Post-Rank SEO After Google Removes num=100: Win AI Overview Citations

What changed and why it matters\n\n- On September 18, 2025, Google confirmed that the results per page parameter is no longer supported. That move effectively ended deep page pulls and disrupted many rank trackers. See the report on the change in Google ends num=100 parameter.\n- In parallel, Google expanded AI Overviews and introduced a distinct mode that answers directly and links to sources. Read the official description of the AI Mode search experience.\n- Many site owners noticed that impressions in Search Console dropped even when traffic and top 10 positions stayed stable. The visibility funnel has shifted toward answer surfaces where links and citations are treated differently from classic organic results.\n\n## Post-rank reality\n\nClassic rank is now a partial proxy for visibility. A meaningful share of attention flows into AI Overviews, AI Mode, and rival answer engines. Your reporting should reflect whether you are cited and how visible those citations are to users.\n\n## Build an AEO-first measurement stack\n\nGoal: quantify how often you are cited, how much share you win among citations, and how visible your links are inside AI answers.\n\n### 1) Answer inclusion rate (AIR)\n\n- Definition: percent of tracked queries where your domain appears as a cited source inside an AI answer surface.\n- Where to measure: Google AI Overviews, Google AI Mode, Microsoft curated answers, Perplexity, and other answer engines.\n- Collection options: use vendors that expose AI answers programmatically. SerpApi and DataForSEO support AI surfaces and asynchronous retrieval. Validate results with headless spot checks.\n\n### 2) Citation share\n\n- Definition: share of all citations on the answer canvas that point to your domain versus your competitor set for a given query list and location.\n- How to compute: per query, count unique citing links to your domain, divide by total citations, then aggregate by topic cluster. Track by engine and by vertical.\n- Why it matters: engines vary in what they cite and how much attribution they expose. Optimize toward the engines most important to your audience.\n\n### 3) AI Overview link prominence\n\n- Definition: a visibility score for your link based on placement and interaction state inside AI Overviews and AI Mode.\n- Starter scoring model: begin at 100 for the first visible source chip above the fold on load. Subtract 20 points per position. Apply a 0.5 multiplier if an expansion or More action is required. Add 10 points if your link is inline within the summary text. Calibrate with user testing and scroll depth.\n\n## Suggested instrumentation and workflow\n\n- Query sets: rebuild tracked sets by intent and journey stage. Include head, midtail, entity, and how-to queries that frequently trigger AI answers.\n- Locations and devices: measure per country and per device because answer frequency and layout vary.\n- Data capture: rely on vendor APIs for AI Overviews and AI Mode where available. Store raw JSON for auditability and reprocessing. Use headless verification to spot check rendering states that affect prominence.\n- Dashboards: chart AIR, citation share, and prominence by engine and topic. Add deltas against classic rank and Search Console clicks to monitor decoupling.\n- Alerts: trigger when inclusion drops for priority queries or when a competitor gains majority citation share on key topics.\n\nTo strengthen licensing and eligibility signals that support trustworthy citations, pair this stack with the guidance in the License-Led AEO with Cloudflare signals.\n\n## Optimization playbook for inclusion and citations\n\n- Make answers easy to lift: lead with concise, fact-rich paragraphs and lists that directly answer target questions. Support with verifiable claims and clear entities.\n- Strengthen source credibility: maintain author bylines, dates, references, and structured data that align with on-page content.\n- Be present where engines shop: rival engines highlight citations and often pull from a mix of publishers, docs, and community sources. See channel-specific tactics in the Perplexity Comet optimization playbook.\n- Design for voice and live surfaces: ensure short, attributable snippets that read well aloud and include visible citations. Use the patterns in Voice AEO for Google Search Live.\n\n## Executive timeline\n\n- Weeks 0 to 2: freeze pre-change baselines. Stand up AIR and citation share collection for AI Overviews, AI Mode, and at least one rival engine. Re-baseline Search Console trends after the September 2025 change.\n- Weeks 3 to 6: ship prominence scoring, add competitive panels, and run 10-query user tests to calibrate above-the-fold thresholds.\n- Quarter 2: migrate reporting from rank-first to answer-first. Tie inclusion and citation share to assisted conversions and content investment decisions.\n\n## Key takeaways\n\n- The end of num=100 reduced deep SERP visibility and broke many trackers overnight. Treat classic rank as directional, not definitive.\n- AI Overviews, AI Mode, and rival engines now mediate a growing slice of attention. Plan to win citations and prominence, not just ranks.\n- Build AIR, citation share, and prominence into your core KPI set to capture the reality of post-rank SEO.