What I Learned From Publishing Thousands Of AEO Articles

Somewhere north of a few thousand articles have now gone through the pipeline I work on, ours and our customers', across more categories than I can keep in my head. That's enough volume to be embarrassed by early opinions, and this essay is the honest ledger: what held up, what evaporated, and what I'd tell the version of me who thought answer engine optimization was mostly about writing more articles. Spoiler on that last one: he was cheerfully funding a very organized, beautifully formatted graveyard.

Lesson 1: The Plan Was Always the Product

Here's the thing nobody wants to hear because it's boring: after the first few hundred articles, writing quality stopped being the differentiator. A competent pipeline produces competent drafts all day. What separated the libraries that earned citations from the ones that rotted was decided before any writing happened, in whether topics came from real buyer questions, grouped into subjects a site could actually own, published in an order that built a center instead of a sprawl, and interlinked by design rather than by memory. I can now look at a content plan and predict the library's fate with uncomfortable accuracy, and the pre-writing checklist we published is basically that prediction formalized. The plan was the asset all along, and the articles were how you cashed it.

Lesson 2: Mentions and Citations Are Different Animals

We learned this one by almost getting it wrong. Running a 90-day file-on-trial experiment, we nearly reported a win because a blended visibility score drifted up, and only splitting brand mentions from actual citations showed the truth: mentions up a touch, citations dead flat. They move independently, they mean different things, an engine naming you is reputation while an engine citing your URL is your page winning a slot, and any tool or report that blends them will eventually tell you a flattering lie. Split them forever. This is the single measurement habit I'd tattoo somewhere visible.

Lesson 3: Every Engine Is Its Own Country

I assumed the engines were dialects of one language. They're different countries with different customs. When we tagged 500 Perplexity citations, the domain overlap with ChatGPT's picks for equivalent queries was around 11%, and the mega-audits say most cited URLs appear in exactly one engine's citations. Perplexity worships freshness and communities. ChatGPT, since August, trusts first and searches second, and loves documentation. Google's AI features run on Google's machinery. Gemini likes your own site. One playbook spread evenly across that map under-serves every country at once, which is why per-engine columns stopped being a nice-to-have in our reporting and became the reporting.

RANKCONTROL

Know exactly what AI says about your competitors.

RankControl's Recon Agent monitors competitor citations across ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode. See where they show up and you don't.

See how it works→

Lesson 4: August Happens, and August Will Happen Again

The week GPT-5.6 reshuffled ChatGPT's sources, community citations collapsed something like 86% in days, documentation jumped from a rounding error to roughly a third of citations, and every strategy built on the old mix aged a year in a week. Watching that on live tracking taught me something. Actually, scratch the specifics, because the real lesson has nothing to do with that particular reshuffle: citation share is a policy, policies change with versions, and the only durable advantage is the instrument. Teams with baselines knew what broke by Friday. Teams without them found out a quarter later, in pipeline numbers nobody could explain. There will be another August, and I genuinely can't tell you what it will break, only that the tracking will see it first.

Lesson 5: Half Your Citations Won't Send Anyone

The encyclopedia problem, and it stung when I first understood it. A large share of hard-won citations function as fact confirmations inside answers, the user reads the answer, nobody clicks, and one team's audit that matched our observations found roughly three quarters of cited pages sending zero visits. I now think of citations the way I think of impressions, with position doing the multiplying, and I build page portfolios accordingly: encyclopedia pages on purpose, doing the fact-delivery and trust work with no traffic expectations, and destination pages the answers can't finish, calculators, living data, templates, decision-stage depth, doing the converting. The only mistake was ever grading them all on the same metric.

Interlude: The Receipts

Opinions are cheap and I'm asking you to trust ten of them, so here are the numbers these lessons rest on, each one either from our own tracking and experiments or from the public audits our samples kept agreeing with.

None of these came from a lab. They came from watching real answers, weekly, across real categories, which is itself the meta-receipt for lesson four.

Three Articles I Regret, As Archetypes

Every large library carries its embarrassments, and ours cluster into three shapes worth confessing because they're the three shapes I now see everywhere.

The trend-chaser. There's a version of an article in our network that sprinted onto a hot model-release rumor, ranked for a week, and then aged like milk when the rumor resolved differently. It taught me the difference between news coverage with a durable angle and news coverage that's just being early, loudly. We still write news; we now ask what the piece teaches after the news cools, and if the answer is nothing, it doesn't run.

The eleventh listicle. A perfectly optimized roundup on a query where ten roundups already existed, saying what they said, ranking behind them forever. Nobody read it, no engine cited it, and it did its real damage upstream, occupying a calendar slot that a deeper page deserved. The same-ten-pages test exists because of articles like this one.

The clever one. An article built around a tactic that toured the industry for a month, written fast to catch the wave, technically accurate, and worthless within a quarter when the tactic died. Its ghost is why lesson eight exists. The boring fundamentals never made me write a retraction.

What Scale Actually Bought

Since lesson ten says start smaller, honesty requires the other side: what the volume genuinely purchased. Pattern visibility, mostly. You can't see an 11% overlap or a 73% zero-click share in a twenty-article library; the sample is too small and every observation is an anecdote. Thousands of articles across many categories turned anecdotes into distributions, and the distributions are where every lesson above came from. Scale also stress-tested the guardrails into existence, because every failure mode you can imagine shows up by article five hundred, and a few you can't. And it bought cluster completion speed, whole subjects owned in months, which is the one advantage of volume that compounds cleanly when the planning is right.

So the honest synthesis: scale is a wonderful instrument and a mediocre strategy. Use volume to learn faster, never to substitute for coherence, and if you have to pick one, pick coherence, because the public audits now publish the distributions scale used to be the only way to see.

Lesson 6: Freshness Is a Subscription

I used to think of publishing as shipping. Ship the article, move on. The citation data beat this out of me, especially on Perplexity, where half of citations point at current-year content and a stale date is a quiet disqualification. Answer engines treat your library like a subscription they're deciding whether to keep paying attention to, and visible maintenance, real updates when facts change, dates that tell the truth, is part of the product. The operational consequence at scale: refresh capacity has to be budgeted like new-article capacity, permanently, or the back catalog becomes an archive that answers stopped visiting.

Lesson 7: The Docs Kept Beating the Blog

Painful for someone who runs a content engine: across the August reshuffle and after, documentation-shaped pages kept out-earning blog-shaped pages for citation slots, sometimes embarrassingly. Factual, structured, unglamorous pages read as trustworthy to verification-flavored retrieval, and the marketing-voiced article next to them reads as persuasion. The lesson turned out to be letting the docs energy into the blog rather than blogging less: opening answers, verifiable specifics, tables that state facts, FAQ blocks phrased like real questions. The most-cited articles in our whole network are the ones a stranger might mistake for documentation with a personality.

RANKCONTROL

26 content formats. Published on your domain. Matched to your brand.

Guides, comparisons, listicles, case studies, and more. RankControl generates content that gets cited by ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode.

Start free→

Lesson 8: The Boring Fixes Beat the Clever Ones, Every Quarter

Tally the citation gains across everything we've shipped and audited, and the leaderboard is humiliating for cleverness. Unblocking crawlers a CDN was silently challenging. Reconciling entity facts across profiles nobody had read since 2023. Restructuring first screens so pages answer before they warm up. Making docs public. Meanwhile the clever things, the schema exotica, the courtesy files, the submit-your-site directories, the prompt-shaped hacks that tour LinkedIn monthly, produced approximately nothing measurable, approximately every time. I keep a whole essay's worth of this in the strategies guide, but the compressed version fits on an index card: access, structure, truth, mentions, measured weekly.

Lesson 9: Guardrails Made the Volume Worth Having

Publishing thousands of anything machine-assisted without guardrails produces a very scalable embarrassment. The validator we run, no em dashes, no cursed vocabulary, no fake experience claims, direct answers up top, real dates, originality checks, started as hygiene and became strategy, because the patterns it blocks are exactly what readers and engines learned to discount. The uncomfortable insight: at scale, your guardrails are your voice. Sound reductive? Watch what an unchecked pipeline converges on by article eighty. Whatever you don't check for, you will eventually publish, and the library converges on whatever the checks tolerate.

The Questions I Still Can't Answer

Ten lessons implies more certainty than I own, so here's the open ledger too. I can't yet measure influence without clicks properly: when an answer describes a brand correctly to a buyer who never visits, something valuable happened, and attribution science for it doesn't exist yet. I don't know how agent browsing reshapes this, since agents reading sites on a buyer's behalf look human in every log we have. I don't know whether citations keep concentrating toward fewer, bigger sources or whether the engines re-broaden. And I genuinely don't know what the next August breaks, which is the most honest sentence in this essay and the entire reason the weekly instrument exists.

Lesson 10: I'd Start Smaller and Deeper

The one I'd actually send back in time. Starting today, I'd publish half as many articles into twice the depth: fewer subjects, owned completely, definition pages first, supporting questions in order, refreshed forever, with measurement running from day zero so every fortnight teaches something. The thousands taught me that the compounding came from coherence rather than count, and coherence is available at any size, which is the genuinely good news for every team that can't publish thousands and shouldn't try.

That's the ledger. If you want the operational version of these lessons, the pipeline, the guardrails, and the per-engine scoreboard run as one system, that's the product we ended up building, largely because we needed it first. And if you're building your own version manually, take lesson 10 and lesson 2 and go: small, deep, split your mentions from your citations, and let the instrument tell you what your market's engines actually reward. The rest of the lessons will find you on schedule.

Your competitors are getting cited by AI. You're not.

Every day without citation tracking is a day your competitors pull ahead in ChatGPT, Perplexity, and Claude.

Show me who's getting cited→2-minute overview · real case-study numbers