SISTRIX yesterday published its April 2026 monthly review, released on April 30, 2026, by Johannes Beus, founder and managing director of SISTRIX GmbH, headquartered in Bonn, Germany. The newsletter covers three substantial topics: a new tool called Prompt Research designed to replace conventional keyword research in AI search environments, a data-driven study of AI Citation Drift conducted across three platforms, six countries, and 17 weeks, and the Q1 2026 IndexWatch results for German Google search visibility. Together, the three sections document a search landscape where the assumptions underlying traditional SEO and content strategy are breaking down faster than most practitioners have adjusted to.

Prompt Research: when keywords become conversations

The core argument behind Prompt Research starts with a simple observation. According to SISTRIX, the keyword "Gasgrill" - the German word for gas grill - generates 86,500 searches per month on Google. In AI search environments, that single keyword fragments into radically different expressions depending on the platform. On ChatGPT, a user might submit the full sentence "I am looking for a compact gas grill with a side burner for the rooftop terrace, budget 500 euros." On Google's AI Mode, the same underlying need might appear as "gas grill vs charcoal for beginners." In AI Overviews, the same topic surfaces as "gas grill winter storage."

These are not three variations of the same keyword. They reflect different intents, different moments in a purchase journey, and different platform conventions. According to SISTRIX, "keyword research at the level of individual prompts is no longer practical." The solution, as the company frames it, is to move up one level of abstraction: from individual prompts to topics.

That is what Prompt Research does. According to the April 30 newsletter, the tool clusters similar prompts into topic groups - which SISTRIX calls Topics - and makes AI search demand analyzable at that aggregate level. The German-language dataset underpinning the tool was built from over 62 million real user queries, organized into more than 1.4 million Topics. That is a substantial corpus. It means the tool is not drawing on synthetic or modeled demand but on actual observed behavior from AI search platforms.

Each Topic in Prompt Research delivers considerably more than a search volume number. According to SISTRIX, every Topic comes with intent classification, a Customer Journey Stage label, a Buyer Persona description, an Emotional Driver assessment, a list of "unasked questions" that users have but do not explicitly submit, and platform-specific prompt examples segmented by AI system - Google AI Mode, ChatGPT, and others. That is a content briefing structure, not a keyword list. The distinction matters. Traditional keyword research tells a content team what queries exist. Prompt Research, as SISTRIX describes it, tells a content team who is asking, why they are asking, where they are in a decision process, and what they are implicitly trying to resolve.

Prompt Research is included within the SISTRIX for AI/Chatbots module and is available at no additional charge to existing SISTRIX customers. A free trial account is also available for those without an existing subscription. SISTRIX had previously rolled out two related tools - AI Check and Prompt Monitoring - in the months before the April 2026 release. Prompt Monitoring, which allows users to track self-defined prompts across AI platforms, is also receiving an upgrade in this cycle: the updated version adds AI-assisted tag suggestions that automatically segment prompts into topic areas, enabling more targeted analysis by Buyer Journey stage, product segment, or other criteria. The updated Prompt Monitoring also introduces a domain-to-brand linking function, allowing citations to be measured alongside prompt visibility.

The launch is significant for the marketing community. PPC Land has been tracking how AI search features are reshaping keyword research and search measurement since AI Overviews began scaling in Germany in 2025. The problem SISTRIX is attempting to solve is real: measurement systems built on keyword-level search volume are structurally misaligned with AI search, where user intent is expressed in full sentences and platform behavior diverges sharply between systems.

AI Citation Drift: how stable are sources in AI search results?

The second major section of the April 2026 newsletter presents SISTRIX's own study of AI Citation Drift - a term that, according to the newsletter, has been discussed since mid-2025, when a study by Profound focused on monthly citation drift in the US market. SISTRIX extended that analysis with its own data, examining three platforms, six countries, and 17 consecutive weeks. The findings are specific enough to be operationally relevant.

The first finding concerns structure. According to SISTRIX, every AI Mode response contains what the company calls a "fixed core" and a "carousel." Across 86% of all prompts, there is a stable core of a small number of domains that appear consistently. The remainder of the citation set rotates - and rotates aggressively. According to the analysis, that non-core portion changes by 89% per week. The practical implication, as SISTRIX frames it, is that the right question for generative engine optimization (GEO) is not "am I appearing in responses?" but "am I in the core or in the carousel?"

The second finding separates the three platforms by architecture. AI Overviews, according to SISTRIX, is "completely stable" for just over half of all queries. AI Mode rotates 56% of citations per week. ChatGPT rotates 74% per week. These are dramatically different stability profiles. Treating AI Overviews, AI Mode, and ChatGPT as interchangeable surfaces for citation strategy would, according to SISTRIX, obscure more than it reveals.

The third finding addresses brand domains specifically. According to the analysis, 43% of brand-related queries maintain the presence of the brand's own domain across all 17 weeks of measurement. Being a recognized brand appears to confer a degree of citation stability. However, the neighboring citations - the other sources that appear alongside the brand domain - rotate at 70% per week. Being cited next to a brand does not confer the same anchoring effect. The co-citation position is highly unstable.

The fourth finding is perhaps the most immediately actionable for publishers. According to SISTRIX, only 1.4% of news articles that are cited as sources in AI responses remain in the citation set on a permanent basis. The newsletter is direct: "Those who plan citation strategy around editorial news content are planning wrong. Evergreen content systematically survives better." For news publishers already contending with the declining click-through rates documented in PPC Land's coverage of SISTRIX's February 2026 analysis, the citation drift data adds another layer of structural difficulty. Capturing an AI citation through a timely news article is not a durable outcome. The citation will almost certainly disappear within weeks.

The fifth finding addresses geographic scope and persistence. Drift rates across the six countries in SISTRIX's study run consistently between 54% and 59%. Crucially, these rates show no sign of declining across the 17-week measurement window. According to SISTRIX, this is not an introductory effect that stabilizes over time. It is a structural characteristic of the platforms themselves. The Profound study that initially sparked interest in AI citation drift focused on the US market and monthly timeframes. SISTRIX's 17-week weekly measurement across six countries provides a broader and more granular picture. The conclusion is the same but more firmly established: AI citation behavior is persistently volatile, globally.

For the marketing community, these findings land against a backdrop of intensifying interest in GEO as a discipline. PPC Land has tracked how top domains in Google AI Mode are selected, and how SISTRIX identified patterns of Google self-preference in AI Mode citations following the feature's German launch in October 2025. The citation drift data complicates the strategic picture. Even if a site achieves citation on a given AI platform, holding that citation reliably over time - at least outside of the stable core - appears to be structurally difficult.

IndexWatch Q1 2026: UGC platforms dominate, migrations destroy visibility

The third section of the newsletter covers SISTRIX's IndexWatch for the first quarter of 2026, an analysis of German Google search visibility conducted by SEO expert Jolle Lahr-Eigen. The headline finding, according to the April 30 newsletter, is that the winners in Google search for Q1 2026 are not traditional publishers. User-generated content (UGC) platforms - YouTube, Reddit, and Facebook - are gaining organic reach at a pace that conventional websites struggle to match.

On the losing side, the IndexWatch identifies two distinct categories of damage. The first is self-inflicted: website migrations handled incorrectly. According to SISTRIX, redirecting migration traffic to generic collection pages rather than to matching content destroys visibility in a way that is very difficult to recover from. The second is regulatory. The newsletter specifically names the Digital Services Act, noting that operators without valid licenses can disappear from search results entirely. This is a concrete and documented consequence. The DSA's enforcement mechanisms have search visibility implications that go beyond compliance risk in the abstract.

The technical dimension of the Q1 visibility analysis is consistent with SISTRIX's longer-running messaging around indexing quality. According to the newsletter, indexation errors and broken redirects are avoidable risks that can eliminate rankings overnight in the worst case. These are not novel observations, but they acquire fresh urgency when placed alongside data showing that organic search recovery from algorithmic or structural damage is slow and uncertain. PPC Land's coverage of the March 2026 core update documented that nearly 80% of top-three URLs changed positions during the cycle - and the SISTRIX data for that update showed how authority and brand strength were increasingly the determining factors.

Additional items from the April 2026 newsletter

The newsletter also summarizes several broader developments. The Google Core Update for March 2026, according to SISTRIX, hit online shops, language and education tools, and recipe portals hardest, with four times more losing domains than gaining ones. The central message SISTRIX draws from the data is that authority beats interchangeability - established brands and official sources gain, while generic providers lose. PPC Land covered the March 2026 core update in detail from its initial release on March 27 through its completion on April 8 and subsequent volatility analysis.

The newsletter also references a phrase from Google's Search Central event in Toronto: "Good SEO is good GEO." According to SISTRIX, this is only partially accurate. While SEO fundamentals and authenticity are relevant to AI search performance, platforms like ChatGPT, Claude, and Google's own AI Overviews operate under their own distinct rules - a point reinforced by the citation drift data in the same newsletter.

On June 15, 2026, Google will begin penalizing sites that manipulate the browser back button to redirect users to unwanted pages. According to SISTRIX, this is a long-overdue measure that will apply either manual or algorithmic penalties. The practice damages user experience significantly. Its enforcement has been absent until now.

SISTRIX also updated its workshop program in response to the pace of change in search. The workshops have been revised to reflect what the company calls "the new era of search," and participation is free for most existing SISTRIX customers. A new SectorWatch analysis on summer holiday 2026 search trends was published by SEO expert Helen Schrader, focusing on intent strategy and domain performance in a year of geopolitical uncertainty.

Timeline

  • Mid-2025: The term "AI Citation Drift" enters industry discussion, prompted by a Profound study examining monthly citation volatility in the US market
  • August 2025: SISTRIX documents AI Overviews appearing on 17% of German keywords, alongside Reddit's surge to 13th place among Germany's most visible domains
  • October 31, 2025: Google launches AI Mode in Germany; SISTRIX publishes same-day analysis showing YouTube at 40.08% of citations and Google.com at 31.69%, documenting systematic patterns of Google self-preference
  • December 11-29, 2025: Google December 2025 core update runs 18 days and produces severe publisher visibility losses; Indian and UK publishers see declines of 65-85%
  • March 1, 2026: SISTRIX February 2026 review published, documenting 265 million monthly organic clicks lost in Germany due to AI Overviews, based on analysis of more than 100 million keywords; top organic position CTR falls from 27% to 11% when AI Overviews are present
  • March 24, 2026: Google releases March 2026 spam update at 12:18 PDT, completing in approximately 19.5 hours - the fastest spam update on record
  • March 27, 2026: Google releases March 2026 core update at 02:14 PDT, three days after the spam update
  • March 31, 2026: SISTRIX March 2026 review published, documenting Google's structural consolidation of AI search and Gemini traffic tripling in H2 2025
  • April 8, 2026: March 2026 core update completes after 12 days; NYT and Guardian identified as major visibility winners
  • April 15, 2026: SE Ranking data confirms March 2026 core update drove 79.5% URL shifts in top-three positions, the highest on record
  • April 30, 2026: SISTRIX April 2026 newsletter published by Johannes Beus; announces Prompt Research tool built on 62 million real user queries in 1.4 million Topics; releases AI Citation Drift study covering three platforms, six countries, 17 weeks; publishes Q1 2026 IndexWatch showing UGC platforms dominating German Google search

Summary

Who: Johannes Beus, founder and managing director of SISTRIX GmbH, a search analytics company based in Bonn, Germany, operating for 18 years in the European search intelligence market.

What: The April 30, 2026 SISTRIX monthly newsletter announces three major developments. First, the launch of Prompt Research, a tool that clusters AI search prompts into topic groups using a database of over 62 million real user queries across more than 1.4 million Topics for the German market, available free to SISTRIX customers. Second, the publication of a 17-week AI Citation Drift study covering three AI platforms and six countries, finding that 89% of the non-core citation set in AI Mode rotates weekly, that only 1.4% of news article citations are durable, and that drift rates hold consistently between 54% and 59% globally with no sign of stabilization. Third, the Q1 2026 IndexWatch, showing UGC platforms including YouTube, Reddit, and Facebook as the dominant gainers in German Google search visibility, with site migration errors and Digital Services Act enforcement identified as primary sources of visibility loss.

When: The newsletter was published on April 30, 2026. The AI Citation Drift study was conducted over 17 consecutive weeks from an unspecified start date in early 2026. The IndexWatch covers the first quarter of 2026, from January through March.

Where: SISTRIX GmbH is headquartered at Thomas-Mann-Str. 37, 53111 Bonn, Germany. The AI Citation Drift study covers six countries across three AI platforms: AI Overviews, Google AI Mode, and ChatGPT. The IndexWatch covers the German Google search market. Prompt Research is currently built on German-language data.

Why: The newsletter responds to a structural shift in how users interact with search. Conventional keyword research was built on the assumption that users type short queries into a search box. In AI search environments, users submit full-sentence prompts with specific context, and each AI platform handles those prompts differently. At the same time, the citation volatility data raises questions about the durability of any content strategy built around achieving AI citations through news or timely content, given that 98.6% of cited news articles do not remain in the citation set permanently. SISTRIX's tools and data aim to give practitioners measurable signals in an environment where measurement itself has become structurally harder.

Share this article
The link has been copied!