. Clean structure, fast pages, and clear schema markup help both. The only thing I’d call new in 2026 is the AI content labeling schema — everything else is table stakes.
- Core Web Vitals: Established LCP under 2.5 seconds, FID under 100ms, CLS under 0.1. Use PageSpeed Insights and the CWV report in Search Console. Mobile failures are penalized harder than desktop.
- Schema markup: Implement FAQ, HowTo, Article, and Organization schemas. The direct ranking boost is debated, but structured data helps AI systems understand what a page is about and who produced it — that’s increasingly the more important function.
- Internal link architecture: Cluster pages should link up to the pillar page and across to closely related cluster pages. Use descriptive anchor text — not “click here,” not just the exact-match keyword. Read the internal linking guide for anchor text patterns that work.
-
URL structure: Keep it clean and topically coherent.
/seo-audit-guide/is better than/blog/post-1847/for both crawlers and human readers who decide whether to click based on the URL alone. - AI content labeling: Speculative Google supports schema for labeling AI-assisted content. Whether this affects rankings directly is unconfirmed. But hiding AI involvement is increasingly detectable, and the reputational cost of getting caught outweighs any short-term upside.
- Crawl budget hygiene: Audit for orphaned pages, redirect chains longer than 2 hops, and near-duplicate URL variants without canonical tags. These drain crawl efficiency and dilute topical authority signals across your cluster.
Produce People-First Content That Demonstrates E-E-A-T
This is the one that everything else depends on. Every other step in this guide is scaffolding — E-E-A-T is the foundation. And I want to be specific about what that means in practice, because most published “E-E-A-T guides” treat it as a checklist you bolt on after writing. It’s not. It’s a production discipline that shapes how you plan, research, draft, and publish.
The sites that survived March 2026 didn’t write better SEO content. They wrote better content, full stop — and then made sure it was findable.
Based on post-update pattern analysis — April 2026Here’s what I actually mean by each E-E-A-T signal at the production level:
| Signal | What Google actually evaluates | What it looks like in production |
|---|---|---|
| Experience | Evidence the author has done what they’re describing — not just researched it | Screenshots of your own results, “I ran this test and found…” framing, before/after case studies with named conditions |
| Expertise | Demonstrated knowledge depth — mechanism-level understanding, not surface summary | Cite primary sources not secondary aggregators; explain why something works, not just that it does; acknowledge edge cases |
| Authoritativeness | Credible external validation of the site and the named author | Named author bios with verifiable credentials; external mentions in reputable publications; guest bylines that link back |
| Trustworthiness | Transparency about limitations, sources, and where you might be wrong | Dated update stamps; methodology disclosures; “What Could Be Wrong” sections; no misleading claims; clear ownership and contact |
The March 2026 update didn’t ban AI content. But it deployed stronger semantic evaluation to identify content that is generic, adds no original perspective, and fails to contribute information beyond what’s already indexed. The test I use: can you answer “what does this page know that similar pages don’t?” If you can’t answer that question in one sentence, the page has a problem — and that problem existed before AI, it’s just more common now.
Optimize for AI Overviews and Search Everywhere
Traditional SERP position is one visibility channel. It’s not even the most important one for informational queries anymore — AI Overviews appear on an expanding share of searches, and as the Seer data shows, being cited there outperforms ranking position 3 in the blue links on the same query.
GEO — Generative Engine Optimization — is the practice of structuring content so it gets cited by AI systems, not just indexed by traditional crawlers. Here’s the thing nobody wants to admit: the standard for GEO is identical to the standard for E-E-A-T. There’s no separate trick. There’s only writing well enough to get cited. Probable
What actually gets cited in AI Overviews
- Pages that answer the query directly in the opening paragraph — specific, factual, no warmup preamble
- Content with clear headings, tables, and structured information AI can extract and present without rewriting
- Sites that have earned mentions in third-party publications as a source — not just backlinks, but citations
- YouTube videos with full transcripts for the same queries (Google surfaces YouTube increasingly in Overviews for how-to content)
- LinkedIn articles for B2B topics — analysis has identified LinkedIn as one of the top AI citation sources for professional queries, though the mechanism isn’t fully public
Build Sustainable Brand and Authority Signals
Links still matter. They’re not going anywhere. But the shift since 2024 is real: brand strength — direct traffic volume, branded search query growth, unlinked mentions from credible sources — carries more weight relative to raw link count than it did even two years ago. Probable I’ve seen sites with 50 highly relevant, topically matched links outrank competitors with 500 generic directory links. Every time. The quality-vs-quantity shift in links is established; the exact weighting is not public.
- Digital PR with original research: One well-covered research piece generates more relevant links than 50 cold outreach emails. Publish a study, survey, or original dataset that journalists and bloggers will cite. Make it specific enough to be quotable, simple enough to explain in a tweet.
- Topical depth over volume: 3 excellent pieces per month on a tight topic cluster beats 20 generic posts per month on ten loosely related topics. Topical coherence is a ranking signal — not just a best practice.
- Community participation: Industry forum discussions, newsletter guest slots, and social commentary generate unlinked brand mentions. Google reads these as entity trust signals. You can’t manufacture them at scale; you can earn them consistently.
- Branded search velocity: Track this quarterly in Search Console. Month-over-month growth in queries that include your brand name is the clearest leading indicator that E-E-A-T work is sticking — before it shows up in blue-link rankings.
Measure, Iterate, and Scale Quarterly
SEO in 2026 isn’t a launch-and-leave system. Core updates happen more frequently and with more granular targeting than five years ago. An AI Overview share that was 30% of informational queries in 2024 is approaching 60%+ now, which means a traffic measurement model from 2024 can badly misread a 2026 performance picture. Two metrics most sites still aren’t tracking — but should be:
- AI citation rate: What percentage of your core topic queries surface your content in AI Overviews? Track manually, or use monitoring tools that scrape AI Overview appearances.
- Branded search velocity: Month-over-month growth in brand-name queries. Leading indicator that trust signals are compounding before rankings respond.
Common Pitfalls to Avoid
Scaled generation without meaningful editorial oversight. The March 2026 update’s semantic filter is specifically calibrated for this. Your voice and judgment need to be visible in every published page — not just in the metadata.
100 thin cluster pages consistently underperform 10 deep, original ones in post-March 2026 data. Depth beats breadth. This is the thing everyone knows and almost nobody does.
60% of searches end without a click. Building an SEO strategy around click projections alone is measuring the wrong output. Brand awareness from AI Overview appearances has real value even without a click — model your attribution accordingly.
In YMYL verticals — health, finance, legal, SEO — unattributed content carries a measurable trust penalty. Named authors with verifiable credentials aren’t optional in these categories. Frankly, they probably shouldn’t be optional in any category.
Google search is one surface. AI Overviews, YouTube, Reddit, LinkedIn, and newsletters are others. Your content strategy needs to feed all of them — or at minimum, not actively undermine them.
Targeted directly by the March 2026 Spam Update. If you’re buying expired domains to build links or host content, stop. Detection has improved significantly and the risk-reward ratio is now completely inverted.
Your 30-Day Launch Plan
Then repeat the cluster process every quarter. Consistency is what compounds. A site that publishes one excellent cluster per quarter for two years will outrank a site that published fifty mediocre posts in the same window — not because Google rewards patience, but because depth of topical coverage genuinely accumulates into measurable authority.
The CTR and citation-benefit statistics (61% CTR drop, 35% more organic clicks, 91% more paid clicks) come from Seer Interactive and Bain Research, but I’m citing them via secondary aggregator sources — not the original methodology documents. The directional findings are consistent with what I observe in client data, but treat the specific percentages as illustrative, not precise benchmarks for your site.
The “entity recognition” framing and “Information Gain re-weighting” descriptions are industry consensus interpretations of Google’s behavior — not confirmed signal descriptions from Google. Google has never published a ranked list of what its systems actually evaluate. Everything in that category is inference.
The March 2026 update is recent enough that definitive post-update analysis is still limited. Some findings may evolve as more data accumulates over the coming months. I’ll update this article in Q3 2026 with revised pattern data.
This guide reflects my experience primarily in B2B SaaS and content-heavy information sites. E-commerce SEO, local SEO, and certain YMYL verticals have materially different dynamics — not everything here transfers one-to-one.
Frequently Asked Questions
The sites that win in 2027 won’t be the ones that published the most — they’ll be the ones that built the right entity foundation before the next update hit and spent less time recovering.
