Part 4 in our BrightonSEO 2026 series. Catch up with the hub post, Part 2 on why your traffic is falling, and Part 3 on getting your brand mentioned by AI before reading this one.
Continuing our series of posts following on from BrightonSEO, the world’s biggest SEO conference, this week we are taking a close look at content and some of the key messages that came up throughout the conference.
And there were a lot of them.
On a personal level, these were some of the most interesting sessions as my knowledge and expertise in SEO definitely leans more into the content side of things. Technical SEO is an important part of what we do here, but I was intrigued to learn more about how others were tackling content creation in an ever-increasing environment of AI slop, and the talks did not let me down.
Content was arguably the most richly debated topic across the two days – probably because it’s the one that sits at the intersection of everything else. The traffic decline we wrote about in Part 2 is partly a content problem. The AI visibility strategies from Part 3 are built, at their core, on content. And the measurement and reporting questions we’ll cover in later posts in this series are ultimately questions about whether the content we’re producing is delivering business outcomes.
The consistent thread that ran through every content-focused session was the same: the way most brands approach content, built around keyword lists, chasing volume, producing at scale, is the wrong model. Not because it never worked, but because it produces exactly the kind of average, interchangeable output that AI now aggregates and makes invisible.
Here’s what the leading practitioners in our industry are doing instead.
The Uncomfortable Truth About Most Content Strategies
Let me start with a statistic that Paul brought back from Day 2, because it’s the kind of number that makes you look at your own client reports with fresh eyes.
Daniel Cartland, Delivery Director at Novos shared the analytics from a client whose blog was generating 50% of their total site sessions – a significant portion of their overall traffic. But when he looked at revenue attribution, that same blog was responsible for just 7% of their revenue.
Half the traffic. Seven per cent of the business value.
The blog wasn’t underperforming because it was poorly written or technically flawed. It was underperforming because it had been built around keyword volume and traffic targets rather than around the audience – who they were, what they actually needed, where they were in their decision-making journey, and what would move them towards a purchase or an enquiry.
This is, if we’re being completely honest, a pattern I recognise across a lot of content strategies I’ve reviewed in my time as an SEO. Truthfully, it’s one we’ve chased ourselves. It’s understandable how it happens: keyword research tools show you volume, so you chase volume. You produce content that ranks for terms with high search counts. The traffic comes in. The dashboard looks healthy. And then you look at the revenue data and realise most of those visitors had no meaningful connection to what the business actually sells. Our historic approach, to be fair, has always been slightly different, and where possible, we would split the traffic streams up so clients could see where the traffic was coming from and what was driving sales and revenue. This is even more important today.
Daniel Cartland’s session title was From Personas to Purchase: Turning Customer Insight into SEO Impact – and his central argument was that the gap between traffic and revenue almost always comes down to persona depth. Content built around a real, detailed understanding of who you’re writing for will almost always outperform content built around a keyword spreadsheet.
Paul’s take:“Daniel’s session landed with me because it wasn’t theoretical. He had the data from a real client account showing the exact commercial disconnect between traffic-led content and revenue. I came away from that session with a very specific action: we need to build proper persona frameworks for every Digital Hothouse client and use them as the lens for every content decision. Not as a box-ticking exercise, but as the actual foundation of strategy.”
Thankfully, we were already ahead of this to some extent and have been working with client customer personas for at least 12 months and in some cases, even longer. We just needed to do more to build these out based on the research presented and the way that content is heading.
Sophie Coley and the Audience Facets Framework
If Daniel Cartland’s session was the diagnosis, Sophie Coley’s (Head of AI & Strategy, Propellernet) was the prescription. Her talk, focused on what she calls audience facets, is one of the most practically useful content frameworks I’ve come across in years, and it’s directly applicable to businesses of any size.
The standard approach to audience research produces broad personas: “women aged 35–55 interested in wellness” or “SME owners in the professional services sector.” These are useful as starting points, but Sophie’s argument is that they’re far too blunt to actually inform content decisions. They describe a demographic, not a decision.
Audience facets go deeper. They’re built around four specific dimensions:
Situations — the specific contexts in which your audience finds themselves when they need what you offer. Not just “looking for a new accountant” but “recently hired as a Finance Manager at a company that outgrew their previous accountant and is under pressure from the CEO to get control of the numbers.” The more specific the situation, the more precisely you can address it.
Triggers — the events that move someone from passive awareness to active search. A trigger might be an anniversary, a complaint from a customer, a staff change, a regulatory deadline, a news story, or a conversation with a peer. Understanding triggers helps you create content that connects at the moment of highest intent.
Labels — the exact language your audience uses when they describe their own problem. Not the industry terminology you use internally, and not the keyword the search tool suggests, but the actual words a real person types into a search box or says to an AI assistant when they are experiencing the problem you solve. If your content uses different language from your audience, the gap in vocabulary becomes a gap in visibility.
Context — the background, beliefs, and prior experiences that shape how your audience interprets information. Someone who has been burned by a previous agency will read your “why choose us” page very differently from someone considering professional help for the first time.
What I love about this framework is that it completely sidesteps the volume-versus-quality debate that tends to dominate content discussions. If you build content around real audience facets, you are almost by definition producing content that a specific person genuinely needs at a specific moment, which is exactly what converts, and increasingly, exactly what AI surfaces.
Chima Mmeje: Stop Chasing Keywords. Start Chasing Demand.
Chima Mmeje’s (Moz) session on demand-led content strategy was one of those talks that changes the sequence of steps you follow when you’re building out a content brief.
Her core argument: keyword research tools are a lagging indicator. They show you what people have been searching for in the past. They don’t show you what your audience actually wants to talk about, what they’re worried about, what they’re excited about, or what questions they’re asking each other that they’d never type into a search engine.
Real demand, Chima argued, lives somewhere else – and it’s more accessible than most brands realise:
LinkedIn conversations. Industry discussions, comment threads on thought leadership posts, questions asked in professional communities. These are real practitioners expressing real needs in their own language, often before those needs have crystallised into search queries.
Customer feedback and support logs. Every complaint, every question to your customer service team, every review that begins with “I wish it did…” is a window into an unmet need that your content could address.
Reddit and community forums. As we covered in Part 3, Reddit is one of the most heavily-weighted sources for AI training data – but it’s also one of the richest repositories of authentic consumer language you’ll find anywhere on the internet. A thread about your category will almost always surface questions, objections, and terminology that a keyword tool would never surface.
Industry events and conferences. What are the questions being asked from the floor? What are the topics generating the most energy and debate? (This series of blog posts, incidentally, is a direct product of exactly this kind of demand-led thinking – we came back from Brighton with a clear picture of what our audience needed to understand, and we’re working through it systematically.)
The practical implication of Chima’s framework is that content ideation starts before you open a keyword tool, not after. You find the genuine demand first, then you figure out how to reach the people who have it. The sequence matters.
Daniel Liddle: The Race to Sameness
Daniel Liddle, Senior SEO Director at iCrossing’s session on AI and content strategy had a slightly different angle from Sophie’s and Chima’s, but it arrived at the same place.
His warning was aimed specifically at the surge of AI-generated content we’re all seeing: “AI aggregates the average. In an emerging channel, average is invisible.”
This is worth unpacking. Large language models are trained on patterns. When you ask an AI to write content about, say, the benefits of business coaching, it will produce something that reflects the patterns in all the content it has ever consumed on that topic. The result is, by mathematical definition, the average of everything that has been written about business coaching. It is not unique. It does not offer a perspective no one else has offered. It does not contain an experience, a data point, or an insight that can’t be replicated by any other brand in your space running the same prompt.
And if your content strategy is built primarily on AI-generated text following templates derived from what your competitors are already doing, you are producing content that looks like everything else, sounds like everything else, and offers nothing that everything else doesn’t already offer. In a traditional search world, this is a problem. In an AI-mediated world, where platforms are deciding which sources to cite based partly on distinctiveness and genuine authority, it’s a serious strategic error.
Daniel’s prescription wasn’t “don’t use AI for content” – it was “use AI for the parts of content production where it genuinely helps (research, formatting, drafting) and protect the parts that only you can provide: the genuine expertise, the first-hand experience, the opinions formed from actually doing the work.”
This aligns directly with what Google has been trying to tell us through E-E-A-T for years. The extra E – Experience – was added precisely to differentiate content produced by people who have actually done something from content produced by people (or tools) who have merely read about it. That signal matters to Google. It matters more than ever to AI systems.
The Post-Purchase Opportunity Most Brands Are Completely Ignoring
One of the sessions Paul flagged as most commercially underappreciated came from Olabinjo Adeniran, Product and Growth Marketing Consultant, whose talk was titled SEO for Retention: The One Thing You’re Missing.
His central point: most SEO content strategies are focused entirely on acquisition – reaching people who don’t know you yet and persuading them to become customers. But the content needs of existing customers are at least as important, often more commercially valuable, and almost always underserved.
The case study he used was The Body Shop. Their retention content strategy included a glossary of ingredients (so customers who had purchased a product could understand what was in it and why), FAQs on every product page (addressing the questions people ask after purchase, not just before), and active encouragement of third-party TikTok content – meaning real customers showing what products look like in real use, in real conditions, by real people.
The outcome: reduced post-purchase churn, increased repeat purchase rates, and a content library that AI platforms actively cite because it answers specific, detailed, real-world questions that the brand’s competitors haven’t bothered to address.
The commercial case is difficult to argue with: a 5% increase in customer retention can lift revenue by anywhere from 25% to 95%. That range varies by business model and sector, but the underlying principle is consistent across virtually every business we work with. Acquiring a new customer costs significantly more than keeping an existing one. Content that supports the customer after they’ve bought – that answers the questions they have post-purchase, that builds confidence in their decision, that reduces the likelihood of buyer’s remorse or support calls – pays back in ways that are genuinely measurable.
For any business that sells something with a learning curve, a maintenance requirement, or a repeat purchase cycle, this is a content opportunity worth exploring seriously.
Paul’s take:“Olabinjo’s talk was one of those sessions where I sat there thinking ‘we should be doing this for half our clients right now.’ The 90-day retention content roadmap he outlined – starting with support data mining to identify the most common post-purchase questions, through to content creation and distribution – is a ready-to-implement framework. We’ll be building this into client strategies where there’s a clear retention opportunity.”
Dr Pete Meyers: Keywords Aren’t Dead – They Just Need a Rethink
I want to close this section with a slightly contrarian note from Dr Pete Meyers from Moz, whose session on The Infinite Tail: Keyword Research for AI was one of both our highlights from Day 2 and was the closing session for the entire conference.
The overall thrust of BrightonSEO’s content sessions was firmly in the “stop chasing keywords” camp, and I think that message is right. But Pete’s point was subtler and worth hearing: keyword research isn’t obsolete. It’s the relationship between keyword research and content strategy that needs to change.
In an AI-driven environment, the long tail of search is getting longer, not shorter. People are increasingly asking AI assistants very specific, contextual, multi-part questions rather than typing short keyword fragments into a search box. The queries are more conversational, more situational, and more nuanced. They sound less like “best running shoes” and more like “what running shoes would work for a nurse who stands for 12-hour shifts, has wide feet, and wants something that looks decent off-duty too.”
This is actually the Sophie Coley audience facets framework expressed as a search query. The more specific and situational the content you produce, the more naturally it maps to the long-tail, conversational queries that AI systems are fielding. Keyword research done well – looking for the specific language, the specific situations, the specific questions – feeds directly into the audience-first approach every other speaker was advocating.
The mistake to avoid is using keyword research tools as a substitute for audience understanding. Used as a supplement – a way of validating that the demand you’ve identified through community research and customer feedback also has a measurable search signal – keyword research remains genuinely valuable.
What This Means in Practice: A Simple Framework
Pulling this together into something actionable, here’s how we’re thinking about content strategy for our clients in the wake of BrightonSEO:
Start with the audience, not the keyword. Before any content brief, the question to answer is: who specifically is this for, what situation are they in, what triggered their need, and what language do they use to describe it?
Find demand where it actually lives. LinkedIn threads, Reddit discussions, customer reviews, support tickets, and conversations at industry events are richer sources of content insight than a keyword tool’s “content ideas” tab.
Build content that only you can produce. First-hand experience, proprietary data, expert opinions, case studies, original research – the elements that can’t be replicated by a competitor running the same prompt. This is the content AI cites. This is the content that converts.
Don’t stop at acquisition. The post-purchase content opportunity is real, underexplored, and commercially significant. Map the questions your customers have after they buy, not just before.
Measure what matters. Traffic from content that doesn’t convert is a cost, not an asset. The metric that matters is revenue (or leads, or enquiries) that can be attributed to content — not the volume of sessions content generates.
The Audit We Run Before Any Content Work
For any client where we’re building or reviewing a content strategy, the first step is always to understand where content currently sits in the commercial picture. That means:
Pulling the traffic data by content type (blog/information vs commercial landing pages vs product pages) and comparing it against goal completions and revenue attribution. If traffic and revenue don’t track together, the content mix needs examining.
Reading the last 50 customer reviews and support conversations to identify the questions that appear most frequently, both pre and post-purchase.
Spending 30 minutes in the relevant Reddit communities, LinkedIn groups, and industry forums to understand what language people actually use, what they’re worried about, and what they’re excited about.
Looking at what AI platforms say when asked about your category, because the questions that AI answers on your behalf are often the same ones your content should be addressing on your website.
If any of this sounds like something worth doing for your business, get in touch with us – it’s one of the most useful conversations we have with clients at the start of a content engagement, and it almost always surfaces opportunities that weren’t visible from inside the business.
Coming Up Next
In the next post in this series, we’re shifting to the technical side, specifically the technical SEO foundations that matter most in an AI-driven world. From JavaScript rendering to schema markup to log file analysis, there are some important things your website might be getting wrong that have nothing to do with your content and everything to do with whether AI can actually read what you’ve produced.
Gavin is Digital Hothouse’s SEO lead. He attended BrightonSEO 2026 with Director Paul Thornton.
