Original insights by Ross Hudgens
30-second rundown
Key learnings:
- What is happening: Across 116 B2B websites, dedicated product-versus-product pages had the strongest relationship with visits referred by AI tools.
- What to do: Prioritize honest pages that compare your product with the specific alternatives buyers already consider, then build a focused library around real sales questions.
- Why it matters: Comparison pages can generate a useful baseline of AI referral traffic, but the biggest gains still depend on whether AI systems recognize the brand and consider it relevant to the category.
This week: Pull five recurring comparison questions from sales notes or search data, check whether each has a dedicated comparison page, and choose the highest-value gap. Create a factual brief for that page with buyer type, use case, price or plan differences, strengths, limitations, evidence, and a clear recommendation so a writer can build it without inventing a knife fight.
Coming 3 months: Set an AI search monitoring workflow to test the same competitive buying questions weekly, record which brands and pages are recommended or cited, and alert one content owner when visibility changes. Review the trend monthly and automatically feed the clearest missing comparisons, weak evidence, and outdated claims into the editorial queue.
Ross Hudgens walked into 116 B2B analytics accounts and found a brawl already in progress. His study argues that when buyers ask an AI system to compare two products, the machine favors pages built in the same shape as the question: one product, one rival, and nowhere for the decision to hide.
The analysis covered 1,112 transactional comparison pages and 739,492 visits referred by ChatGPT, Perplexity, Claude, Gemini, Copilot, and Meta AI during a trailing 90-day period. ChatGPT supplied most of those visits, so the whole desert still tilts toward one very large casino.
1. The Machine Wants the Fight Named Clearly
Siege Media compared three page types: direct “A versus B” pages, alternatives pages, and “best software” roundups. A Spearman correlation measures whether websites that rank higher on one measure also tend to rank higher on another, and versus pages produced the strongest relationship with AI-referred visits at 0.65, about twice the next-best page type.
That does not prove the pages caused the traffic. It does suggest a brutally tidy match between the question and the evidence: when somebody asks which of two products wins, a page devoted to exactly those two products gives the machine a convenient witness instead of a crowded lineup outside a county jail.
2. The First Twenty Comparison Pages Change the Weather
The study grouped websites by the size of their comparison-content library. Median AI-referred visits were 350% higher for sites with six to 20 comparison pages than for sites with one to five, while sites with 21 or more comparison pages recorded 900% more median visits than the smallest group.
Hudgens’ practical conclusion is that the early library delivers the sharpest lift. You do not need to manufacture a hundred thin duels in a content factory run by nervous ferrets, but one lonely comparison page is unlikely to cover the real set of competitors buyers bring into the room.
3. Content Raises the Floor, Not the Roof
The most important restraint is buried under the exciting numbers: comparison content explained about 28% of traffic differences between the B2B sites. The remaining variation came from factors such as brand demand, category fit, and the wider network of pages that mention and link to a company.
Build the comparisons buyers genuinely need, make the evidence specific, and admit where the rival is stronger. The machine is not asking for a victory speech. It is looking for a usable record of the fight, and it has no patience for a referee who clearly favours one of the boxers.
