Two recent Google updates point in the same direction for B2B marketers: measure your content wherever it appears in Search, and invest in one well-built website rather than parallel versions for machines. Search Console now reports on social and video posts, and Google's John Mueller has advised against building separate markdown copies of your site for AI systems.
Both stories were covered in Search Engine Journal's SEO Pulse. Here is a short summary of each, followed by what we think B2B teams should do about them.
Search Console now reports on social posts
Google is rolling out a new Search Console property type, called platform properties, that shows how your posts on Instagram, TikTok, X and YouTube perform in Google Search and Discover. It works even if you do not have a website. According to the report, the feature was previously a limited experiment and is now rolling out to all users.
- Performance reports: clicks and impressions, with filtering and export.
- Insights: recent traffic and top posts.
- Achievements: milestone tracking over 28-day periods.
Related: 5 Best Social Media Channels for B2B Marketing (2026)
Why social reporting matters in B2B
B2B buyers increasingly find companies through founder videos, product walkthroughs on YouTube and short clips on social platforms, and those posts often surface in Google results for branded and problem-based searches. Until now, most teams had no clean way to see which social content Google was actually showing.
With platform properties, you can answer practical questions: which YouTube videos appear for your category terms, whether your founder's posts on X get any Search visibility, and which posts deserve to be turned into full articles on your own site.
- Add platform properties for every social account your company actively uses.
- Export the top posts by impressions and look for topics that also appear in your website's Search Console data.
- Where a social post ranks but your site does not, write a deeper page on that topic and link to it from the post.
- Review the report monthly alongside your website data, not in a separate silo.
Mueller's warning about markdown for AI
The second story concerns a tactic that has spread quickly: publishing markdown copies of web pages, or files such as llms.txt, specifically for AI agents. Mueller cautioned against building separate markdown versions of a site for AI, arguing that a well-designed website already works for AI agents, search engines, language models and, most importantly, people. He and Google's Martin Splitt also noted that a parallel version is one more thing to maintain, and that a broken page only machines see has no human visitor to notice the problem.
The recommendation is to put the effort into fixing the HTML you already have instead.
Related: 8 Proven Tips for Getting More Leads via Social Media (2026)
What "fix the HTML you have" means in practice
For most B2B websites, the pages that matter are product, pricing, integration and use-case pages. Making them readable for AI and search systems is the same work as making them clear for people:
- Render content in HTML, not only through client-side scripts or images of text.
- Use a logical heading structure so each section's topic is obvious.
- State facts plainly: what the product does, who it is for, what it costs and how it integrates.
- Keep one canonical version of each page rather than near-duplicates for different audiences.
- Mark up key entities with structured data where it genuinely describes the page.
What about llms.txt?
The llms.txt idea, a plain-text file that summarises a site for language models, is attractive because it looks cheap. The advice from Google's side is to weigh it against simply improving the pages you already have. Google has not presented these files as something that helps a site in Search, and every extra file is something else that can drift out of date.
A reasonable position for a B2B team is this: do not build parallel content for machines as an SEO tactic. If you already have a spare afternoon, spend it on the pages buyers actually read. Rewrite a vague product page so it states what the product does in the first paragraph, add a clear pricing explanation, or fix a page whose main content only appears after JavaScript runs. Those changes help every visitor, human or not, and they never need a second copy maintained.
Where the two stories meet
Both updates reward the same discipline: publish content once, in the right place, and measure it properly. Social posts are now measurable in Search Console, so they can be part of your search strategy rather than an afterthought. Your website should be the single, well-structured source of truth that those posts point back to, not a collection of machine-only mirrors.
That also simplifies maintenance. A small marketing team cannot keep a website, a markdown twin and an llms.txt file in sync while also posting on four social platforms. One clean site plus measured social distribution is a workload a small team can actually sustain.
Connect Search visibility to real visitors
Search Console tells you how often you appeared and how often people clicked. It cannot tell you who those people were. For B2B teams, the useful next step is knowing which companies and buyers arrived from Search, social or AI tools, and whether they match your ideal customer profile.
Happierleads' website visitor identification fills that gap by identifying the person and company behind a visit, so you can see whether a ranking post or page is attracting the right accounts. Pair it with engagement data such as average time on page to judge whether those visitors read what they found.
The takeaway
Add Search Console platform properties for your social accounts and fold that data into your content planning. Skip building markdown copies of your site for AI; spend the time making your real HTML pages clear, fast and well structured instead. One strong website, measured distribution and a way to see who actually arrives will serve search engines, AI systems and buyers at the same time.
