{"id":51058,"date":"2025-12-19T07:00:35","date_gmt":"2025-12-19T12:00:35","guid":{"rendered":"https:\/\/www.demandgenreport.com\/?p=51058"},"modified":"2025-12-10T14:01:27","modified_gmt":"2025-12-10T19:01:27","slug":"demand-gen-report-2025-reflections-branchs-adam-landis","status":"publish","type":"post","link":"https:\/\/www.demandgenreport.com\/demanding-views\/demand-gen-report-2025-reflections-branchs-adam-landis\/51058\/","title":{"rendered":"Demand Gen Report 2025 Reflections: Branch\u2019s Adam Landis"},"content":{"rendered":"

Marketers Were Blind to AI in 2025. In 2026, It\u2019s Back to Basics<\/strong><\/em><\/h1>\n

Marketing operates a predictable loop: measure performance, optimize based on data, repeat. But in 2025, we\u2019ve watched the fastest-growing technology adoption in human history unfold while our ability to measure it lagged years behind.<\/p>\n

The pace caught marketers flat-footed. 2026 is when they catch up.<\/p>\n

2025 Reflections<\/strong><\/h3>\n

Artificial intelligence (AI) growth is unprecedented.<\/strong> Mary Meeker, the referee on industry growth benchmarks for the last decade, published an authoritative 340-page report on AI trends<\/a> that makes all previous digital adoption curves look positively sluggish. Case in point: PCs took 20 years to reach 50% adoption in the US; AI is estimated to take three. OpenAI now predicts their models will run autonomous companies and simulate human minds within a decade.<\/p>\n

We lost control of the funnel.<\/strong> \u201cLLMs are not search engines, they\u2019re answer engines\u201d\u2014 a quip I first heard this year\u2014 perfectly captures an emerging problem for marketers. By design, large language models (LLMs) provide a contained experience that decreases the need to click out to other sources. As LLM usage grows, we\u2019re hemorrhaging organic search and sitewide traffic because discovery is happening entirely within the chatbot experience.<\/p>\n

A SEMrush study<\/a> found that an LLM referral is worth 4.4x more than someone coming from organic search. This makes sense; traffic that clicks out of an LLM is much higher intent. The same study estimates LLM traffic will overtake organic search by 2028, which means this visibility problem is only going to get worse. The reality is the top level of your funnel is happening inside ChatGPT, Claude, and Gemini. You just can\u2019t see it yet.<\/p>\n

LLM measurement is in the dark ages.<\/strong> We have almost no ability to measure the impact of LLMs. Right now, we\u2019re stuck literally asking the LLM itself where we rank. But getting back a simple list of where you rank against competitors feels a lot like the early days of search: interesting, but not very actionable. The SEO adage from the early 2000s, \u201cI know where I rank. I just don\u2019t know what to do about it,\u201d is back in a big way.<\/p>\n

Just like the days before Google Analytics, the most advanced marketers have resorted to combing through server logs<\/a> to find the LLM search-agent traffic, which still doesn\u2019t show up reliably in most analytics platforms.<\/p>\n

2026 Predictions<\/strong><\/h3>\n

OpenAI will start advertising \u2014 and create a land-grab moment.<\/strong> Analysts are already lambasting OpenAI for taking too long to implement an ad model. In 2026, OpenAI will finally move into advertising. Sam Altman expects this to be some type of affiliate fee, but what matters to marketers isn\u2019t the mechanics; it\u2019s what it represents: the birth of an entirely new paid media channel.<\/p>\n

All models will explore some type of monetized traffic or commerce model. Remember when Facebook’s e-commerce advertising spawned an entire DTC category and kicked off a performance marketing boom? LLM-sponsored placements will create the same land-grab opportunity for brands willing to experiment early. And critically, once LLMs start monetizing traffic, they’ll be forced to create methodologies to help brands measure the media efficacy.<\/p>\n

The brands testing placement strategies, bidding approaches, and learning how to measure conversions will have years of advantage over those who wait.<\/p>\n

All major analytics platforms will start measuring AI traffic. <\/strong>Yes, platforms technically track LLM click-out traffic today using generic UTM codes like \u201cchat.openai.com<\/a>,\u201d but most referrals still come through AI web crawlers<\/a> \u2014 and because those are classified as bots, analytics platforms miss the full picture.<\/p>\n

In 2026, expect LLM crawler identification, AI agent referral classification, visibility into event logging, and the emergence of new attribution models (conversation to conversion) \u2014 provided LLMs emit the necessary metadata. This will accelerate as they test and scale monetization and advertising.<\/p>\n

Measurement is step one. What to do about it comes next.<\/p>\n

Marketers will start to optimize for LLM traffic. <\/strong>Here\u2019s the good news: Emerging academic research<\/a> shows us what influences LLM recommendations, and it\u2019s surprisingly familiar territory. LLMs have human-like biases that respond to classic optimization techniques:<\/p>\n