The June 2026 edition of Réacteur is online!

On the program for this edition Reactor from June 2026:

We warmly thank our 6 authors for their responsiveness and above all the quality of the articles published: Indiana Aflalo, Lou Pichard, David Groult, Erwann Cardon, Killian Le Moal and Sylvain Deauré.

As a bonus : as usual, this new edition also welcomes our eve of June 2026 and a presentation tools that caught our attention these last few weeks!

👉 Access the full edition 🔒

And, if this is not yet the case, subscribe !

SEO & Salesforce B2C Commerce: taming the power of SFCC for search

By Lou PichardSEO director at Eskimoz Bordeaux

To discover in this article:

• SFCC is everywhere (Adidas, Lacoste, Sandro, Maje…), but remains one of the least documented CMS on the market. 2025 is a game changer: redesign of the help center and arrival of the AgentForce AI agent directly in the doc.
• SFRA vs. Composable Storefront : two architectures, two radically different SEO logics. Sitemap, canonicals, SSR, Core Web Vitals… everything changes depending on the Core Model chosen.
• SFCC’s most powerful native modules finally explained clearly : URL Rules, Page Meta Tag Rules, and especially the three redirection systems with their order of priority (the one that unlocks debugging hours).
• No .htaccess, no native Schema.org : the real limits of the platform and how to get around them with targeted development.
• The Agentic Commerce shift is coming. With the Stripe-OpenAI partnership, your SFCC catalog will soon have to be readable by AI agents even before being read by Google.

💜 What we liked:

A real work of demystification of a platform deemed closed and where knowledge is transmitted mainly internally or in agencies. Lou Pichard’s article goes into operational detail: order of priority of redirection modules, dependency between catalog data and meta tags, methodology for dividing SEO specifications into three bricks (native, configuration, custom development). Exactly the kind of benchmarks you look for in vain in the official Salesforce documentation. No hollow theory, but precise thresholds and actionable tools. An uncompromising article, which will make more than one think!

👉 SEO & Salesforce B2C Commerce: taming the power of SFCC for search

By David GroultHead of SEO and Erwann CardonSenior SEO Consultant @NOISE

To discover in this article:

• 8 brands, 8 sectors, 8 different logics. So many concrete examples of product sheets that really transform the purchasing decision.
• The ASOS waist simulator analyzed in detail. How a brand transforms size uncertainty into personalized social proof, with a direct impact on returns and margin.
• LDLC or how a technical characteristic becomes an SEO lever. Linking each product specification to a facet page creates a mesh that serves both the user and the crawler.
• A reading grid in 5 types of information gains to know precisely what is missing on your files.
• A complete audit checklist to evaluate your product sheets with a fresh perspective, beyond the title and H1.

💜 What we liked:

The angle adopted is rare: instead of a list of generic best practices, the article dissects eight real cases and explains precisely why each device works, without falling into superficial name dropping. The distinction between the different types of information gains provides an immediately actionable reading grid, and the final warning on the risk of overload avoids the classic pitfall of “always add more content”.

👉 E-commerce product sheet: why gaining information becomes a real SEO, Shopping and AI lever

Becoming a source that AIs cite: method, engine-by-engine analysis and limits

By Indiana AflaloSEO/GEO consultant and founder ofIndHack

To discover in this article:

• A victory at 139 mentionsbut increased to 80% by only two out of five AIs. Engine-by-engine analysis reveals radically different behaviors between Claude, Mistral, ChatGPT, Perplexity and Gemini.
• The seven pillars of a complete GEO method…with supporting code examples.
• “Fan-out” explained through concrete architecture. A hub and ten pages aligned with the sub-intentions that AIs generate in the background to construct their responses.
• The “winner-takes-all” effect on Mistralobserved in real conditions. How an engine can cite up to six URLs from the same domain in a single response, once the source is identified as dominant.
• Failures documented without filter. Reddit banned, Wikidata refused, robots.txt which blocks AI bots for 24 hours… And what to remember.

💜 What we liked:

Rare transparency on what did not work as much as on what did. Indiana Aflalo does not just announce a victory, it dissects why Claude and Mistral reacted while Gemini and Perplexity remained hermetic, with precise hypotheses for each engine. The distinction between what is transposed to a real site and what is not transposed (the artificial context of the competition) gives the article a real methodological value, beyond the simple performance story.

👉 Becoming a source that AIs cite: method, engine-by-engine analysis and limits

Why do you continue to produce content… for nothing [partie 2]

By Killian Le MoalLead GEO-SEO Strategist & Innovation @ Peak Ace agency

To discover in this article:

• A “stress test” in 3 questions to pass to ChatGPT on your first 200 words.
• The Anatomy of a Quotable Chunkwith a stunning before/after example. Two versions of the same length, radically different citation potential.
• The fan-out figures by LLM that change everything. From 22.6% at ChatGPT to 98.7% at Gemini, with concrete consequences on the way you structure your pages.
• Three “false good ideas” from GEO deconstructed with data…What really works and what is a waste of time.
• A new KPIs grid to measure visibility in generative AI, in addition to classic SEO.

💜 What we liked:

The very operational side of this second part: each concept is accompanied by a protocol, a concrete example or a sourced figure, up to the actionable checklist at the end of the article. The deconstruction of GEO myths (llms.txt, Markdown, long pages) contrasts with the ambient discourse and is based on recent studies rather than intuitions. This is the kind of article you keep on hand to audit your own pages.

👉 Why do you continue to produce content… for nothing [partie 2]

By Sylvain DeauréCo-founder of 1492.vision

To discover in this article:

• More than 20 Discover pipelines with radically different logicsbased on the analysis of 42 million cards. The real question is no longer “ am I in Discover “, but ” in how many pipelines am I visible “.
• The multiplier effect of multi-pipeline. 58% of French URLs appear in 2 or more pipelines, with outliers reaching 12 to 14 simultaneously.
• The West-France model deciphered : how a double local + national anchorage makes it possible to reach 25 distinct pipelines, a record for a traditional publisher.
• Which structurally limits the multi-pipeline : pure product testing, daily sport, lifestyle… with the concrete levers to get out of it.
• An emerging signal to watchiller: a social intake pipeline grew 33x in three months, 72% powered by X.com.

💜 What we liked:

The scorecard by publisher profile (national press, regional press, tech/review, sport, lifestyle, video) which precisely distinguishes natural pipelines and those to be conquered: it is immediately actionable, we can situate ourselves and identify our margins for progress. The concrete cases (Le Monde, Ouest-France, Frandroid, L’Equipe, programme-tv.net) with their detailed pipeline footprints offer a reading grid that cannot be found anywhere else.

👉 How to multiply your presence in Google Discover: the multi-pipeline effect

Like every month, this new edition of Reacteur is accompanied by a section dedicated to the latest tools to test And the articles that caught our attention on the web !