BrightonSEO talk roundup October 2025

This October, some of the Varn team headed to Brighton for the world’s biggest search marketing conference, brightonSEO.

Alongside supporting two of our own speakers, we attended a wide range of insightful sessions across the two days. We’ve rounded up summaries of our 10 favourite talks in this article, so you can catch all the key takeaways from the best of the event.

Becky Simms explores search journeys through the lens of behavioural science. She details the emotional and cognitive reasons why users behave differently on different platforms, and how it can help your search strategy.

Key Takeaways: 

• Different platforms encourage different types of search behaviour. A white paper posted on Instagram does not align with the user’s most natural behavioural preference. Position your content to optimise for the platform’s preferred search behaviour. 
• Put thought into who tells your story. While your brand can tell its own story, users on platforms like Reddit may prioritise user-generated content, or content by trusted voices. There are ways to create opportunities for user-generated content through outreach, which is important for credibility and brand image.

In today’s platform-dominated digital ecosystem, the purpose of a website as a one-stop hub for information, engagement and conversions is rapidly disappearing. Rachael Murdoch details the shifting purpose of websites in an age where users are increasingly making purchases and resolving queries off-platform. 

Key Takeaways:  

• Users are turning to more personalised platforms like AI and social media to purchase goods and services, and to engage with content.
• Websites are now being used as a data source, making it an important anchor in a fragmented ecosystem. We need to be comfortable with less traffic and re-set website goals to accommodate this change.
• Adapting your website to these changes will require some experimentation. Optimising for humans: test scenario-based content and test formats that support long-tail conversational search. Optimising for bots: reduce JavaScript dependency, optimise your structured data and check log files to monitor bot behaviour. 

With the rise of zero-click search, publishers and content-centric websites increasingly struggle to drive traffic and monetise. Jack Chambers-Ward shares practical advice to optimise your content and business strategy to drive revenue and connect with your audience.

Key Takeaways: 

• With the rise of search in sites like YouTube, Reddit, and TikTok, content needs to be diversified to meet the needs of users. The rise of these search platforms also brings the opportunity to build your brand outside of traditional search.
• People are interested in people. Once your audience is interested in your story, create premium content that users are willing to pay for.

Inyene Udom highlighted the invisible gaps on your website that can stop traffic from converting, even when SEO is strong.

Key takeaways:

• Tools don’t reveal everything — a visible site isn’t always a converting site.

• Above-the-fold clarity is critical — show what you do, who you help, and why visitors should stay within 3 seconds.

• CTAs aren’t just buttons — they guide users toward conversion.

• Footers are a final handshake — a well-designed footer can secure conversions.

Anthony Tattum explored how behavioural science shapes decision making, showing that website conversions are driven more by emotion than logic.

Key takeaways:

• Understanding Capability, Opportunity and Motivation (COM-B) unlocks buyer behaviour.

• Use the EAST framework. make actions Easy, Timely, Social and Attractive.

• Personalisation builds ownership and reduces abandonment.

• CRO is psychology in action. Focus on human behaviour, not just design tweaks.

Jérôme Salomon’s session kicked off the first day at brightonSEO, setting the bar incredibly high and delivering one of the most insightful talks of the event. He looked into how server logs can reveal your presence in AI-driven search and focused on practical ways to use this data, like filtering key status codes or identifying missed opportunities where content is crawled but never cited or clicked.

 Key Takeaways:  

• We need to differentiate between training bots (GPTBot, ClaudeBot, CCBot) and AI search bots (oai-searchbot, perplexitybot) as their purposes are independent and so are the insights we can gain with this separate segmentation.
• Connecting log data with SEO performance will give us a sharper view on opportunities, like spotting pages that are frequently crawled but never clicked, or high-performing SEO pages that AI bots haven’t visited at all.
• AI bot server hits could pass as AI visibility (impressions), but it is important to note that not all impressions come from citations.

Another fascinating presentation was Adam Gent’s, on how Google’s decides why some pages might never make it into the index. It was one of those sessions that really helped connect dots between crawling and indexing. What caught my attention was his point that “Crawled – currently not indexed” (GSC) is not passive at all. It was a great reminder that quality and engagement is what keep pages in the index.

Key Takeaways:  

• Google is actively removing pages from the index, based on technical errors, duplication and quality. 
• Page quality, measured through content, engagement, and user satisfaction, is the most common reason for deindexing.
• “Crawled – currently not indexed” often means Google has removed a page from the index due to quality, not neglect.

Ulduz Ismayilova shared a highly practical intent analysis, demonstrating her process for classifying search intent at scale without coding, using only scraping tools and AI, to better tailor keyword research. A particularly interesting point was that this approach also serves as a powerful system for analysing competitor performance, boosting both content and SERP strategies.

Key Takeaways:  

• Intent analysis can provide actionable insights like identifying competitors’ performance, who are the dominant domains in the SERPs, and whether content is B2B or B2C to inform content optimisations.

An insightful talk from Tom Rawle looking at the drawbacks of using AI in your outreach and press release content with views from those in the press on what they would think if they received a press release written by AI. 

Key Takeaways:  

• Double down on audience first digital PR & focus on relevancy and quality
• 96% of AI citations originate from PR content (earned and paid) as opposed to brand-owned content.
• You need to be 100% confident in the data behind the stories you are pitching.

An insightful talk from Mary Hickey of Silverbean Looking at a study into a 822 real-world press release headline test, revealing the importance of testing and making sure your headlines and subject lines meet the mark. 

Key takeaways: 

• Subject lines with “Revealed” in had very low engagement with journalists 
• Citing experts within your subject lines is a great way to increase open rates 
• If you have a data-campaign, expect to get very low traction on the locations and cities that rank low within the study 
• Try and get TikTok and celebrities in your subject lines where you can. 

A great talk from Gareth Simpson of Seeker looking at how you can leverage the knowledge graph to increase the authority of your brand and appearances within LLMs and the wider knowledge base of the web. 

Key takeaways: 

• Informational retrieval is moving to a question answer concept with many users now being more interactive with the web in the way they answer questions 
• A lot of what we are seeing now with increased AI adoption could be an indication of a move towards a ‘gentle singularity’ according to Open AI’s Sam Altman. 
• Make sure you are positively contributing to the web to future-proof your appearances in AI and search.

PR has gotten progressively more competitive but Luke Cope’s talk dives into tactics you can employ to generate coverage and links for category pages with creativity and reactive. 

Key Takeaways: 

• Predictable reactive: What events happen every year (e.g an airline going bust) that you can have stock content and ‘reactive’ content for as a brand. 
• Cause and trend analysis: If certain events are coming up or happening, what could this cause and how can you position your client to benefit? E.g the surge in popularity of Barbies post Barbie movie launch (data shows what the most popular barbies are following movie launch). 
• Take your idea a step further: The example Luke gave was of a campaign looking at which areas of the UK are losing the most public toilets, stage 2 of that idea was to start a petition to get it debated in parliament, what could your stage 2 be? 

A great talk from Katie Thompson explored different trends within reactive marketing and how brands can benefit. Katie shared how to take advantage of trends in a world where many people are pulling back from the news. She covered the differences between newsjacking, adding brand commentary to stories gaining press attention, and trendjacking, which focuses more on social media trends to drive client coverage.

Key Takeaways:

• Short-term fame doesn’t always lead to long-term success for brands or individuals.

• Before reacting, fact-check that the story is genuine news.

• Never be afraid to respond effectively to bad press (for example, KFC’s FCK campaign following the chicken shortage).

• Make use of existing communities, such as answering the most common Reddit questions on a topic with expert insight.

An insightful talk from Roman Leliukh looking at why ChatGPT may be promoting your competitors whilst ignoring you. 

Key takeaways:

• LLMs scan early more heavily, focus more on what’s included above the fold, so make sure you summarise content early on. 
• When looking at how you are appearing within prompts, look at the citations and number of facts served within those prompts. 

In this talk, Gijs de Groot shared data-backed insights into what actually drives visibility in AI-powered search engines. Using real statistics from prompt analytics, website traffic and server logs, Gijs took us on a journey of prompt discovery. 

Key takeaways:

• Reddit has dropped massively in citation popularity amongst ChatGPT

• Crawler analytics and log file analysis can show you what AI LLMs are crawling your site and when. 
• Connecting the journey from prompt feature to visitor behaviour is a great way to add to your LLM analytics visibility.