Chasing Trends Is for Losers

If the Semrush AI citation chart has you redirecting your content strategy toward Reddit and LinkedIn, you are chasing a trend.

I have been a brand story and ops consultant for almost two decades, and one of the most inefficient traps I watch companies fall into is exactly this one. Most brands are at least starting to move the right direction, expanding an SEO strategy into the infrastructure to win in GEO and AEO. But if you skimmed past what the chart actually means before reacting to it, you may have just taken several steps back instead of forward.

As soon as I saw the chart, I went deeper to figure out whether it changed anything about a long-form SEO and AEO strategy. Here is what I learned.

What the data actually shows

First, the numbers are real. The chart is not fake, and I am not here to tell you it is.

Semrush ran the analysis across 325,000 prompts on ChatGPT Search, Google AI Mode, and Perplexity. Reddit cited at roughly 11.29 percent of prompts. LinkedIn at 11.03 percent. Wikipedia at 9.53 percent. The chart is accurate.

The problem is not the data. It is what people do with it. Three things the chart does not tell you.

First, citation frequency is volatile. The same dataset, broken down month by month, shows ChatGPT cited Reddit in close to 60 percent of responses in early August 2025, then collapsing to around 10 percent by mid-September. Six times less, in six weeks, because OpenAI changed how the model weights Reddit content. The chart you saw was a snapshot of a moving target. A snapshot is not a strategy.

Second, citation is not endorsement. When an AI cites a Reddit thread or a LinkedIn post, it is just as likely using it as a complaint or a counterpoint as a recommendation. The system pulls content based on how closely it matches the meaning of the question, not on whether the content speaks well of you. A scathing review gets cited the same way a glowing one does.

Third, the content getting cited is the opposite of trendy. Up to 80 percent of Reddit threads cited by AI have fewer than 20 upvotes. The average age of a cited Reddit post is roughly 900 days. The machines are not surfacing this week's hot post. They are surfacing the settled, low-key, high-information thread from two and a half years ago that nobody hyped at the time.

That is the whole thing. AI search rewards consensus, age, and substance. It does not reward virality. If anything, it works against it.

How it actually works, in plain terms

If you want to know why, it helps to know what happens when you ask one of these tools a question.

The system turns your question into a kind of mathematical fingerprint of its meaning. Then it searches a giant library of web content that has been turned into the same kind of fingerprints, and pulls the pieces whose meaning lines up most closely with what you asked. Those pieces get handed to the AI to write its answer from.

The part that matters: that matching step does not care how viral something is. It cares how well the content actually answers the question. A clear, thorough piece that addresses the exact thing someone asked is easier to surface than a hot take that mentions the topic in passing. A piece from 2022 that nailed the established take on a subject is easier to surface than a post from this week still finding its angle.

There is a ranking layer on top that weighs a few more things. How authoritative the source is. Whether it lines up with other good sources. Recency, but much more weakly than old-school Google did. None of those layers reward chasing a trend either.

The systems are built to surface the canonical answer. The answer that has been said clearly, said often, said over time, and backed up by other credible voices. That is the literal opposite of trend-chasing.

What the data actually validates

Here is the part that should let you exhale.

The same Semrush study, read carefully, does not tell you to pivot. It tells you to keep doing the boring, durable thing. It found that long-form articles between 500 and 2,000 words drive the largest share of AI citations from LinkedIn. Educational content, the kind focused on actually teaching something, makes up 54 to 64 percent of cited posts. Reshares are rarely cited. And how often you publish original work matters more than how many followers you have. Authors with under 500 followers were cited as often as authors with thousands, when the work was good and the cadence was steady.

Read that back. Write your own thinking. Publish it consistently. Focus on substance. Do not optimize for virality. That is not a new strategy somebody discovered in a chart. It is the oldest one there is. Search Engine Land made the same point in a piece they bluntly titled stop chasing Reddit and Wikipedia. AI citation data is easy to misread, and acting on the surface reading produces noise instead of visibility.

So if you saw the chart and felt behind, you are not behind. Good content delivered consistently over time still wins. The chart is evidence for that, not against it.

Why we chase the chart anyway

Here is the part that is psychological, not technical.

Chasing a chart feels like work. You see the data. You make a plan. You start posting where the data says to post. You feel busy. You feel sharp. You feel like you are doing the thing the smart people are doing.

You are not doing the thing the smart people are doing. You are doing the thing that feels like it. The smart move is boring. Publish one substantive piece on a steady cadence on something you own, share it to a small set of places, and wait long enough for it to compound.

The compounding is invisible while it happens. That is exactly why most people cannot stay with it. They get six months in, the numbers have not moved, and they pivot to the next trend. The pivot feels like adaptation. The pivot is the constraint.

This is what I mean by avoidance masquerading as effort. The trend-chase is busy work that feels productive and builds nothing. Two years from now, the people who chased the LinkedIn chart, then the Reddit chart, then the Quora chart, then whatever comes next, will have a portfollio of scattered, mediocre content and no real position on anything. The people who picked a few questions and answered them honestly and consistently will own those questions. The effort is the same. The result is not.

The principle underneath all of it

You build a position by saying clear things, consistently, for long enough that they compound. Platforms change. Algorithms change. Trends change. The position is what survives all of it.

If you write about what you actually know, in your real voice, on a regular cadence, your work slowly becomes part of the canonical answer to the questions you care about. The machines surface canonical answers. You get cited because you are the answer, not because you cracked the algorithm.

If you write whatever the chart says this week, you are not the canonical answer to anything. You are noise that briefly matched a snapshot. The next snapshot looks different, and your work has not moved.

The honest work is harder, because it does not give you a hit at the end of every post. It gives you a body of work over months and years, most of it invisible until suddenly it is not. The trend-chase gives you a feeling of motion that resets every Monday.

What this looks like in practice

Stop reading viral posts about AI search optimization for two weeks. They are not telling you anything the systems actually reward.

Instead, pick three to five questions in your world that you are genuinely positioned to answer. Not topics. Questions. The specific things a customer or a peer would type into ChatGPT expecting a real answer.

Write one solid long-form piece a week answering one of them, on something you own, in your real voice. Cross-post to a place or two with proper attribution back to the original. Do not check your citation numbers for six months. They will not have moved, and that is fine. You are building a position, not chasing a metric.

In two years, the questions you committed to will have your name on them in AI search. The people who chased the chart will be on to the next one, with a pile of forgettable posts and no position to show for it.

Why I am betting my own work on this

Here is what I am doing.

I am translating my life's work into long-form, blog-style content. Almost two decades of presentations, pitches, strategy decks, creative briefs, wins, losses, lessons, podcast episodes, and a stack of personal notes, organized into something consistent and digestible, and published in public on a steady cadence. The whole point is to win in AEO over a long horizon. It is the number one goal. And like anything worth building, it takes time. Slow at first, then not.

When the chart crossed my feed, it would have been easy to abandon that plan and start chasing citations on whatever platform was trending. I did not, because everything I have learned in twenty years of this work, and everything the data above actually says, points the same direction. Build the position. Keep showing up. Let it compound.

That is the difference between chasing and building. The systems reward one. The dopamine rewards the other. Chasing trends is for losers. Good content delivered consistently over time always wins, and it always will. AI search did not change that. It just proved it again.

If you're a founder building the brand

If you are early and you want your brand cited in AI search, the pull is to chase whatever platform the latest study crowns. Resist it. The brand-story work that actually moves you into the canonical answer is narrower and slower. Pick the handful of questions only you can answer from real experience, and answer them in your own voice on a surface you own. A founder building the brand wins AI search the same way they win a room, by being the clearest, most consistent voice on a specific thing, not the loudest voice on everything.

If you're a CEO whose company has outgrown you

When the company gets big enough, trend-chasing stops being one person's habit and becomes the whole org's. Marketing chases the chart. The board asks why the citations are not moving. Everybody is busy and nothing compounds. When I am doing constraint coaching with a CEO, this rarely starts with the marketing tactic. It starts with the avoidance underneath it. Avoidance masquerading as effort is almost never one person's problem. It is a culture that rewards visible activity over the invisible work of building a position. The fix is leadership choosing the boring, compounding work out loud, so everyone else is finally allowed to do the same.

Frequently asked questions

Q: Is LinkedIn really the second-most cited domain by AI search engines? A: Yes. Semrush analyzed 325,000 prompts across ChatGPT Search, Perplexity, and Google AI Mode and found LinkedIn cited in about 11 percent of AI responses, just behind Reddit at 11.3 percent. The number is real. What is misleading is the conclusion people draw, that pivoting your strategy toward LinkedIn produces citations. The underlying patterns reward substantive long-form content on a steady cadence, on any platform, not posting volume.

Q: How do LLMs decide which content to cite in AI search results? A: The system turns your question into a mathematical representation of its meaning, searches a library of similarly-encoded web content, pulls the closest matches, and writes an answer from them. The matching step rewards how well the content answers the question, plus source authority and density. It does not strongly reward recency or virality. The average age of a Reddit post cited by AI is roughly 900 days, which is why treating trend-chasing as strategy is avoidance masquerading as effort.

Q: Should I post more on LinkedIn to get cited by ChatGPT and Perplexity? A: More is not the move. Substantive and consistent is the move. The Semrush data shows long-form articles between 500 and 2,000 words drive most LinkedIn-sourced citations, that publishing cadence matters more than follower count, and that reshares are rarely cited. If you are publishing real work consistently, LinkedIn is a useful distribution surface. If you are posting shallow content for the algorithm, it will not help you.

Q: Why don't AI search engines reward viral content? A: Because the retrieval layer is built to surface the canonical answer to a question, not the most popular post about it. Canonical answers tend to be older, more cross-referenced, denser, and lower-engagement-but-higher-information than viral posts. Up to 80 percent of Reddit threads cited by AI have fewer than 20 upvotes. The systems lean toward settled consensus over current attention.

Q: What is the long-term strategy for getting cited in AI search results? A: Pick three to five specific questions in your field that you are genuinely positioned to answer. Publish one substantive long-form piece a week answering one of them, on a domain you own, in your real voice. Cross-post with proper canonical attribution. Do not measure citations for the first six months and trust the compounding. Chasing the chart instead is avoidance masquerading as effort, and after two years it leaves you with no position to show.

So much respect.