Ross Hudgens's Recent LinkedIn Posts

“5% of our closed-won deals come from LLMs.” Wrong. 5% of your currently attributable deals come from LLMs. The real number is far more likely to be in the 30–60% range. Why? You’re probably looking at referral traffic to infer attribution. Instead, you should allow users to self-attribute. Add a prompt to your lead form that asks: “How did you hear about us?” This will better capture sources from people who Googled you after using an LLM or typed your brand name directly. For Siege, 5% of our traffic comes from LLMs. The actual number attributable to LLMs via self-attribution? 50%! 10× the default assumption for many. There are hundreds of SEOs struggling to show the value of their work as clicks decline. This is one path to fixing that. If conversion rate is a concern, make the field optional or place it post-conversion. Neither is perfect, but both are far more helpful than relying solely on referral traffic to justify GEO work to executives.

"SEO for AI" is the slight naming preference of marketers we surveyed. In polling 353 marketers in our most recent trends report, we asked the group what of popular acronyms they preferenced. Turns out the overall favorite is, by a slight margin, a subcategorization of SEO itself rather than an all-new acronym. Of course, it's worth noting this sample size could be skewed towards GEO with only a few more responses. The margin is narrow between all of them, perhaps most accurately showing not what term will win, but that the battle over the preferred term is likely to continue for the months (and perhaps years) to come. It would not shock me if an influential publication, person or event skewed the story in one direction. Or of course, how Google's own AI approach evolves and if they are dominating in a year or two's time. If they do end up as the market dominant source, it would not shock me if we end up as "SEO for AI" after all.

Default LLM tracking is a major strategy mistake. The reason? It often incudes equal weighting across all LLMs. Imagine in 2020 if your CEO called a meeting because your Yahoo! rankings went down. That’s what we’re setting ourselves up for by using an equal weighting of every major LLM in our LLM visibility trackers. The only LLMs that matter are AI Mode and ChatGPT, yet most are tracking across several LLMs and reporting on the aggregated score. Even worse, this is presenting inaccurate prioritization to us since the source URLs and domains on these secondary tools are *not* where we should be spending our energy, it should be where the puck is and is also going in ChatGPT and Gemini/AI Mode. Add a few others if you really care about it, but know that would be better done via complete segmentation and unique reporting rather than a mixed view. We all understandably started in this “let’s see how it’s all doing” direction, but as the market evolves, it’s clear where it is. Ahrefs doesn’t normalize traffic metrics for Yahoo! and DuckDuckGo in its search data. There’s a reason for that, and there’s a reason we shouldn’t do the equivalent for our LLMs either.

Digital PR media citations do not drive direct LLM visibility increases. And we sell digital PR as a service! There is some misguided snake oil being passed around by people who have something to gain by telling you a random mention in news publications leads to direct visibility gains. Digital PR *can* and *does* have impact on LLM visibility, but it's not because the mentions get cited. It's because the rankings of product pages and comparison listing pages increases. And that overall brand awareness improves. If you dig into any set of bottom funnel prompts' source URLs from media, they are almost always the same: affiliate roundups. The best "X" articles taking advantage of that site's authority to rank. Not an article covering a news story. In fact, looking at a client's list just now, 100% of the media URLs (into the 100s for this site) are affiliate roundups. Graphs like this can be misguided because they drive site owners to believe a random mention drives visibility, measuring the success of their work incorrectly. It may also lead them to investing in things like syndication for high-volume "citation increases" in the media. You should not invest in syndication for that reason. You should invest for additional reach and actual brand awareness improvements from doing so. You can definitely get that from syndication, but it's important that you start by assessing its value from the right place.

LinkedIn post search traffic is up 400% in the last quarter! For the first time last week, I saw LinkedIn rank first in a B2B site's most sourced LLM domains. The trend is clear: LinkedIn is the next Reddit in the world of SEO. In September, I wrote about the first burst in traffic these posts were seeing on Google. Since then, the trend has accelerated. The likely reason? LinkedIn is the B2B version of what Google looked for in Reddit - a firehose of "human" content in a world where that's hard to come by. What does that mean for brands? Leaning search *slightly* in LinkedIn approach might be worth considering. The value of LinkedIn content has gone up from its already high position. The longer-feeling engagement curve to our LinkedIn content? That might be Google's doing, not LinkedIn's. Using an employee-first content creation strategy to scale may have outside effects in the months and years to come. I don't think low-engagement content will work here -- i.e. Google may be able to suss out and set a bar around a high volume of quality engagements, rather than low-value content generated by AI. Those are just initial impressions. Lots more to watch in this space!

ICP mapping is a critical, GEO-first tactic for B2B. What is ICP mapping? Creating landing pages that give you full coverage of the industries, solutions and business sizes you support. Within LLMs, long-tail searches and user customization will look for signals like this so clear signaling will help significantly. Where before you could "vibe" communicate this with logos, design signaling and case studies in many circumstances, it now pays to very clearly make it obvious who you support in the form of direct landing pages. These pages will in turn appear in query-fan-outs for LLMs, making sure you appear for the queries that matter. This is also an alternative, brand-first alternative to the "best X software" approach -- by having these long-tail pages mapped, you have a better shot at ranking for landing pages affiliate sites won't touch, or won't touch well. Finally, it's worth noting that this is also just good best practice for CRO independent of LLMs. The more content you have that speaks tightly to a given ICP, the better you'll convert within that audience. Before LLMs were a thing, industry pages were a highly recommended strategy for most SaaS companies. Now it's a critical one.

SEOs, welcome to the world of the long, long tail. What is the long long tail? The thinned out, multi-faceted variations on searches users, and in turn LLMs, perform to find the absolute best result for that prompt. As Clearscope founder Bernard Huang described on our most recent podcast, this has implications for practitioners to map their topic universe to the possible variables query-fan out could have for the bottom-funnel searches they make. I recently called this "product page / ICP mapping" in my recent LinkedIn post. Where previously we could stop at a term like "project management software", now it behooves us to have pages for: "project management software for small business" "project management software for retail" "project management software for IT" "organizational planning software" "project intake software" .. and all the additional permutations off that. We must now anticipate the fan-out queries and meet them where they are in order to best optimize for the future in front of us (AI Mode as default). The best way to think about this is to build pages that have minor long-tail search volume already, that also have a high probability of better speaking to your ICP. If we do that, we have multi-faceted value that builds defensibility against changes to come.

Is your content strategy actually "fresh," or just aging gracefully? We analyzed 100 SEO heavyweights across 10 industries to see how often the best of the best actually update their content. We measured the top 200 posts from every player to establish both the average freshness across the board, and also the "freshest five" -- the aggressive benchmarks set by the best of the best. We plan to update this quarterly to measure how competitiveness changes and freshness increases (or decreases) with improvements in technology.

Here’s how to successfully promote content on Reddit: Represent yourself authentically, as the brand. Attach a premium image that makes sense out of context of the post and would work as a standalone asset even if you never visited the website. Include a title that speaks in Reddit’s voice, while also intriguing the reader. Share content that’s unique and Redditors would love (unique data, news, behind-the-scenes takes), not a recycled keyword-focused post. Submit to niche, hyper-relevant subreddits only, and don’t overdo submission volume. Emphasize quality, not quantity. Overall, The Athletic does this well via their owned account. The submissions don’t always hit a home run, and that’s okay. They add enough authentic value that The Athletic never needs to be worried about being banned for their approach. That’s how you successfully scale a content distribution approach on the network.

Reddit's influence on LLMs may be overrated for many. Something I've pattern matched recently: Reddit almost always appears in the most cited domains, but it rarely appears in the most cited URLs. This hints at the fact that Reddit supports LLMs with breadth, but it does not provide it depth. Anecdotal research notes that Reddit's partnerships with these tools mean it may be providing it with an initial list of reference points to dig into, but it leans on live web search and deeper, more credible pages like the list on the right to determine its final ranking. This means a few things: 1) It seems that LLMs generally like referencing the opinion of the crowd (Reddit) to inform its outputs, but skips it as a consistent source of trust. 2) If you're a prominent enough brand that discovery is not a problem, Reddit is likely not adding much value for you. 3) The fact that it uses a long-tail of Reddit threads rather than a short-tail means that scaling this outcome effectively is problematic, especially when we consider that doing so requires gray-hat behavior for the most part. Of course, Reddit is seeing massive Google visibility, so the reasons for caring are larger than LLMs alone. This said, there are now a few datapoints out there that point to this channel being a bit overstated. A parable I'm considering: if Reddit is the equivalent to something like a Robots.txt file in SEO. You have to be aware of it and should occasionally make tweaks to the file (strategy), but for the most part it's set and forget. Currently, the crowd is overindexing because we *know* it's a citation factor. But just because it's a proven citation factor does not mean it's a large one, like many other SEO ranking factors before it. We should act accordingly.

Community is not a Slack channel. In-person events are not for every ICP. Qualitative datapoints must be considered and integrated into reporting if you are to build a truly defensible content strategy. These were just some of the takeaways I took away from my recent conversation with fractional CMO/ex-Animalz CEO Devin Bramhall. It was a needed mentality shift around what it takes to build a community. A community is not (despite my premonitions) limited to a Slack channel everyone's invited to. Community is instead the investment in relationship building with your ICP, wherever and wherever they live. That can sometimes be within someone *else's* community. Or breakfasts, for busy executives. Or happy hours, for young professionals. The medium changes, but the need for investment is all the same. Devin calls this "media-first content marketing strategy", the backbone of her new B2B book on the same subject. Highly suggest checking it and the conversation out!

Content marketing spend is growing in some areas – but bottoming out in others. That’s the insight from our latest trends report, where we surveyed 353 marketers in SEO and content marketing about their 2026 budgets, amongst several other questions that were top of mind for all of us this year. The bottom end of the budget arena is getting smaller, which may speak to those companies using AI instead, or otherwise abandoning content marketing as AI summaries proliferate and clicks consolidate within the enterprise. In the middle area there are increases, which is a good sign for companies (and their team members) taking content marketing seriously. At the top end, despite low sample size, the data shows an increase as well. This is the datapoint I question the statistical significance of most, as I do think it’s getting harder and harder for many enterprises to spend $50k/month plus on content, even if it’s addressably possible for their market. That’s due to executive AI pressure on line items like that, as well as the reality that there is less top-funnel traffic to go around than there was several years ago. This means the true content “TAM” is smaller for the average business, making million-dollar plus yearly spends a bit more rare to see. In the report, we also asked preferred AI content tools, marketers’ favorite AI search acronym, outsourcing trends and a whole lot more. This was our fourth year in a row creating this report. It's wild to see just how much change has happened in that time period, and it's safe to say we're not done yet.

Can’t prove the value of your GEO work? You probably need self-attribution. Self-attribution is a lead form step that allows buyers to input how they heard of you. Without this, it’s nearly impossible to adequately capture the value you are creating on Gemini, ChatGPT, Claude and other LLMs. Why? LLMs rarely link. When they don’t, consumers will end up Googling your brand and clicking an ad or your homepage, areas that historically have not been credited to search practitioners, or are otherwise blended with other channels. Self-attribution is better than a multi-choice selection because it’s more precise. Lazy buyers who don’t have an incentive to answer this well will often just pick something, while an open-ended solution allows for specificity and also, unique responses that will often lead you to new, high-converting sources worth looking into. If lowering conversion rates is a concern, make the field optional. This is especially critical in B2B right now because so much purchase behavior is happening on LLMs. If you don’t have this in place, things are likely looking bad for you due to the decline in clicks.

Content marketers' trust in Claude increased by 96% in 2025. We surveyed 353 SEOs and content marketers in our most recent trends report. One of things we were curious about: what LLMs content marketers trust most for content creation. In a multiple choice poll, Claude's reputation significantly improved year-over-year, nearly doubling from the year prior. The overall theme, however, is that trust has generally increased across the board. Gemini is up more than double as well, and Perplexity is too. ChatGPT had the shortest distance to improve, but even they went up. Using AI for content marketing is becoming more popular as these tools get better and better at what they do. That's the core change that we're seeing as an industry year-over-year.

More and more companies are outsourcing their content marketing. This is a good and bad thing for freelancers, agencies, and the market as a whole. The good: obviously, freelancers and agencies like us may be positioned to benefit from this shift. The bad: this, in my opinion, reflects the current anxiety around the state of content marketing in 2026. And for in-house teams, agency or freelancer gain is their (potential) loss as well. Fewer companies feel that overstaffed content departments will hold up long-term, and instead may decide to outsource that function rather than overhire as they wait for the dust to settle. It may also be a byproduct of a broader, connected malaise around hiring as AI accelerates around us. Overall, content marketing is clearly still in high demand, and it always will be. But we’re seeing a shift in how the market perceives it. It behooves us all to shift with those changes—and ideally to do so before the waves move themselves.

The right type of affiliate review can make or break a business on LLMs. We’ve entered a period equivalent to New York Times giving a positive or negative single review of a local restaurant pre-Yelp. That could close or grow that business. There are industries where there is one dominant URL source in the space. What if that same source gave you a negative review, instead of a good one? Your entire visibility would be thrown. An example of this is Bankrate in Fidelity’s market. Bankrate is everywhere across the prompts that matter to them. Incentives are aligned for Fidelity (and others like them) to have a good score. Because Bankrate gets paid when they get paid as an affiliate business. But what if that dynamic changed? These affiliate sites are getting hammered from a traffic standpoint, but they still carry clear influence in terms of the “votes” they offer these companies via LLMs. Don’t be surprised if the business models change with them. Behind-the-scenes dealings around the sentiment of the reviews offered. Affiliates charging hand over fist to be listed, rather than getting paid per lead.   Attribution will get muddy, but the influence they have is clear. If we can associate the value LLMs create for us through self-attribution, we can tie a dollar value to the right LLM sources. Prediction: Affiliate sites will find a second wind in the future as this all becomes clear, and the brains behind the monetization shift with it.

Want a quick way to (potentially) pay for gifts? Take our survey! We're looking for more responses to our content marketing trends report to fully capture the state of the market as we head into 2026. Two survey responders will be randomly selected for a $500 Amazon Gift card, and the survey only takes ~3 minutes to complete. If you've ever found something Siege has created helpful, a moment of your time would be greatly appreciated as a thank you. Fill out the survey here: https://lnkd.in/gkw633Ni Happy holidays!

"Commit to publishing one LinkedIn post a day." This is the wrong thinking, but it's close to the right idea. What if we instead changed the output goal to an input instead? "Commit to spending one hour on LinkedIn content per day." This changes the framing from needing to post once per day to instead spending one hour of energy towards high-quality content. This might mean only one post per week. Or even one post per month! But it means we've allocated energy consistently to the outcome, vs arbitrarily forcing ourselves towards an output metric that will inevitably lead to diluted content quality over time. It's easy to see how this same framing applies to content creation beyond LinkedIn, too. It's about showing up and spending consistent time towards making something great, not hitting arbitrary frequency goals. The returns get exponential when you cross the quality chasm to a place few others are willing to go in terms of time invested.

New data: Despite an increase in manipulation, trust in Reddit is up 19%! We surveyed 2,000 Americans to get an up to date sense for how users perceive different online sources in 2026, including Reddit. The original hypothesis? Reddit manipulation due to recent integration in Google may be driving a decrease in trust. I could not have been more wrong – trust is up YOY. What did I get wrong in that original assessment? Google’s impact is mostly on older results, and shopping recommendations are still a small overall percentage of the social network’s content. Also, Reddit is an offset to AI. There is much less reason to use AI-generated content there, and the respite from slop makes it a place that continues to get more and more popular. The channels actually seeing a decline in trust, especially with earlier generations? Search engines and traditional media. The move to originally surface Reddit was Google’s alone, and if the results they push ultimately end up manipulated as well, it’s possible this ties to some of that perception. Social media and AI is generally more trusted with younger audiences as well, with the biggest gaps in old and new being with social media (less trusted with old, well trusted with young) and Reddit (mildly trusted with young, not at all with old). The full study can be found in our top navigation!

We shouldn’t over-optimize for personalization in GEO. The reason? Gemini, unlike OpenAI, requires a multi-step user opt-in (that hasn't yet worked for me) to retain information about the user. This is because Google takes a buttoned up approach and is understandably sensitive to possible user blowback from collecting personal information at scale. This means the default prompt response will be generalized for most. If we think that Gemini is going to win, which they are now on pace to do, the idea of synthetic personas may be overstated and unnecessary for the majority. However, user behavior is still likely to change over time. Short keywords will be swapped out for longer prompts, especially for users who realize the results will not automatically be customized to them. This means that there will still be a need for longer-tail landing pages than there was previously. But there may not be as massive of a long-tail as an infinite web of personalized outcomes may have previously projected, if the winners and the losers stay on pace with recent changes.

What blogs generate the most organic links? That's what we set out to discover in an analysis of organic link generation by industry. Certain industries can generate these at scale, and others require manual outreach via digital PR (or other methods) to make a dent in their link footprint. Where companies go wrong strategically is completely missing this as a forest of opportunity available to them. In certain spaces, tying content (and data) to the right search terms can unlock a long-term waterfall of link generation. The initial promotion nudge is still helpful, but it's aim of developing compounding snowball that often generates organic links into the hundreds and sometimes thousands per month for many of these websites. Looking across the winners, we see data-rich real estate websites like Zillow, and then complex but important technology spaces like SOC2 compliance with Vanta and Secureframe. Evaluating them deeper, it's almost certain that these sites *also* do outreach. For example, I'd bet Zillow's press team shares that information with media every time it's published. But that initial outreach is still in aim of igniting the snowball -- and it's often the biggest link earners organically that may also benefit the most from manual outreach, given the overall value of the content they're creating. Full report can be found in the top nav of our website!

BrightonSEO's early bird rates expire Thursday at 12pm PT! Brighton is a must-attend SEO/GEO conference in San Diego that I recommend to all marketers across the United States. I'm honored to be speaking and teaching a course again this year (this time on Generative Engine Optimization). The event is one of my favorites, with great networking, speakers, and of course, San Diego weather to cap it off. Looking forward to meeting many old friends and meeting several new ones, too. See you in September?

LLM visibility has the potential to improve faster than SEO ever could. We had a client increase visibility by 330% in just 150 days! Why? The ability to impact or replace source URLs almost immediately. If you previously lacked a transactional content strategy but had strong authority, sprinting in that direction can drive immediate gains in visibility. Similarly, if you didn’t have an affiliate motion or awareness of the most frequently cited affiliate and partnership source URLs, you could sprint there as well, going from visibility in only a few source URLs to nearly all of them almost overnight. That creates a before-and-after visibility shift unlike the SEO strategies of years past. This doesn’t mean this level of improvement is possible for everyone, or even most brands. It’s most achievable for brands with specific gaps and existing authority, growing from a low visibility baseline to a strong one. Still, it’s encouraging that we now have the ability to move the needle quickly for some companies. Where “wait six months” was once common in SEO, being able to demonstrate real traction in three to four months is a compelling reason for more executive teams to commit and stay invested for the long term.

LinkedIn article SEO traffic is up 296% QoQ! Here's how to create LinkedIn posts that rank in order to take advantage: Put the keyword in the first sentence of the post. LinkedIn uses the first sentence as the title tag and URL slug, critical SEO elements. To add to that, Google extracts the first ~72 characters into their "What people are saying" area, making it a pivotal CTR element if you want search visibility. You should also add an image that's compelling, driving additional intrigue. Anecdotally, it does seem that showing the keyword on the image may also contribute to surfacing more often, although it might be a correlative rather than causative effect. There are some examples of pictureless posts surfacing, but they are rare and understandably have few reasons to surface when compared to more engaging, multi-modal content created on the platform. LinkedIn Pulse (blog post type formats) and LinkedIn posts both show up in this carousel. However, typical posts like this are the format seeing explosive search growth/showing up more often outside of "What people are saying", while Pulse is actually declining in search performance. Finally, freshness is critical. Most of these feature areas include content less than a month old. Evergreen LinkedIn content can pop into the organic results, but it's these feature blocks that are most reliably going to drive performance for you. Tying creation to trending topics can increase ROI probability for this reason. LinkedIn is on the up and up, and I only expect that to accelerate. Give it a year or less, and we'll be talking about the network's B2B search visibility being equatable to Reddit's, at least in proportional impact.

Something I haven’t seen LLM trackers solve for yet: prompt difficulty. It’s an amorphous concept because difficulty is relative: for certain top-funnel terms, the difficulty of getting mentioned is effectively infinite, or otherwise stated, impossible. But on bottom-funnel (BOFU) terms, brands *will* get mentioned. So the relative difficulty of being shown relative to someone else is still a real thing. For practitioners thinking about GEO, this is important to understand because modeling visibility improvements is difficult when the baseline is hard to understand. Imagine trying to show up for “best soda brands” compared to “best SaaS tools for California private equity firms”. The first would be wildly difficult to penetrate, the second, potentially not as much once a product was up and running. How I would start thinking about this: Determine the relative timeline to appear frequently amongst the top source URLs. For e-commerce brands, I often see Amazon reviews driven by most popular placements. These sometimes have thousands of reviews. Breaking in here may be very difficult for a new CPG startup. In B2B, I’ll see BOFU roundups created by first parties ranking. If we’re relevant and authoritative, it may be possible to appear here within three months. For travel brands, I see first party landing pages ranking, likely driven through Google search results. This is a brand authority/digital PR play, that again may take years to unfurl. It’s feasible a trendy, Reddit-loved startup could disrupt rankings quickly due to its ability to appear amongst Reddit results: again, only if Reddit is amongst the top source URLs for the prompts that matter. Building on this, we’d need to look at the relative topical authority of the site and likely also consider traditional search metrics like Domain Rating, given the overlap is still so strong. It may be more useful for these trackers to actually account for brand-by-brand difficulty to rank using topical assessments rather than generalized ranking. With the power of AI, this is more feasible today and definitely something we will add to our internal datapoints for client prioritization in the future. Of course, this is just the top of the iceberg here – what would you add?