Context Window

If you open ChatGPT, Claude, or Perplexity and type: "Explain Jobs to Be Done theory," you'll get Harvard professor Clayton Christensen's name mentioned in the first few sentences. Every time. Every model. Ask the same tools to explain the core idea without naming it: "Why do people buy products based on the progress they're trying to make rather than demographics?" The answer gets muddier, less attributable, and anchored to no one. The AI tool will construct something useful from general...

We've been talking about execution and attribution of the Source Signal Stack for several weeks now, so today, I want to zoom out a bit and take a look at the data around why this work is so important in 2026, and why LinkedIn is a prime platform for leveraging executive thought leadership, employee experts + B2B creators. What Today's Edition Is Below, you'll find a condensed synthesis of published research from Profound, SEMrush, SE Ranking, ConvertMate, Yext, and Search Engine Land....

In the last two editions of this newsletter (here and here), I made the case for two of the most underused Source Signal investments that help with LLM citation rate in B2B SaaS: Internal SMEs publishing under their own names, and long-term engagements with external creators bringing independent, third-party authority. You understand the theory now, but next, we have to talk about outcomes and attribution of these efforts. Because if you’re going to invest in this work, you need to be able to...

Last week, I made the case that external B2B creators on LinkedIn can anchor and supplement Layer 3 of the Source Signal Stack (the layer where AI search systems look for independent, named-human entities) for companies that don't have internal SMEs willing or able to publish under their own names. Today, I want to go deeper on the B2B creator piece. In this edition, I’ll be answering questions like: How to identify a valuable Layer 3 creator/influencer How to structure partnerships that...

Last week, I published a piece on the Source Signal Stack, a four-layer framework for thinking about which content signals LLMs actually trust when they decide who to cite in ChatGPT, Perplexity, Claude, and Google AI Mode. The TL;DR: the stack starts with brand-owned content at the bottom and ends with earned community mentions at the top, and the further a signal originates from brand control, the more weight LLMs give it. That piece argued Layer 3, with internal subject matter experts...

You've read the AEO playbooks, and you understand what needs to happen: Your employee subject matter experts (not just the CEO, but also the PMs, solutions engineers, and VPs of customer success) should be publishing under their own names, consistently, with a defensible point of view. The problem is that you're running a team already stretched past its limit. Your SMEs have jobs to do, and nobody has thirty free hours a month to figure out which experts to activate, what they should say, how...

Here’s a stat from AirOps that should make every B2B content team genuinely uncomfortable: 85% of AI citations come from third-party platforms, not brand-owned sites. Not your blog. Not your resource hub. Not the content you've spent years building, optimizing, and promoting through every owned channel you have. Third-party platforms. Sources you don't own and often can't control. Now pair it with this finding from Moz’s analysis of 40K queries: 88% of Google AI Mode citations aren't even in...

AI content is still newish on the scene, but it's already making B2B content all blend together. You know what I’m talking about: The long-form articles that all follow the same structure and tone, with definitions, tables, comparisons, and stat-heavy paragraphs all built for LLM extraction. It’s a formula…and sure, it might appease the LLM method of information parsing. But for the human readers (you know, the ones making the decisions and spending the money to buy your thing?) They’re...

Last week, I ran a quick experiment. I pulled the top 20 traffic-driving pages from a client's blog. Every single one ranked on page one for its target keyword, had solid click-through rates, and good engagement metrics. By every traditional SEO measure, these pages were performing well. Then I tested how often those same pages were cited in AI-generated answers across ChatGPT, Perplexity, and Google's AI Overviews. The results were not great. A handful showed up occasionally, but most didn't...