Jun 5 2026
AI Literacy Is Not Prompt Literacy. Ann Handley Says It’s Judgment Literacy via @sejournal, @gregjarboe

Ann Handley posted something on LinkedIn last week that stopped me mid-scroll. She’s a Wall Street Journal bestselling author and one of the most respected voices in marketing, and she wrote:

“AI literacy is not prompt literacy. It’s judgment literacy.“

Her post went on to ask a question that nobody in the AI training industry seems to be asking: “Why do we keep teaching people how to use AI – without ever teaching them when not to?”

I messaged her. I had to know where someone would go to learn that.

Her honest answer: “I don’t know of a course that teaches exclusively this. At MarketingProfs, our sessions about AI typically include a few slides that touch on when not to use AI, or how to protect against hallucinations, but I don’t know of a whole session or series.”

She added, “I think that’s actually the story, and why I wrote what I wrote. We have an entire industry built around AI skills training – prompt engineering bootcamps, certification programs, tools tutorials, a million LinkedIn posts about the perfect prompts you need to do this or that or else you’re falling behind. What we don’t have is anything that asks: when should you put the tool down? When does using it cost you something you didn’t mean to give up?”

That gap is real, and it matters more than the AI training industry currently acknowledges.

Prompt Literacy Takes An Afternoon. Judgment Literacy Takes Years

The distinction Ann draws is not subtle once you see it. Prompt literacy is teachable in an afternoon. You learn the syntax, the structure, the iterative refinement loop. You learn to be specific, to add constraints, to tell the model what not to do as well as what to do. This is genuinely useful and genuinely learnable quickly.

Judgment literacy is something else entirely. It is knowing when the speed of AI output is actually eroding something you needed to build slowly. It is recognizing when the struggle itself is the point, when the friction of not knowing the answer yet is what produces the expertise that will matter later. It is understanding, as Ann put it, “when AI helps and when it shortcuts the very struggle that teaches us something.”

One commenter on her post put it precisely:

“Prompt literacy is teachable in an afternoon and judgment literacy takes years, because judgment is mostly knowing the value of the struggle you’d be skipping.”

I’ve been teaching an online course on AI content that audiences actually trust for several years. And I’ve spent recent months analyzing what the AI training landscape actually offers practitioners. The pattern is consistent. The courses that exist (and there are now many of them) teach you what tools can do. The better ones teach you how to deploy them strategically. Almost none of them teach you when to put them down.

This is not a minor gap in the curriculum. It is the central question of the current moment.

Why The Gap Exists

The AI training industry has a structural incentive problem. Courses that teach you to use tools generate demand for more tools, more courses, more certifications. There is no business model for teaching restraint. Nobody is building a prompt engineering bootcamp whose primary lesson is “sometimes don’t.”

But the cost of skipping the judgment question is real and measurable. Anthropic’s own research found that junior engineers who leaned heavily on AI coding agents demonstrated weaker understanding of their work when tested afterward. When the tool produced output, their struggle that would have built expertise did not happen. The output and the expertise are not the same thing.

For SEO professionals and content marketers specifically, the exposure is direct. MIT’s AI Labor Exposure Map, which I wrote about last week, found that nearly three-quarters of the time a marketing specialist spends at work goes to tasks that AI can already handle. The question is not whether to use AI for those tasks. For many of them, you should. The question is which tasks in that 74% are actually the ones where the doing is the learning, where outsourcing the execution also outsources the understanding you needed to build.

That question requires judgment. It cannot be answered by a prompt.

Culture, Not Coursework

When I asked Ann where practitioners should go to develop this judgment, her second message reframed the question entirely.

“Do we actually need a course? What we need instead is permission and better modeling. Leaders who visibly choose the long road. Managers who say out loud when they are not going to use AI for certain things, and here’s why. Individuals who see the value. Said another way: culture not coursework.”

That reframe is worth sitting with. The judgment about when not to use AI is not a skill that gets transmitted through a certificate program. It is a professional norm that gets transmitted through observation, through watching someone you respect make a deliberate choice to do something the slow, human-fumbling-in-the-dark way, and then explaining why.

Ann has a book coming out in February 2027 from Penguin Random House called “ASAP (As Slow As Possible): When to Take the Long Road in a Shortcut World.” The title captures the tension precisely. In a professional culture that has made speed the primary virtue, choosing slowness requires not just judgment but courage: the willingness to be seen taking longer when everyone around you is accelerating.

What Practitioners Can Actually Try Right Now

Ann’s point about culture rather than coursework is correct in the long run. But while that culture is still forming, practitioners need something concrete. Here is a workflow worth replicating, drawn from an experiment I ran with the editorial team at The Acton Exchange, a nonprofit community newspaper in Acton, Massachusetts, in November 2025.

The team faced a deadline problem. A steering committee had just held a three-hour working session on a critical school district reorganization question, reviewing 156 pages of materials. The meeting wasn’t recorded, which meant no transcript was available. But the 101 pages of supplemental information and 55 pages of public comments the committee had received ahead of time were accessible.

So, the team tried something new. We crafted a detailed prompt specifying what the article needed to accomplish: accurate and trustworthy information, a compelling story, relevant to residents. We uploaded all 156 pages to four AI engines simultaneously: ChatGPT, Gemini, Perplexity, and NotebookLM. Each engine took a different route from the same prompt and the same source material. ChatGPT produced 748 words focused on data and process. Gemini produced 712 words focused on why the status quo was no longer viable. Perplexity produced 1,232 words focused on what the options meant for residents. NotebookLM produced 1,506 words organized around five surprising truths.

We reviewed all four drafts together at an all-hands editorial meeting. Perplexity’s draft was the most accurate and the most useful as a foundation. We chose it as our starting point. Then we did what no AI engine could do: We added direct quotes from people who were in the room, reflecting the community voices that the Acton Exchange exists to represent.

The key lesson from this experiment is not which engine performed best. It is what the process revealed about judgment. Town Manager John Mangiaratti had observed a few weeks earlier that the tools were helpful for the first 75% of content, but that “the remaining 25% of details, nuance, and context are either missing or incorrect.” Superintendent Peter Light agreed, adding that quality improves with better input prompts.

That 75/25 split is a practical frame for any content workflow. Use AI to get 75% of the way there quickly. Then apply human expertise, primary source verification, and direct observation to close the gap. The 25% that requires a human is not a bug in the workflow. It is where the judgment lives.

Before adopting any AI tool in your content process, have an explicit conversation with your editor or team about which tasks the AI will handle and which require human oversight. Document your prompt. Run the same prompt through more than one engine when the stakes are high. Verify outputs against primary sources before publishing. And disclose your process to your audience, as the Acton Exchange did at the foot of this published article.

Ann Handley is right that the real skill is judgment: knowing when speed is useful and when it actually erodes something you needed to build. The Acton Exchange experiment didn’t resolve that question. It made the question visible in a way that a prompt engineering course never would.

Prompt literacy gets you to 75%. Judgment literacy is what closes the rest.

More Resources

Featured Image: Yuriy2012/Shutterstock

Ecommerce MGMT 0 Comments
Jun 4 2026
You Can Finally Measure Content Alignment. That’s The Dangerous Part via @sejournal, @DuaneForrester

We have always been approximating relevance. Every keyword list, every TF-IDF score, every editorial judgment about whether a page “covers the topic” has been an attempt to answer a single question: is this content about the thing the user is looking for? The tools changed. The question did not. What changed, meaningfully, is the resolution of the instrument. Keyword research approximated relevance through lexical overlap: If the words match, the topics probably align. Vector-based semantic analysis approximates it through meaning overlap: If the concepts are close in embedding space, the content is probably relevant regardless of whether the exact terms appear. That is a genuine, material upgrade, but it is not a move from guessing to knowing.

The reason that distinction matters is that a significant portion of the SEO and content strategy community is right now treating it as if it were. They are looking at alignment scores, cosine similarity outputs, and semantic proximity metrics and reading them as ground truth. A high score means aligned. A low score means not aligned. Optimize until the number goes up. And the number, because it is a number, feels like it has settled the question that keyword research always left open. It hasn’t. It has given you a higher-resolution version of the same approximation, and the higher resolution is exactly what makes it dangerous, because it removes the humility that low resolution used to enforce.

Precision Is Not Accuracy

Gerard Salton’s SMART system at Cornell introduced the vector space model for document retrieval in the 1960s. The core insight then was the same insight powering today’s embedding models: represent both the query and the document as vectors, measure the angle between them, and use that angle as a proxy for relevance. What has changed across 60 years is the sophistication of how those vectors are constructed. Salton used term frequency. Modern embedding models use transformer-derived representations that encode semantic relationships, contextual meaning, and conceptual proximity across hundreds or thousands of dimensions. The measurement got dramatically better. But the thing being measured, the angular distance between two vector representations, is still a proxy for a relationship that exists outside the math.

This is where the Netflix research team landed in their 2024 study on cosine similarity in embedding models. Steck, Ekanadham, and Kallus demonstrated that cosine similarity applied to learned embeddings can produce results that are, in their framing, arbitrary. The way an embedding model is trained, the regularization applied, the data it saw, all shape the geometry of the space in ways that make a raw cosine score unreliable as an absolute measure of semantic similarity. A high score in one embedding space is not equivalent to a high score in another. The score is real. The similarity it claims to represent may not be.

For practitioners optimizing content, the implication is direct. When you score your content’s alignment to a query using an embedding model, you are measuring semantic proximity inside that specific model’s representation of language. You are not measuring how Google’s retrieval infrastructure or OpenAI’s RAG pipeline or Perplexity’s index would evaluate the same relationship. Those systems use their own embedding models, their own retrieval architectures, and their own reranking layers. A score of 0.92 in your measurement space might correspond to strong retrieval in one system, weak retrieval in another, and irrelevance in a third.

What Kind Of Wrong Are You?

This is the axis that matters, and it is not the one most practitioners are thinking about. The question is not whether keyword research or vector alignment is the better method. The question is what kind of error each method produces, because the error type determines whether you can correct for it.

Keyword research, for all its limitations, produces a known unknown. You know you are approximating. You know that matching terms to a page does not guarantee topical coverage, does not guarantee user satisfaction, and does not guarantee that a search engine will judge the page as relevant. The imprecision is visible, and because it is visible, it keeps you honest. Practitioners who grew up in keyword-driven optimization learned to over-cover, to build supporting content, to triangulate intent from multiple angles, precisely because they understood the instrument was blunt. The bluntness was a feature. It forced humility.

Vector alignment scoring, by contrast, can produce an unknown unknown. The number is precise. It has decimal places. It can be tracked over time, graphed, compared across content assets, and optimized against. And that precision creates a psychological trap: it feels like the question has been answered. The content is 0.89 aligned to the query. That must mean something definitive. But what it actually means is that in one specific embedding space, using one specific model’s learned representation, the angular distance between two vectors falls within a certain range. The score says nothing about whether the production retrieval system that will actually serve your content uses a compatible embedding space, applies the same tokenization, or weights semantic similarity the same way during reranking.

The MTEB benchmark leaderboard illustrates this concretely. The performance spread across current embedding models is not small. A content asset that scores well against one model’s embedding space may score materially differently against another, not because the content changed but because the geometry of the space changed. And the embedding model your scoring tool uses is almost certainly not the one any given AI platform uses in production. There is no public registry of which model powers which system’s retrieval layer. You are measuring in a space that is representative of the general problem but not identical to the specific system where your content will be evaluated.

That is not an argument against measuring. It is an argument against reading the measurement as settled fact. The distinction between a directional signal and a definitive answer is the entire discipline.

The Instrument Got Better. The Old One Is Not Enough

None of this rescues keyword-only optimization as a sufficient strategy. It is not sufficient, and the reasons are structural, not sentimental.

LLMs and AI retrieval systems operate in semantic space, not lexical space. They process meaning, not strings. A page can score perfectly against a keyword target list while being semantically adrift from the actual intent the query represents, because keyword presence and semantic coverage are different things. Conversely, a page can use none of the target keywords and still be strongly aligned semantically, because it covers the same conceptual territory through different vocabulary. The paraphrase and synonym space that LLMs operate in is structurally invisible to a keyword-based evaluation. You cannot see what you cannot measure, and keyword tools cannot measure semantic proximity.

Consider a practical case. Keyword research correctly identifies “customer churn prevention strategies” as a high-value target. The content team builds a thorough, intent-appropriate piece around it. It covers the topic, uses the target terms naturally, and would pass any keyword audit without issue. But an alignment score reveals that the content’s semantic center of gravity sits closer to “measuring churn” than to “preventing churn,” because the piece leans heavy on diagnostic framing, identifying at-risk accounts, calculating churn rates, segmenting by behavior, and lighter on intervention framing, what to actually do once you have identified the problem. Both treatments are on-topic. Both satisfy the keyword target. But the semantic distance between the content and the query as a retrieval system represents it is larger than the keyword coverage suggests, and keyword research has no instrument to surface that drift. The alignment score does. Not because the keyword research failed, but because it was never built to see at that resolution.

This is not a criticism of people who focus on keyword research. Those practitioners are not wrong. They are working at the resolution the available instruments allow. Intuiting alignment between content and query intent is a real skill, and the best keyword strategists are doing something genuinely sophisticated: they are approximating semantic relevance through lexical indicators, using editorial judgment to bridge the gap the tools could not cross. The tools can now cross a version of that gap. The editorial judgment still matters, but the gap it has to bridge is different.

The danger is the practitioner who decides that because keyword research is no longer sufficient, vector alignment scoring is the complete replacement. That practitioner has traded one approximation for a better one while losing the awareness that it is still an approximation. They have upgraded the instrument and downgraded the literacy, which is a net loss.

The Discipline Is Knowing What The Number Is Not Telling You

Goodhart’s Law, the observation that when a measure becomes a target, it ceases to be a good measure, is not just an aphorism for economists. It is the exact failure waiting for any team that treats an alignment score as a target to optimize against rather than a signal to interpret. The moment the score becomes the goal, the content starts drifting toward the score’s geometry and away from the actual relevance it was supposed to approximate. You start writing for the embedding model instead of the reader and the retrieval system, and the embedding model you are writing for is not the one any production system uses.

The real discipline, the one that did not exist when practitioners were navigating by keyword intuition alone, is understanding what an alignment measurement is and is not telling you. It is telling you that in a given embedding space, your content’s vector representation is geometrically close to a query’s vector representation. That is useful. That is more information than keyword presence gives you. It is telling you something about semantic coverage that lexical analysis cannot. But it is not telling you whether the production system’s embedding space has the same geometry. It is not telling you how reranking will treat the result. It is not telling you whether the LLM’s generation layer will interpret your content as authoritative, complete, or worth citing. Alignment is a retrieval-adjacent signal. It says nothing about interpretation.

The practitioner who can hold those two realities, the signal is real and the signal is incomplete, is the one operating with genuine literacy about the systems they are trying to influence. The one who collapses them, who reads a high alignment score as confirmation that the content is “optimized,” is operating with a more sophisticated version of the same overconfidence that made people think a keyword density of 3% meant their page was relevant. The number got better. The mistake is the same.

Representative, Not Identical

The honest framing is not “right space versus wrong space.” That binary invites paralysis: If no measurement space is the production space, why measure at all? The best framing, in my opinion, is a spectrum of representativeness. Some measurement spaces are closer to what production systems use than others. Some embedding models share more architectural DNA with the models powering major AI platforms than others. Some scoring methodologies account for the gap between measurement and production better than others. The question is not whether your measurement is perfect. It never will be. The question is how representative your measurement space is of the systems you actually care about, and whether you are treating the score with appropriate directional respect rather than absolute faith.

This is the actual work. Not chasing a number. Not abandoning measurement because it is imperfect. Building enough literacy about how these systems work to know which signals to take seriously, which to discount, and which to combine with other indicators before making a content decision. That literacy was optional when the only instrument was keyword research, because the instrument was so obviously blunt that nobody mistook it for truth. It is not optional now. The instruments are precise enough to fool you, and the cost of being fooled is optimizing content for a geometry that does not represent the system where your brand needs to be visible.

I wrote about a related dimension of this problem in the vector index hygiene piece last year, focusing on how the quality and maintenance of the index itself shape retrieval outcomes. This article is the other side of that coin: not the index, but the measurement you use to evaluate whether your content belongs in it. And both connect to a larger question I will return to in future work, which is a gap most people aren’t talking about yet.

Start With What You Can See

If you are still running keyword research as your primary content alignment method, you are working with a blunt instrument in an environment that now demands more resolution. If you are running vector alignment scoring and reading the output as settled truth, you have the resolution but not the literacy to use it safely. Both are correctable. The path forward is not choosing one over the other. It is layering them, understanding what each can and cannot tell you, and building the organizational capacity to treat precise measurements as what they are: directional signals produced inside a specific space that may or may not represent the systems where your content competes.

The gut feeling was never the enemy. The illusion that you have moved past the need for judgment is.

For a broader look at how AI search visibility is reshaping the work of being found, “The Machine Layer” covers the structural shifts that make this kind of measurement literacy essential.

More Resources:

This post was originally published on Duane Forrester Decodes.

Featured Image: Luke Jade/Shutterstock; Paulo Bobita/Search Engine Journal

Ecommerce MGMT 0 Comments
Jun 2 2026
Why Great Content No Longer Works: MIT Research Shows The Shift Reshaping SEO Strategy via @sejournal, @gregjarboe

“Many of the truths we cling to depend greatly on our own point of view.” said Obi-Wan Kenobi. It came back to me this week when I read a LinkedIn post from Rand Fishkin, which opened with a sentence I’ve never seen him write before: “I almost never write blog posts anymore, but this one felt necessary.”

Screenshot from LinkedIn, May 2026

I’ve been reading Rand’s blog posts for more than 20 years, and when he says something feels necessary, it’s worth stopping for.

The TL;DR for his article is this:

“Ignore traffic. Make inimitable products. Shift your priorities away from ‘great content’ on your own site and toward ‘great marketing’ on the platforms where your audience pays attention. Influence is the new traffic.”

What Rand Is Actually Saying

For 25 years, Google told websites to make great content, and they’d sort out the rest. Rand’s argument is that this was always incomplete advice, but it kinda, sorta worked – until now. Google’s future, as he sees it, is no longer indexing the web and making information universally available. It is what he calls “the great digital enclosure of publishing”: extracting content to fuel AI answers, reducing the need for users to ever click through to the original source.

The result is a zero-click web where content becomes a commodity and creators lose direct user engagement. His response is two-pronged.

Rand’s first solution is collective action. For SEO professionals and content creators, the question is whether the collective action path is realistic given their market position – and for most individual practitioners or small agencies, the honest answer is that it isn’t. Which makes Rand’s second solution the more immediately actionable one.

Solution two is what the piece is really about: building inimitable products. Things AI cannot replicate, Google cannot summarize away, and no algorithm can disintermediate. His examples are evocative. Ultrasonic chef’s knives. Made-to-measure suits with oceanic personality. WWI-era Armagnac sourced to serve someone’s 98-year-old grandfather something older than him. The point is that physical craft, genuine curation, deep expertise, and irreplaceable human judgment cannot be scraped and served in an AI Overview.

For digital practitioners who don’t make knives or suits, the harder question is what the inimitable version of their work actually looks like. Rand’s nearly universal advice: “Build an audience on a platform you don’t own. Publish there. Engage there. Use it to drive interest in your inimitable product.”

What The MIT Map Confirms

If Rand’s post tells you where the pressure is coming from, a new tool from MIT’s Work Analytics Lab/MIT CTL tells you how much pressure you’re personally under.

The AI Labor Exposure Map, reported by Hiawatha Bray in The Boston Globe this week, is a point-and-click resource that breaks down specific workplace tasks and shows which of them AI can already perform. It draws on methodology from MIT’s Work Analytics Lab/MIT CTL and data from Anthropic’s own AI Economic Index, measuring penetration scores for the share of each task currently capable of being automated or significantly assisted by AI.

The finding for marketing specialists is direct: 65% of the time a marketing specialist spends at work goes to tasks that today’s AI systems can handle. Market research, competitor analysis, campaign planning, data interpretation. Separate Anthropic research ranks marketing specialists fifth among the occupations most exposed to AI, ahead of customer service representatives and data entry workers.

MIT’s Pierre Bouquet, the doctoral candidate who developed the map, is careful to note it wasn’t designed as a doomsday prediction. AI capable of performing tasks and AI that will actually replace workers are not the same thing. But for SEO professionals, content marketers, and digital strategists reading Rand’s argument alongside the MIT data, the combination is clarifying: The content tasks that have defined these roles are precisely the ones most exposed to automation. And Google’s AI features are the delivery mechanism for that exposure.

Two Hard Choices, One Honest Assessment

Rand’s solutions map onto two genuinely different strategic paths, and they are not equally available to everyone.

The collective action path requires scale, coordination, and willingness to absorb short-term traffic loss in exchange for long-term leverage. It is more realistic for large publishers with established audiences than for individual practitioners or small agencies who cannot afford to gate their content and wait. The sites that tried withholding content from AI crawlers discovered quickly that the traffic cost arrived immediately while the negotiating leverage did not.

The inimitable product path is available to more people, but it requires a different kind of honesty about what you actually do. If 74% of your current tasks can be handled by AI, the question isn’t whether to use AI – it’s what the remaining 26% is, and whether you can build something valuable enough around it that people will pay for it regardless of what Google does to the click economy. That 26% is where Rand’s advice is pointing. Original research. Direct access to sources and communities. Judgment formed through years of pattern recognition that AI has not yet replicated.

Major brands are already reorganizing around this reality. Large agencies are facing account reviews. This will not be a quick or easy transition for anyone.

Advice For An Epic Journey

If you are about to navigate this transition, three things are worth carrying.

The first is a clear map of your own exposure. Know specifically which of your tasks are exposed before deciding which ones to protect, automate, or eliminate. You cannot navigate from a position you haven’t honestly assessed.

The second is Rand’s distinction between tasks and identity. The tasks that are being automated are not the same as the expertise that made you good at them. The SEO professional who understands why content earns trust is not the same as the workflow that produced that content at scale. The former survives.

The third is the oldest advice for any long journey: Travel with people who are honest about the terrain. Rand Fishkin is being honest about the terrain. So is the MIT map. The practitioners who read these sources carefully, test their conclusions against their own data, and update their strategies accordingly are the ones who will still be doing meaningful work when the transition is further along.

The point of view you cling to right now depends greatly on which data you’re willing to look at.

More Resources:

Featured Image: Roman Samborskyi/Shutterstock

Ecommerce MGMT 0 Comments
May 27 2026
Google’s Standards Haven’t Changed But AI Is Making That Harder To Ignore via @sejournal, @gregjarboe

Recently, Sam Sifton, who hosts The Morning newsletter for The New York Times, published a letter to his readers with an unusual subject line, “Who’s Writing This?”

His prompt was a new book called “The Future of Truth,” written by Steven Rosenbaum with significant AI assistance. The Times reviewed the book and found more than half a dozen misattributed or entirely fabricated quotes conjured by the AI, including one attributed to tech journalist Kara Swisher. Swisher’s response said not only was the quote wrong, but “I also sound like I have a stick up my butt.”

Rosenbaum’s defense that the hallucinations “serve as a warning about the risks of AI-assisted research and verification” is the kind of sentence that would be more convincing if it appeared in a different book.

Sifton used the moment to tell his readers something he clearly felt they deserved to hear directly. The Morning is built by humans, for humans. His team may use AI to find information that gets verified elsewhere. They may use it for editorial logistics, buying time for more reporting, but the thought-making, the question-asking, the deep reading, and the writing that follows – those are tasks performed by journalists free of chips. “I write fueled by adrenaline and fear of errors,” he told his readers. “And I promise you that will never change.”

What Google’s Guidance Actually Says

In February 2023, Danny Sullivan and Chris Nelson published Google’s guidance on AI-generated content. The position, which has not meaningfully changed since and was reinforced again recently in Matt Southern’s reporting on Google’s new AI search guide, is this: Google’s ranking systems aim to reward original, high-quality content that demonstrates E-E-A-T (expertise, experience, authoritativeness, and trustworthiness). The focus is on the quality of content, not how it is produced.

That sounds, on a quick reading, like a green light for AI content. It isn’t, or at least it isn’t a green light without conditions that matter enormously.

Google’s guidance specifically says that using automation to generate content with the primary purpose of manipulating search rankings violates its spam policies. And it draws an analogy that SEO professionals should analyze and evaluate: about a decade before the 2023 guidance was written, there were understandable concerns about content farms, which mass-produced large volumes of human-generated content. No one thought it reasonable to ban all human-generated content. Instead, Google improved its systems to reward quality. The helpful content system, the E-E-A-T framework, the information gain patent, the ongoing Quality Rater Guidelines updates through 2025 – all of it is the same enforcement mechanism, applied again, at greater sophistication.

Rosenbaum’s book is exactly the kind of content that Google’s systems are designed to identify and discount. Not because it used AI, but because it used AI carelessly, without the verification, the original reporting, and the editorial accountability that Google’s quality signals are trained to detect.

Sifton’s newsletter is exactly the kind of content those same systems are designed to reward. Not because it is human-generated, but because it is produced by people with genuine expertise, direct experience, and accountability to a specific audience. It is built by humans, for humans, in precisely the sense Google’s helpful content guidance has always intended.

Will Sifton’s Letter Change Anything?

The question at the center of this commentary is whether Sifton’s look at AI’s expanding role will change what Google is doing, change how practitioners write for AI, or change how they win in AI visibility.

The honest answer is no, not directly, and that’s the point.

Google’s guidance has been consistent since February 2023. It was consistent before that in spirit, through Panda in 2011, through E-A-T, through the Helpful Content Update in 2022, through the transition to E-E-A-T later that year. What changes is only the acuity with which people spot it on the horizon.

What Sifton’s letter does, that Google’s technical documentation cannot, is make the human cost of the alternative legible. Rosenbaum’s Kara Swisher hallucination is not an edge case or a technical failure. It is what happens when the thought-making is outsourced entirely, when the question-asking stops, when no one is writing fueled by adrenaline and fear of errors. It is a book about the future of truth that cannot be trusted.

For SEO professionals, the practical implication has not changed since Amit Singhal’s 23 Panda questions in 2011. Does the article provide original content or information, original reporting, original research, or original analysis? Does it have the kind of quality you’d expect to see referenced by a magazine, encyclopedia, or book? Would you be comfortable giving this to your editor and putting your name on it?

Sifton’s promise to his readers is that he would. That accountability is not a stylistic choice. It is the entire mechanism by which trust is built with an audience, and by which Google’s systems learn to surface content worth surfacing.

The Real Lesson

AI is not indifferent. It is responsive, adaptive, and improving faster than any previous technology transition in the industry’s history. That’s exactly what makes it useful and exactly what makes the question of how you use it so consequential.

But the standards that determine whether content earns trust, from readers and from Google’s ranking systems alike, do not move on AI’s schedule. They have been moving in the same direction for as long as Google has existed. Every approach that has assumed those standards would yield to scale, to automation, and to the next optimization trick has found the same thing.

They don’t yield. They move right along as though nothing happened.

More Resources:

Featured Image: Roman Samborskyi/Shutterstock

Ecommerce MGMT 0 Comments
May 22 2026
3 Unrelated Stories About AI & Writing Tell The Same Story via @sejournal, @gregjarboe

I stumbled upon three separate articles about writing and AI in the same week, each from a completely different angle, and all describing the same thing.

A novelist turned MIT writing lecturer confronting students who outsourced their essays to AI. A new Graphite study showing AI-generated articles now make up roughly half of all new content on the web and have plateaued there. And fresh data from The Accountancy Partnership showing that half of freelance creatives say rising stress is affecting their work, as client budgets for human creative services shrink.

One data point is a fact. Two is a coincidence. Three is a trend.

When read together, these articles formed an argument that every SEO professional, content marketer, and creative freelancer should take seriously, acknowledging the content divide that is happening and asking, “Which side are you on?”

The First Story: What Happens When Students Outsource The Struggle

On May 10, Micah Nathan, a novelist and MIT lecturer in fiction and non-fiction writing, published a piece in The Guardian about confronting his creative writing students over their AI use. The confession session that followed, he wrote, became one of the most productive teaching moments of his eight years at MIT.

His key insight wasn’t about academic honesty. It was about what writing actually does. “Writing isn’t just the production of sentences,” he told his students. “It’s the training of endurance by way of sustained attention. It’s a way of learning what one thinks by attempting to say it. An LLM can reproduce the appearance of that activity, but it can’t replace it, because the value lies not only in the object produced but in the transformation that occurs during its making.”

He described AI prose as “faultily faultless, icily regular, splendidly null,” borrowing Tennyson’s description of a beautiful but empty face, producing what he called “simulacra of thought, generated via pattern recognition learned from millions of human-penned words, rooted in no particular experience by no particular person.”

Insightful readers, he argued, feel that emptiness even if they can’t articulate it.

For SEO professionals, this is not an abstract literary concern. It is a precise description of the content quality problem that Google’s helpful content systems have been trying to solve since 2022. The signal Google is hunting for is exactly what Nathan identifies as the thing AI cannot produce – evidence of a mind actively grappling with a specific problem from a specific experience. Pattern recognition learns from what humans wrote. It cannot replicate why they wrote it. 

→ Read More: Why Great Content Is No Longer Enough & What Beats It In AI Search

The Second Story: The Feared Takeover Hasn’t Happened – Yet

On May 15, Megan Morrone reported for Axios on new data from digital marketing agency Graphite, which analyzed 55,400 online articles and listicles published between January 2020 and March 2026, running each through three AI-detection tools. The finding was more nuanced than most AI content coverage has been about the share of primarily AI-generated content, which has held near 50% for more than a year and appears to have plateaued.

The feared takeover hasn’t materialized. AI content briefly surpassed human-authored content in late 2024, but the two have stayed roughly equal since.

The important caveat Morrone included is that many articles are no longer written purely by humans or AI. A human may use AI for outlining, drafting, rewriting, or editing, making the line genuinely blurry. Dan Klein, a UC Berkeley professor and AI model CTO, flagged the feedback loop risk. Once models train heavily on AI-generated content, the internet could become a machine that produces low-quality content that trains models that produce more low-quality content.

For SEO professionals, the plateau is reassuring and cautionary in equal measures. The volume panic was overstated. But the quality dilution problem is real and growing, and it creates the same opportunity Nathan identified from the other direction. In a web that is roughly half AI-generated content, content that carries genuine human experience and specific expertise becomes more differentiating, not less.

→ Read More: AI Platform Founder Explains Why We Need To Focus On Human Behavior, Not LLMs

The Third Story: The People Producing This Content Are Under Serious Stress

On May 13, Emma Hull at The Accountancy Partnership directly emailed me data from a new report on creative freelancers across PR, marketing, performing arts, graphic design, photography, and adjacent industries. Half of freelance creatives (50.7%) say rising stress levels are affecting their work. Half (50.2%) say client budget cuts are the biggest challenge they faced in 2025. Over two in five (43.3%) believe AI will negatively affect their sector. Nearly half regularly work unpaid hours each week.

Lee Murphy, Managing Director at The Accountancy Partnership, put it plainly: “Creative work is often closely linked to marketing budgets and discretionary spending. When businesses begin tightening costs, creative services can sometimes be one of the first areas to see reduced investment.”

The irony embedded in these three numbers together is worth reflecting on. Clients are cutting budgets for human creative work at the same time AI is generating roughly half the content on the web, while a professor at MIT is documenting the specific cognitive cost that outsourcing the writing process extracts from anyone who does it, whether a student or a professional.

The freelancers under the most pressure are the ones most tempted to use AI to produce more content faster to compensate for lower rates. The content they produce that way becomes part of the 50% that is indistinguishable from machine output. And content that is indistinguishable from machine output is exactly what the Graphite data and Google’s quality systems are training users and algorithms to discount.

→ Read More: Relying Too Much On AI Is Backfiring For Businesses

What The Pattern Actually Means

The three stories, read together, describe a market in the process of bifurcating. On one side sits high-volume, low-differentiation content produced quickly, priced cheaply, and increasingly hard to distinguish from AI output, regardless of who generated it. On the other sits content that carries specific expertise, direct experience, and the editorial judgment that Nathan’s students were trying to skip past. Content that takes longer, costs more, and is increasingly the only kind that earns meaningful search visibility and reader trust.

This is not a new argument in SEO. What is new is the empirical clarity with which three independent sources from three entirely different disciplines – literary education, web content analysis, and freelance labor economics – are all pointing at the same conclusion in the same week.

Shelley Walsh made the point in her recent Search Engine Journal piece on scaling AI content that the commodity versus non-commodity divide is where the real strategic question lives. The three stories above are evidence that the divide is already here, already measurable, and already affecting people’s livelihoods.

The writers who understand this, and produce accordingly, are the ones who will still have work worth doing when the budget cycles turn again.

More Resources:

Featured Image: SvetaZi/Shutterstock

Ecommerce MGMT 0 Comments
May 20 2026
Can A 300,000-Influencer Network Built On AI-Generated Content Work? via @sejournal, @gregjarboe

When Unilever CEO Fernando Fernández stood before investors and declared that the era of expensive corporate brand advertising was over, calling traditional TV-heavy campaigns “lazy marketing,” the shockwave through the agency world was immediate. Half of Unilever’s massive global advertising budget would shift to a “social-first” strategy. Creator collaborations would scale by 20 times. The target would be an army of over 300,000 influencers, including a micro-influencer in every postal code in key markets like India.

Traditional advertising agencies that had spent decades building relationships around six-figure production budgets and a handful of celebrity partnerships suddenly faced a client with an operationally impossible mandate. Manual sourcing, onboarding, and content approval at 300,000-creator scale simply does not exist as a human workflow. Specialized creator agencies picked up business that legacy agency-of-record relationships had assumed were locked in.

The panic was understandable. It was also aimed at the wrong target.

The More Important Question

A March 2026 Adobe Express study surveyed video creators across YouTube, TikTok, and Instagram and found that 71% have now adopted AI video generation or editing tools. Of those, 41% deploy them on a weekly basis. 56% of creators using AI tools report saving over 30 minutes per video on average, with 10% shaving more than four hours off their production time. On the performance side, they’re seeing a 19% average increase in audience watch time and a 17% boost in community engagement. Half plan to increase their AI tool spending over the next year.

So, Unilever is building an army of 300,000 creators, and 71% of creators are now using AI to produce their content. The math is straightforward, and what Unilever is actually building is a massive distributed network for the production and distribution of AI-assisted content at a scale the marketing industry has never seen.

The question that hasn’t been answered yet is whether any of it will work.

→ Read More: The State Of AI In Marketing: 6 Key Findings From Marketing Leaders

Will It Work?

Unilever’s 300,000-creator network is generating content at a scale that makes traditional test-and-learn frameworks difficult to apply cleanly. When hyper-local micro-influencers are producing AI-assisted videos for niche audiences across hundreds of markets simultaneously, the signal-to-noise problem becomes acute. Individual pieces of content may perform well in isolation while the overall brand narrative diffuses into incoherence. Or the personalization may be exactly what audiences want, and the aggregate effect may be stronger than anything a single high-production campaign could achieve. Right now, the honest answer is that nobody knows with confidence.

Where DAIVID And ADIN.AI Come In

On April 27, 2026, two companies that many SEO professionals and digital marketers haven’t heard of yet announced a partnership that addresses the exact problem Unilever’s strategy creates.

DAIVID is a creative intelligence platform whose AI models, trained on tens of millions of human responses to ads, predict in seconds how any piece of ad creative will perform – measuring attention, 39 distinct emotions, memory encoding, brand recall, and likely next-step actions – without requiring human panels. ADIN.AI is an AI-native operating system for enterprise marketing that sits above an organization’s existing tools and provides a unified intelligence layer across channels, budgets, and decisions.

The partnership embeds DAIVID’s creative effectiveness models directly into ADIN.AI’s platform, creating what they describe as a live loop between creative intelligence and media execution. Before a campaign launches, marketers can identify which creative is most likely to succeed and allocate budget accordingly. While campaigns run, they can scale high-performing assets and pause underperformers in real time. After campaigns end, the historical performance data becomes benchmarks that guide future creative and media planning.

Ian Forrester, CEO of DAIVID, described the core problem the partnership solves: “Creative is a key driver of advertising outcomes, but for too long it has been measured in isolation, disconnected from media results.” The first live client is Ajinomoto, the global food and nutrition company.

Why This Matters For SEO And Digital Marketing Professionals

The traditional advertising agency’s anxiety about Unilever’s creator pivot was understandable but slightly misdirected. The real disruption isn’t that Unilever is working with 300,000 influencers instead of three ad agencies. The real disruption is that when 71% of those creators are using AI tools to produce content at speed, and that content is being distributed across dozens of platforms in hundreds of markets simultaneously, the evaluation infrastructure that used to separate good creative decisions from bad ones stops working.

Human panels are too slow. A/B testing individual pieces of content across a 300,000-creator network is logistically impossible. Traditional brand-tracking surveys capture what happened last quarter, not what’s working right now.

What DAIVID and ADIN.AI are building is the kind of infrastructure that makes the Unilever model actually governable – a system that can score creative at scale, link those scores to media performance in real time, and surface the signal from the noise before the budget has already been allocated to the wrong places.

Shelley Walsh made the point in her recent Search Engine Journal article on AI content scaling that enterprise brands face a specific trap: They know what they want to do (scale content production) but not how to do it without sacrificing the quality signals that make the content worth producing. The DAIVID and ADIN.AI partnership doesn’t solve the content quality problem. But it does solve the evaluation problem – which is arguably more urgent when you’re managing 300,000 creators rather than three.

For SEO professionals and content marketers, the practical implication is familiar. The distribution channels are changing, the production tools are changing, and the volume is increasing. What stays constant is the need to measure what’s actually working and make decisions based on that measurement rather than assumptions. That’s true whether you’re optimizing for search citations or creator content performance. Ground truth it, as always.

More Resources:

Featured Image: elenabsl/Shutterstock

Ecommerce MGMT 0 Comments
May 19 2026
Creating ‘Non-Commodity’ Content That Cuts Through The Noise

Google’s recent definition of commodity vs. non-commodity content is a bit meh. Meh if I’m being kind. Downright useless if I’m being more reasonable.

Complete and utter rubbish if I’ve had a drink.

Google's @dannysullivan on commodity vs non-commodity content https://t.co/lOelMIgQtP via @marthavanberkel and @gaganghotra_ and others
Image Credit: Harry Clarkson-Bennett

They all read like headlines you’d see in Discover and scroll past very quickly.

Maybe in a few years, that’ll be all that’s left, and that’s what Googlers are prepping us for. Personally, I think it’s far more likely their idea of quality, interesting content is just a bit rubbish.

Marble vs. grape juice – what a stupid title. Although interesting that they specify this is a video. Don’t hate the shoe one. No idea how that will make money for anyone, however… Doesn’t matter to Google.

Anyway, here’s how I think you can create unique, interesting content that still drives actual value to your business. (Hint: It’s not about grape juice).

TL;DR

  1. Commodity content is doomed for two reasons: It is easily summarized (because it has been done to death), and it doesn’t make (as much) money in a zero-click world.
  2. If you are creating content just for SEO and have nothing unique to offer, stop. You are throwing money down the drain.
  3. Be more than an SEO. Help other teams structure their workflows to generate the maximum value from all channels, with things like demand analysis.
  4. Google calculates the uniqueness of a document using a custom “information gain” score at a query and document level.

Why Commodity Content Is Doomed

People are like water. We take the easiest possible route. One that really doesn’t include clicking to find an answer, even if said answer is riddled with BS.

Commodity content – content that has been the bedrock of evergreen search strategies for years – can be very effectively summarized and synthesized by answer engines. So effectively that people will be satisfied with said clickless search.

Direct from the greedy horse’s mouth:

“Focus on making unique, non-commodity content that visitors from Search and your own readers will find helpful and satisfying. Then you’re on the right path for success with our AI search experiences, where users are asking longer and more specific questions — as well as follow-up questions to dig even deeper.”

Succeeding in AI Search

This means we have to focus our efforts elsewhere.

We have to focus our time and efforts on content more likely to drive legitimate value. Content that cannot easily be summarized by AI adds something of real value to the user and hasn’t already been thrashed to death by savvy SEO teams.

If you’re unsure whether to create content or not, ask yourself two questions:

  1. Are we creating this just for SEO?
  2. Are we adding anything unique to the existing corpus of information?

If you answered 1. Yes and 2. No, throw it straight in the bin.

You do not have the time, money, or resources anymore to spend time on content that doesn’t drive value.

Does This Mean Things Like Search Volume Are Useless?

At an individual keyword level, search volume has been declining in value for a long time. We just can’t generate the value we once could, and it isn’t coming back.

But search volume just indicates demand. If you’re savvy and use monthly data, you can help content, social, paid marketing, and editorial teams understand when users really care about a topic.

In this capacity, your job is to help teams understand when to create or update content, what that content should cover, and crucially, why it’s spiking in search at this particular time.

Searches for family holidays on Google Trends in the UK market
Five years’ worth of searches in Google for [family holidays] (Image Credit: Harry Clarkson-Bennett)

If we take searches for [family holidays] in Google Trends as an example, there is clear and obvious consistency. Searches spike every January as people plan their family holidays for the year ahead in the bleak midwinter.

So you should still get your core family holiday content ready for January. But as we shouldn’t operate in a silo, you should share this with social and travel teams so they know what time of year this type of content will generate the most value.

Planning and structure take center stage.

It is no longer about “Create x, get y.” That click-based marketing is dead.

Commodity Or Not Commodity

Loosely, this header was a Shakespearean-based to be or not to be joke, which is a. clunky and b. outside of my wheelhouse.

A picture of Shakespeare saying commodity or not commodity, that is the question
Image Credit: Harry Clarkson-Bennett

Now I’ve had to explain it.

I wrote about this in “How to do evergreen content in 2026 and beyond.” Which is, ironically, quite a commodity topic. But it has evolved. There’s new stuff to share. You can make commodity, non-commodity.

But you need to have a level of understanding and expertise that can really elevate a topic. That requires experience, a level of uniqueness, and a platform. Your content needs to be found, and what we have always done in search is unlikely to be anywhere near as valuable.

The Pillars Of Non-Commodity Content

  • Uniqueness.
  • E-E-A-T.
  • Engagement.
  • Structure.

Uniqueness

Uniqueness is the bedrock of everything when it comes to content that will continue to drive value. Without uniqueness, there’s no E-E-A-T. You won’t generate any shares, likes, comments, or links. Certainly not any good ones.

You can make this as fancy as you like.

If you’re lucky enough to have access to high-quality data sources like Similarweb, you can create some truly brilliant proprietary metrics that elevate your content above and beyond.

Let me give you an example.

Similarweb gives excellent engagement data at a site level. App-level too. If I was to combine these three metrics (pages per session, session duration and bounce rate) I have a composite engagement score.

Something no one else has.

If I took that engagement score and correlated it with third-party traffic data or something like branded search/backlinks, I could correlate engagement data with traffic from search over time.

A composite engagement score of newspapers broken down by type - young vs old
This is part of our audience engagement index (coming soon!) Image Credit: Harry Clarkson-Bennett

This is what stands out. This is what audiences will read, share, and crucially, remember. It requires more effort.

And as we know from the Google Leak (this brilliant warehouse from Daniel Foley Carter is superb), effort is quite literally estimated and scored by Google. Things that are difficult to replicate are rewarded.

Unless they’re absolutely insane. Then probably the opposite.

You don’t get good at this overnight. But Google has been prepping us for this for some time. If you look at the declining youth engagement in the above graph, maybe people have, too.

Not everyone is fortunate enough to have access to Similarweb data. But that doesn’t matter. Creativity and quality research is more important (and more readily available) than ever.

There are so many quality free data sources – Google Trends (combined with Glimpse), Keyword Planner, some free plans on tools like Ahrefs or Similarweb etc. You just need to identify metrics and combine them to make something bigger and better.

Google Attempts To Quantify Information Gain

Google has a patent (US20200349181A1) called Contextual estimation of link information gain that shows how the search giant may score the added value each document provides when compared to other similar documents.

How Google's attempt to quantify information gain works in practice
Documents are identified against a topic, scored, compared and presented based on the user’s likely need (Image Credit: Harry Clarkson-Bennett)

“In some implementations, information gain scores may be determined for one or more documents by applying data indicative of the documents, such as their entire contents, salient extracted information, a semantic representation across a machine learning model to generate an information gain score.”

Patents aren’t absolute. Just because a patent is present, it doesn’t mean it is always in use. If they’re frequently cited, recently updated, and have worldwide applications, that’s usually a very good sign they have a level of importance.

Screenshot from Google's information gain patent showing worldwide usage
This patent is all of those things (Image Credit: Harry Clarkson-Bennett)

But “ranking factors” aren’t absolute either. SERPs and topics are vastly different. It’s why we have subtopics like local SEO, YMYL, et al.

What matters for one term or topic may not matter as much, if at all, for another. It’s the nuance of the job and why trial and error is so important.

You don’t know until you know.

Consider The Four E’s

Your content needs a purpose.

Yes, it needs to convert. That is a business purpose. But it needs a purpose for people. Is it designed to entertain? Educate? As audiences turn away from news (and probably more widely, commodity content), this matters more than ever.

What we now term as commodity content was never designed to do any of the above. It was just designed to make money. Over the years, anything substandard propped up by Google just to make money has died.

This is the next cab of the rank.

E-E-A-T

E-E-A-T has taken a bit of a kicking recently. Not without reason.

The premise is sound. Not unreasonable for readers to expect the author to be, you know, a real person, who knows something and has some kind of online presence. And Google absolutely does track authorship and entities. Plenty of evidence of that.

Google has built and maintained its Knowledge Graph for decades, and entities have been the bedrock of news SEO for years. But E-E-A-T requires you to join the dots. To remove ambiguity – something we call disambiguation.

Google's knowledge panel with Florence Price
The Knowledge Graph and disambiguation in action (Image Credit: Harry Clarkson-Bennett)

Doesn’t mean doing this is incredibly valuable, but it’s foundational. Particularly in this modern-day iteration of the internet.

Remember, E-E-A-T Projects Have To Add Value

The problem with the whole – use experts, showcase expertise, prove you test everything, create video, make an effort in the industry, etc. – is now twofold:

  1. It’s expensive.
  2. And less valuable than ever.

Having that person build some kind of profile in the industry. A platform that their content can be shared from and that reduces reliance on search can only be a good thing.

A moat, if you will.

If they’re a legitimate expert on the topic, know how to structure great content and effectively showcase expertise, then you’re onto a bloody winner.

Which is why commodity content is doomed. Because people don’t care about it, and now it doesn’t drive value.

We need to find ways to make non-commodity content truly valuable to the business. If it isn’t driving some kind of trackable value, ignore it. Move on.

Be ruthless, brave and interesting.

Content just for SEO has diminishing returns. It’s almost certainly a bad idea IF you do it the same way you have been for the last 10 years.

Engagement

I have always felt that links should be a happy byproduct of creating and sharing brilliant stuff.

Leadership in SEO backlink overview from Ahrefs
Make me an offer, link sellers. I’m all ears. (Image Credit: Harry Clarkson-Bennett)

I’ve never made an effort to build links. I have just made an effort to write stuff I think is interesting, made some semi-libelous jokes, and got out there in the industry.

That is, more or less the Google definition of link building. In their world of sunshine, links are just earned by doing beautiful things. I am, in this scenario, the poster boy for white hat SEO.

The problem is, people need to make money, and links still drive rankings. So there’s a market there. And if you’re a student of the scriptures like I am, you’ll know the buying and selling of links is the oldest recorded job.

Either way, my inbox is full.

Anyway, your content has to fulfill a need. We’re moving away from straight-laced content, being able to do that for you as a publisher. Traditional ad revenue and the volume model sucks, and you sure as hell aren’t going to drive any subscriptions with what time is x or how to tie your shoes.

I really hope this is a good thing for SEOs and publishers. I want us to focus on content that really makes a difference to people’s lives. Content that makes them smile or think.

Content that makes people angry has been a big hit when it comes to numbers for a long time. But I don’t think anger is the emotion you should shoot for.

Measurement

You need to measure quality engagement, on and off-site. That means:

On-Site

No need to overcomplicate it for now.

  • Session duration.
  • Bounce rate.
  • Link clicks.
  • Pages per session.
  • Comments.
  • Read time.

Off-Site

Very much depends on the platform and the purpose, but I would focus on:

  • Links.
  • Shares.
  • Comments.
  • Saves.
  • Watch time.

You need to track metrics that tell you clearly whether people truly care about what you are creating. Clicks are dying, so I’d rather be measured against something a. more valuable and b. less miserable.

Create a composite metric(s) that gives you and your creators something to clearly focus on. Make their job easy by guiding their content with simple, straightforward metrics. Metrics that don’t just chase page views.

Structure

Structure’s not sexy. Let’s be honest.

But it matters. If, for some reason, you think LLMs are the zenith of society and content consumption, then you should know that models are more likely to cite or reference content from the top or bottom of the page, thanks to their inability to properly follow an argument.

This is known as the lost in the middle effect.

An overview of page structure
Semantic markup is still the foundation of a well-ordered page (Image Credit: Harry Clarkson-Bennett)

Unless, of course, the entity and topic are repeatedly referenced throughout.

I shouldn’t have to tell you that this is a bad idea and your content will become unreadable to living, breathing people.

But maybe you don’t care about that anymore.

Proper structure really matters. People have expectations (and accessibility needs). In more traditional commodity content, they want their question answered immediately. If you satisfy that – and the intro to your article isn’t abysmal – you might generate a longer session, a click, or hell, maybe even a conversion.

Theoretically, non-commodity content accessed via search should still be intent-driven. Possibly more so if we’re to believe the more qualified users with longer tail queries theory Google espouses.

So you still need to follow a similar, highly coherent page structure:

  • Answer the question.
  • Some form of TL;DR article summary.
  • Argument.
  • Concluding thoughts.
  • Coherent FAQs (if applicable).

One that logically answers queries in the appropriate format – text, video, image, list, etc. – and is highly consumable.

The argument section is where LLMs tend to lose their ability to accurately and appropriately cite and reference content. Which is not at all dissimilar to people.

I am not saying you need to continually refresh and restate the entity in question. That may be construed as keyword stuffing. It needs to read well for people. But you need to be clear, concise and accurate to make consuming your content simple.

Don’t People Consume Content In Different Ways?

You’re absolutely right, my pedantic friend, they do. Broadly, I think there are four types of consumption:

  1. Scanners: The vast majority. Too lazy or illiterate to read the whole thing, but will be satisfied from a headline, bold text, bullet points, and headers. They treat a page like a map, not a story.
  2. Answer seekers: They find what they want and leave. But still leave satisfied.
  3. Visual/audio consumers: A cohort that either refuses to or cannot read, but will stare at a pretty picture for 60 seconds.
  4. Deep readers: A small cohort, but a deeply engaged one, desperate for you to get something wrong.

I suspect these groups cover more than 90% of people. There are also fact-checkers – who skip the narrative and head straight for the citations, data points, or the “About Us” section before deciding if the content is worth their time.

And community-readers, who scroll to the bottom of the article to see the community reaction before deciding whether the content is worth their time. This is (obviously) more of a social trait. Particularly from younger audiences.

Your content can and should satisfy all of these people. It must:

  • Answer the question.
  • Be highly scannable.
  • Broken up with clear, distinct headers.
  • Form a concise, easy-to-follow narrative.
  • Be highly scannable.
  • Easy to share.
  • Visually appealing (audio and video options available).
  • Cite sources and clearly explain your methodology if appropriate.

You might think it’s beneath you, but if you don’t optimize for scanners and answer-seekers, you risk losing up to c. 80% or more of your potential audience within the first few seconds.

This is why front-loading (putting the most important info at the top) and using clear hierarchies is so vital in modern writing.

Anyway, that’s it. Thanks for reading as always!

More Resources:

Read Leadership In SEO. Subscribe now.

Featured Image: Roman Samborskyi/Shutterstock

Ecommerce MGMT 0 Comments
Apr 22 2026
Does AI Actually Reward Quality Content?

For well over a decade, SEOs and marketers have debated the importance of high-quality, original content. After just about every major update, the message from Google was clear: If you want to rank, cut it out with the derivative listicles and other quick-churn assets that are big on keywords and light on substance.

More recently, our current understanding of how LLMs select which sources to cite in responses has SEOs and content marketers championing high-quality, original, and in-depth content with renewed fervor. If you want AI to identify your content as the best source with which to answer a user’s query, logically, it must be among the best online content available on the topic.

While that’s all great in theory, I’m sure many of you reading this have experienced that crushing disappointment on publishing, only for it to sink like a stone with barely a ripple. Somehow, your magnum opus languishes on page 4 of the relevant search results, outranked by content that, in your humble opinion, isn’t that remarkable.

Can we really call something high quality if it doesn’t achieve the strategic outcome that led us to create it?

Even when our content succeeds, there’s still the nagging worry that we might perhaps be investing too much time and money trying to achieve content perfection. Did that white paper really need to be 10 pages? Or would a simpler, five-page version have done just as well?

Might it be possible to achieve the same results with a little less quality? How do we find the sweet spot? In short, what’s the minimal viable product?

I’m not going to pretend to have the answer. And that’s because the question isn’t clear on what we mean by quality content.

A Question Of Quality

I’m as guilty as anyone of writing about the need for high-quality content as if it’s obvious what it is and how to achieve it without any further explanation. It’s a form of industry shorthand that has become increasingly meaningless through overuse.

Ask 10 CMOs, SEOs, and content marketers to define what they mean by high-quality content, and you’ll probably get 15 different answers.

Is “quality” determined by thought leadership and subject matter expertise? Or can a few average thoughts be elevated to high quality with skilled writing, a strong layout, and some clever design work?

Is “depth” characterized by longer word counts and more detailed research? Or is it really about demonstrating a superior understanding of a topic by exploring more nuanced or highfalutin’ ideas? Never mind the graphs, can you somehow weave in some Ancient Greek philosophy to get the point across?

And how much originality adds up to “original”? If you reference someone else’s work, are you somehow detracting from your own originality score?

While I can’t confidently give you a single, unambiguous definition of what high quality is, I can tell you what it isn’t: While it may be important, high-quality content is no silver bullet.

Just because your content is meticulously researched and extremely well executed doesn’t mean it’s somehow entitled to high rankings.

Does Original Content Actually Perform Better?

I tasked my team with conducting some qualitative research to answer the question: Does original content perform better than repurposed, unoriginal content, in both traditional search and AI-generated responses?

Of course, the internet is a big place (who knew?). So, for the purposes of this study, we restricted the definition of “search” to Google’s search results and to citations within AI platforms Gemini, ChatGPT, and Perplexity.

Similarly, because you’ve got to compare apples with apples, the team focused on popular search queries in the B2B SaaS and professional services space; mid-funnel, informational queries like “marketing automation tools” and “email deliverability tools.”

The team then identified and analyzed the top-ranking URLs for each query before assigning each one a score from 0 to 3 in five different categories.

  • Primary contribution.
  • Structural novelty.
  • Interpretive depth.
  • Source dependence.
  • Contextual insight.

With a maximum total score of 15, each page was then classified as follows:

  • 12-15: Group A (Original).
  • 7-11: Borderline (Excluded).
  • 0-6: Group B (Repurposed).

When the data came back, it appeared at first glance that URLs with higher originality scores (Group A) do tend to rank more consistently in Google and appear more frequently in AI responses than repurposed or derivative content (Group B).

However, before all the content marketers scream “I told you so” at anyone in earshot, you might want to read this next bit first.

Data analysts are notoriously skeptical of knee-jerk first glance conclusions (again, who knew?). The team crunched the data further, using data sciency techniques involving far more Greek letters than I’m used to seeing. They concluded that, while the correlation exists, it’s weak. Strong performance in one part of the dataset doesn’t reliably predict strong performance elsewhere in the dataset. The relationship simply isn’t consistent enough to say with any confidence that highly original content performs better every time.

Even so, while the correlation may be weak, it doesn’t appear to be entirely random. Looking at the overall averages, stripped of extreme cases that might skew the results, we did detect a pattern.

For example, original content appeared to perform better in relation to queries requiring interpretation or judgment, such as “benefits of marketing automation” or “email marketing best practices.” But that relationship virtually disappeared for more straightforward requests for information like “what is marketing automation.”

This makes sense. When the answer is factual, being original matters less than being accurate. When the answer requires perspective or judgment, originality becomes more valuable.

So, where does that leave us? We can’t confidently prove that original content always outperforms repurposed content. On the other hand, we can rule out the idea that originality has no impact at all. Therefore, what we can say is that original insight helps in some contexts, for some query types. It just isn’t a guaranteed lever you can pull for predictable results.

When Mediocre Content Has The Edge

Back in the 2010s, the API industry was booming. And that meant lots of content being published on every aspect of how APIs function. At the very least, a software company would need to publish detailed documentation for each of its APIs, from technical specifications and structures to implementation guides and walkthroughs.

This created a problem for one of our clients, a small startup of 10 people: How could they compete for visibility in search, let alone attract positive attention, when the entire conversation around APIs appeared to be dominated by industry giants? The competitors already had massive online footprints, larger content budgets, established domain authority, and significantly more comprehensive resources. How could we ever outrank them?

Conventional wisdom might have seen us attempt to fight quantity with quality by creating the best possible online resource on the topic of APIs. If we could publish content that goes far deeper and offers more value than the competition, we might gradually earn trust and authority through original, detailed research and thought leadership.

With enough budget and a long-term commitment, you could definitely build a strategy around such an approach. Except, of course, we would have needed both quality and quantity to have any chance of overtaking their competitors.

Trying to compete for visibility in every relevant subtopic and keyword on the subject of APIs would mean fighting on way too many fronts at once. How could we find an original angle on a topic that’s already well served online? How could we talk about APIs in a way that would differentiate their software from everyone else’s?

Short answer: We couldn’t. So, we flipped the problem. What if, instead of being last to join the race for the most relevant keywords today, we could be first out of the blocks in the race for whichever keyword might become relevant tomorrow?

I sent out a survey to the relevant audience, asking a bunch of typical users what search terms they would use in certain scenarios. The results revealed a plethora of short- and long-term keywords, but when we looked for any common themes, two words stood out. One was “API,” naturally. The other was “design.”

“API design” hadn’t cropped up in our initial keyword research as a potential opportunity. But as the search volume for “API design” was practically zero, that’s hardly surprising. Yet we now had clear evidence that, as the industry matured, so too would the search terms people used.

And because very few currently search for “API design,” none of the competitors appeared to be targeting the keyword or publishing content on the topic at all.

This was our window of opportunity. Never mind original content: We had an original keyword, an entire topic niche, to ourselves.

However, we also knew the value of that keyword would evaporate overnight if one or more competitors got there before us.

Forget spending six months developing an award-winning whitepaper series. We didn’t need perfection – with all the time, expense, and effort that entails – because we were staring at the SEO equivalent of an open goal.

In just a few days, we threw together a simple landing page focused on API design. It wasn’t exceptional. At only about 1,500 words, it wasn’t comprehensive. As content goes, it was pretty mediocre. But that’s all it took.

About 12 months later, just as predicted, the search volume materialized. Our single modest page continued to outrank every major competitor, even when they started chasing that new search volume with their own landing pages and content hubs.

Within two years, the keyword “API design” was worth approximately £200 per click. But our client didn’t need to pay for clicks. In effect, we won the space before anyone else even realized there was a space worth winning.

Perfection Is The Enemy Of Good

Striving to achieve the best possible iteration of your content, endlessly refining and polishing and second-guessing every detail, can get in the way of just getting it out there. Sometimes, good enough really is good enough.

I’m not arguing that we should stop striving for excellence in our content. As I hope our little study demonstrated, there are situations where well-researched, original content can give you an advantage. And, of course, success doesn’t end with rankings, citations, and clicks. Once they land on your content, you still want visitors to be wowed, persuaded, and motivated into action.

But like so many things in life, success depends on timing at least as much as it does on quality or originality. In a way, that’s what originality is all about; not necessarily being best but being first.

The API design landing page didn’t succeed because it was mediocre. It succeeded because they got there first. Quality mattered, but not in the way most content strategies define it.

This matters even more in AI search. LLMs can curate ideas and summarize information, but they can’t have original thoughts, provide firsthand experiences, or offer up fresh perspectives (as of now). While there are no guarantees, as our limited research shows, in AI at least, being the original source has influence.

Start asking what your content can say that hasn’t already been said, and then say it before someone else does.

More Resources:

Featured Image: ImageFlow/Shutterstock

Ecommerce MGMT 0 Comments
Apr 14 2026
Shorter, Focused Content Wins In ChatGPT via @sejournal, @Kevin_Indig

Boost your skills with Growth Memo’s weekly expert insights. Subscribe for free!

For years, SEOs have operated on a simple assumption: The more ground your content covers, the more likely it is to surface in AI-generated answers. In fact, every “best practice” in classic SEO content pushes you toward more: more subtopics, more sections, more words. Build the “ultimate guide.”

An analysis of 815,000 query-page pairs across 16,851 queries and 353,799 pages says otherwise:

  • Fan-out coverage is nearly irrelevant to citation rates.
  • Two signals actually predict whether ChatGPT cites your page.
  • Six concrete changes to your existing content library help.

1. The Study

AirOps ran 16,851 queries through ChatGPT three times each through the UI, capturing every fan-out sub-query, every URL searched, every citation made, and every page scraped. Oshen Davidson built the pipeline. I analyzed the data.

Each query generates an average of two fan-out queries. ChatGPT retrieves roughly 10 URLs per sub-search, reads through them, then selects which ones to cite. We scored how well each page’s H2-H4 subheadings matched those fan-out queries using cosine similarity on bge-base-en-v1.5 embeddings. That score is what we call fan-out coverage: the share of subtopics a page addresses at a 0.80 similarity threshold. (The 0.80 similarity threshold cutoff was used to decide whether a subheading counts as a match to a fan-out query. Think of it as a relevance bar.)

The question: Do pages with higher fan-out coverage get cited more?

You’ll find even more information in the co-written AirOps report.

2. Density Barely Moves The Needle

Across 815,484 rows, the relationship between fan-out coverage and citation is weak.

Covering 100% of subtopics adds 4.6 percentage points over covering none. That gap shrinks further when you control for query match (how well the page’s best heading matches the original query). Among pages with strong query match (>= 0.80 cosine similarity):

Image Credit: Kevin Indig

Moderate coverage (26-50%) outperforms exhaustive coverage. Pages that cover everything score lower than pages that cover a quarter of the subtopics. The “ultimate guide” strategy produces worse results than a focused article that covers two to three related angles well.

3. What Actually Predicts Citation

These two signals dominate: retrieval rank and query match.

1. Retrieval rank is the strongest predictor by a wide margin. A page at position 0 in ChatGPT’s web search results (the first URL returned by its search tool) has a 58% citation rate. By position 10, that drops to 14%. We ran each prompt three times consecutively for this analysis, and pages cited in all three runs have a median retrieval rank of 2.5. Pages never cited: median rank 13.

Image Credit: Kevin Indig

2. Query match (cosine similarity between the query and the page’s best heading) is the strongest content signal. Pages with a 0.90+ heading match have a 41% citation rate compared to the 30% rate for pages below 0.50. Even among top-ranked pages (position 0-2), higher query match adds 19 percentage points.

Fan-out coverage, word count, heading count, domain authority: all secondary. Some are flat. Some are inversely correlated.

4. The Wikipedia Exception

One site type breaks the pattern. Wikipedia has the worst retrieval rank in the dataset (median 24) and the lowest query match score (0.576). It still achieves the highest citation rate: 59%.

Wikipedia pages average 4,383 words, 31 lists, and 6.6 tables. They are encyclopedic in the literal sense. ChatGPT cites Wikipedia from deep in the search results where every other site type gets ignored.

This is density working as a signal, but at a scale no publisher can replicate. Wikipedia’s content is exhaustive, richly structured, and cross-linked across millions of topics. A 3,000-word corporate blog post with 15 subheadings is not the same thing.

5. The Bimodal Reality

58% of pages retrieved by ChatGPT in this dataset are never cited. 25% are always cited when they appear. Only 17% fall in between.

The always-cited and never-cited groups look nearly identical on most content metrics: similar word counts (~2,200), similar heading counts (~20), similar readability scores (~12 FK grade), similar domain authority (~54). The on-page signals we can measure do not separate winners from losers.

What separates them is retrieval rank. Always-cited pages rank near the top when they surface. Never-cited pages rank in the bottom half. The retrieval system, whatever signals it uses internally, is the gatekeeper. Everything else is a tiebreaker.

6. What This Means For Your Content

Conventional SEO content writing wisdom says cover more subtopics, add more sections, build density. The data says the conventional approach produces “mixed” pages, the 17% in the middle that get cited sometimes and ignored other times.

Mixed pages have the highest word counts, the most headings, and the highest domain authority in the dataset. They are the “ultimate guides.” They are also the least reliable performers in ChatGPT.

The pages that win consistently are focused. They:

  • Match the query directly in their headings,
  • Tend to be shorter (the citation sweet spot is 500-2,000 words), and
  • Have enough structure (7-20 subheadings) to organize the content without diluting it.

Build the page that is the best answer to one question. Not the page that adequately answers 20.

Featured Image: Tero Vesalainen/Shutterstock; Paulo Bobita/Search Engine Journal

Ecommerce MGMT 0 Comments
Apr 1 2026
How To Do Evergreen Content In 2026 (And Beyond)

Fair to say the majority of evergreen content will not drive the value it did five years ago. Hell, even one or two years ago. What we have done for the last decade will not be as profitable.

AIOs have eroded clicks. Answer engines have given people options. And to be fair, people are bored of the +2,000-word article answering “What time does X start?” Or recipes where the ingredient list is hidden below 1,500 words about why daddy didn’t like me.

In response to this, publishers say it will be important to focus on more original investigations and less on things like evergreen content (-32 percentage points).

So, you’ve got to be smart. This has to be framed as a commercial decision. Content needs to drive real business value. You’ve got to be confident in it delivering.

That doesn’t mean every article, video, or podcast has to drive a subscription or direct conversion. But it needs to play a clear part in the user’s journey. You need to be able to argue for its inclusion:

  • Is it a jumping-off point?
  • Will it drive a registration?
  • Or a free subscriber, save or follow on social

More commonly known as micro-conversions, these things really matter when it comes to cultivating and retaining an audience. People don’t want more bland, banal nonsense. They want something better.

The antithesis to AI slop will help your business be profitable.

Inherently, nothing. It’s a foundational part of the content pyramid.

In most cases, it’s been done to death, and AI is very effective at summarizing a lot of this bread-and-butter content.

Over the last 10 years, it’s been pretty easy to build a strategy around evergreen content, particularly if you go down the parasite SEO route. Remember Forbes’ Advisor and the great affiliate cull?

The epitome of quantity over quality; it worked and made a fortune.

But I digress.

An authoritative enough site has been able to drive clicks and follow-up value with sub-par content for decades. That is, slowly diminishing. Rightly or wrongly.

And not because of the Helpful Content stuff. Google nerfed all the small sites long before the goliaths. Now they’ve gone after the big fish.

We have to make commercial decisions that help businesses make the right choice. Concepts like E-E-A-T have had an impact on the quality of content (a good thing). It’s also had an impact on the cost of creating quality content.

  • Working with experts.
  • Unique imagery.
  • Video.
  • Product and development costs.
  • Data.

This isn’t cheap. Once upon a time, we could generate value from authorless content full of stock images and no unique value. Unless you’re willing to bend the rules (which isn’t an option for most of us), you need an updated plan.

It depends.

You need to establish how much your content now costs to produce and the value it brings. Not everything is going to drive a significant conversion. That doesn’t mean you shouldn’t do it. It means you need to have a very clear reason for what you’re creating and why.

If particular topics are essential to your audience, service, and/or product, then they should at least be investigated.

One of the joys of creating evergreen content has always been that it adds value throughout the year(s). A couple of annual updates, even relatively light touch, could yield big results.

Commissioning something of quality in this space is likely more expensive. It needs to be worth it; it has to form part of your multi-channel experience to make it so.

  • Unique data and visuals that can be shared on socials.
  • Building campaigns around it (or it’s part of a campaign).
  • You can even build authors and your brand around it.
  • And if it resonates, you can rinse and repeat year after year.
Ahrefs created demand for their brand + an evergreen topic – AIOs (Image Credit: Harry Clarkson-Bennett)

And this type of content or campaign can increase demand for a topic. You can become a thought leader by shifting the tide of public opinion.

For publishers and content creators, that is foundational.

Two broadly rhetorical questions:

  1. Do you think in a world of zero click searches, clicks and reach are sensible tier one goals?
  2. Do you want to be targeted against a metric that is very likely to go down each year?
Like it or not, people really do use AIOs (Image Credit: Harry Clarkson-Bennett)

I don’t – on both counts. We should want to be targeted on driving real value for the business.

Something like:

  1. Tier 1: Value – core, revenue, and value-driving conversions.
  2. Tier 2: Registrations (and things that help you build your owned properties), links, shares, and comments.
  3. Tier 3: Page views, returning visits, and engagement metrics.

Micro-conversions over clicks. We’re focusing on registrations, free or lower-value subscriptions. Whatever gets the user into the ecosystem and one step closer to a genuinely valuable conversion.

The messy middle has changed, and it is largely unattributable (Image Credit: Harry Clarkson-Bennett)

Now, could a click be a micro-conversion? If you know that someone who reads a secondary article (by clicking a follow-up link) is 10x more likely to register, that follow-up click could be a sensible micro-conversion.

This type of conversion may not directly drive your bottom line. But it forces you and your team to focus on behaviors that are more likely to lead to a valuable conversion.

That is the point of a micro-conversion. It changes behaviors.

You can tweak the above tiers to better suit your content offering. Not all content is going to drive direct tier one or even two value. You just need to have a very clear idea of its purpose in the customer journey.

If what you’re creating already exists, you’d better make sure you add something extra. You’ve got to force your way into the conversation, and unless you can offer something unique, you’re (almost certainly) wasting your time IMO.

I’ll break all of these down, but I think (in order of importance):

  1. Writing content for people.
  2. Information gain.
  3. Getting it found.
  4. Creating it at the right time.
  5. Structuring it for bots.

Everyone is obsessed with getting cited or being visible in AI.

I think this is completely the wrong way of framing this new era. Getting cited there, or being visible, is a happy byproduct of building a quality brand with an efficient, joined-up approach to marketing.

The more you understand your audience, the more likely you will be to create high-quality, relevant content that gets cited.

If you know your audience really cares about a topic, that’s step one taken care of. If you know where they spend time and how they’re influenced, that’s step two. And if you know how to cut through the noise, that’s step three.

Really, this is an evolution in SEO and the internet at large.

  • Invest in and create content that will resonate with your audience.
  • Create a cross-channel marketing strategy that will genuinely reach and influence them.
  • Share, share, share. Be impactful. Get out there.
  • Make sure it’s easy to read, share, and consume.

Your content still needs to reach and be remembered by the right people. Do that better than anybody else, and wider visibility will come.

In SEO, we have a different definition of information gain than more traditional information retrieval mechanics. I don’t know if that’s because we’re wrong (probably), or that we have a valid reason…

Maybe someone can enlighten me?

In more traditional machine learning, information gain measures how much uncertainty is reduced after observing new data. That uncertainty is captured by entropy, which is a way of quantifying how unpredictable a variable is based on its probability distribution.

Events with low probability are more surprising and therefore carry more information. High probability events are less surprising and novel. Therefore, entropy reflects the overall level of disorder and unpredictability across all possible outcomes.

Information gain, then, tells us how much that unpredictability drops when we split or segment the data. A higher information gain means the data has become more ordered and less uncertain – in other words, we’ve learned something useful.

To us in SEO, information gain means the addition of new, relevant information. Beyond what is already out there in the wider corpus.

A representative workflow of Google’s Contextual estimation of link information gain patent (Image Credit: Harry Clarkson-Bennett)

Google wants to reduce uncertainty. Reduce ambiguity. Content with a higher level of information gain isn’t only different, it elevates a user’s understanding. It raises the bar by answering the question(s) and topic more effectively than anyone else.

So, try something different, novel even, and watch Google test your content higher up in the SERPs to see if it satisfies a user.

This is such an important concept for evergreen content because so many of these queries have well-established answers. If you’re just parroting these answers because your competitors do it, you’re not forcing Google’s hand.

Particularly if you’re still just copying headers and FAQs from the top three results. Audiences are not arriving at publisher destinations through direct navigation at the same scale. They encounter journalism incidentally, through social feeds, not through habitual site visits.

Younger audiences spend less time on news sites and more time on social every year (Image Credit: Harry Clarkson-Bennett)

You’ve got to meet them there and force their hand.

According to this patent – contextual estimation of link information gain – Google scores documents based on the additional information they offer to a user, considering what the user has already seen.

“Based on the information gain scores of a set of documents, the documents can be provided to the user in a manner that reflects the likely information gain that can be attained by the user if the user were to view the documents.”

Bots, like people, need structure to properly “understand” content.

Elements like headings (h1 – h6), semantic HTML, and linking effectively between articles help search engines (and other forms of information retrieval) understand what content you deem important.

While the majority of semi-literates “understand” content, bots don’t. They fake it. They use engagement signals, NLP, and the vector model space to map your document against others.

They can only do this effectively if you understand how to structure a page.

  • Frontloading key information.
  • Effectively targeting highly relevant queries.
  • Using structured data formats like lists and tables, where appropriate (these are more cost-effective forms of tokenization).
  • Internal and external links.
  • Increasing contextual knowledge gain with multimedia (yes, Google can interpret them).

The more clearly a page communicates its topic, subtopics, and relationships, the more likely it is to be consistently retrieved and reused across search and AI surfaces. This has a compounding effect.

Rank more effectively (great for RAG, obviously) – feature more heavily in versions of the internet – force your way into model training data.

If you need to get development work put through, frame it through the lens of assistive technology. Can people with specific needs fully access your pages?

As up to 20% need some kind of digital assistive technology, this becomes a ‘ranking factor’ of sorts.

I won’t go through this in much detail, as I’ve written a really detailed post on it. Basically:

  • Track and pay very close attention to spikes in demand (Google Trends API being a very obvious option here).
  • Make sure you’re adding something of value to the wider corpus.
  • If quality content is already out there and you have nothing extra to add, consider whether it’s worth spending money on (SEO is not free).
Create and update timely evergreen content (Image Credit: Harry Clarkson-Bennett)

While this is primarily for news, you can apply a similar logic to evergreen content if you zoom out and follow macro trends.

Evergreen content still spikes at different times throughout the year. Take Spain as an example. There’s much more limited interest in going to Spain in the Winter months from the UK. But January (holiday planning or weekend breaks) and summer (more immediate holiday-ing with the kids) provide better opportunities to generate traffic.

You’re capturing the spike in demand by updating content at the right time. Particularly if you understand the difference in user needs when this spike in demand happens.

  • In January, get your holiday planning content ready.
  • In the summer, get your family-friendly and last-minute holiday content up and running.
Image Credit: Harry Clarkson-Bennett

Demand for evergreen topics can be cyclical. In this example, you would want to capture the spike(s) with carefully planned updates, so you have up-to-date content when a user is really searching for that product, service, or information.

Well, what matters to your brand and your users? Have you asked them?

By the very nature of new and evolving topics and concepts, not everything “evergreen” has been done.

New topics rise. Old ones fall. Some are cyclical.

My rule(s) of thumb would be to establish:

  • Is the topic foundational to your product and service?
  • Does your current (and potential) audience demand it?
  • Do you have something new to add to the wider corpus of information?

If the answer to those three is a broad variation of yes, it’s almost certainly a good bet. Then, I would consider topic search volume, cross-platform demand, and whether the topic is trending up or down in popularity.

There are some things you should be doing “just for SEO.” Content isn’t one of them. You can yell topical authority until you’re blue in the face. If you’re creating stuff just for SEO – kill it.

IMO, these plays have been dead or dying for some time. The modern-day version of the internet (in particular search) demands disambiguation. It demands accuracy. Verification that you are an expert. Otherwise, you’re competing with those who have a level of legitimacy that you do not.

Social profiles, newsletters, real people sharing stories. You’re competing with people who aren’t polishing turds.

If all you’re thinking about is search volume or clicks, I don’t think it’s worth it.

YouTube and TikTok are flying. The young mind cannot escape big tech’s immeasurable evil.

They’re bored with reading the news, but they really, really like video. They will watch it.

TikTok and YouTube dominate (Image Credit: Harry Clarkson-Bennett)

The good news for you (and me) is that platforms like YouTube are still very viable opportunities to build something brilliant. Memorable even. They’re also far more AI-resilient – even if Google desperately tries to summarize everything with AI.

And this brings me nicely onto rented land. Platforms you don’t own.

We’ve spent years creating assets (your websites) to deliver value in search. Owning all of your assets and prioritizing your site above all else. But that is changing. In many cases, people don’t reach your website until they’ve already made a purchasing decision.

I think Rand has managed this transition better than anybody (Image Credit: Harry Clarkson-Bennett)

So, you have to get your stuff out there. Create large, unique studies. Cut them into snippets and short-form videos. Use your individual platform to boost your profile and the content’s chances of soaring.

This is, IMO, particularly prescient for publishers. You’ve got to get out there. You’ve got to share and reuse your content. To make the most of what you’ve created.

Sweat your assets. Even if senior figures aren’t comfortable with this, you need to make it happen.

People have been espousing how important it is to feature as part of the answer. And that may be true. But you’re going to have to be good at selling your projects in if there’s no clear attribution or value.

It might not have the spikes of news, but evergreen interest still spikes at certain times in the year.

Get people – real people – to share it. To have their spin on it.

Outperform the expected early stage engagement and maximize your chance of appearing in platforms like Discover with wider platform engagement.

You have to work harder than before.

I shared an example of this around a year ago, but to revisit it, I now have 11 recommendations from other Substacks.

You can’t do this alone (Image Credit: Harry Clarkson-Bennett)

They have accounted for over 40% of my total subscribers. Admittedly, mainly from Barry, Shelby, and Jessie. But they are, if I may be so bold, superhumans.

And when our main driver of evergreen traffic to the site (Google) has really leaned into the evil that surrounds big tech, we’ve got to be cannier. We have to find ways to get people to share our content.

Even evergreen content.

If we’re being honest, a lot of SEO content has been rubbish. Churned out muck.

People are still churning out muck at an incredible rate. When what you’ve got is crap, more crap isn’t the answer. I think people are turned off. They’re tuning out of things at an alarming rate, especially young people.

It is all about getting the right people into the system. Evergreen content is still foundational here. You just have to make it work harder. Be more interesting. Be shareable.

Hopefully, this makes decisions over what we should and shouldn’t create easier.

More Resources:

Read Leadership In SEO. Subscribe now.

Featured Image: str.nk/Shutterstock

Ecommerce MGMT 0 Comments