TechCrunch covers 0.03% of the companies that pitch it. This is for the other 99.97%.

A founder we know refreshes her inbox four times a day. Not for customers. For a reply from a TechCrunch reporter she pitched nineteen days ago.

She did everything the guides tell you to do. Found the right beat. Read the reporter's last six months of stories. Framed the round around an unusual investor instead of the number. Kept it to four sentences. Followed up once, politely, on day ten.

Nothing. Not a no. Just silence, which in PR means no.

Here's the part that should make her feel better and doesn't: it was a good pitch. TechCrunch's own advice to founders is close to what she did - find a real story, not just a company update, and test it like a reporter would. She passed that test. She still didn't get in.

That's not a TechCrunch problem. That's arithmetic. The outlet has described receiving thousands upon thousands of submissions for a single pitch competition alone, before you even count the daily news tips. There is one front page. There are twenty-some slots on it a day. There are, conservatively, thousands of funded companies who'd like to be on one of those slots this week.

Somebody has to lose that lottery. Almost everybody loses that lottery.

The uncomfortable twistThe part nobody says out loud

Here's the twist, and it's a good one: TechCrunch coverage might matter more now than it did five years ago, not less. Research on AI citation patterns has found a strong bias toward earned media over anything a company publishes about itself, and TechCrunch sits near the top of that pile - reportedly among the most-cited editorial sources when AI systems answer questions about technology companies. A hit there doesn't just spike traffic for a day. It gets indexed, trusted, and pulled from for months.

So the prize is real. It's just that almost nobody gets to claim it, and the ones who do got there through a process that has nothing to do with whether AI needs to know about their company. It has to do with whether a specific person, on a specific day, with fourteen other emails in their queue, decided your round was the interesting one.

Think about what that means at scale. Roughly five million new businesses get started every year in the US alone. TechCrunch, at full output, runs a few thousand stories a year across every beat it covers - funding, product, policy, mobility, security, all of it combined. Even in the generous case where every single one of those stories covers a different company, that's coverage for a rounding error of the total.

Stretch that gap out over five or ten years and you get an answer engine with rich, frequently updated knowledge about a tiny sliver of the economy, and almost nothing structured or current about everyone else. Not because everyone else doesn't matter. Because nobody wrote it down anywhere a machine could read it.

The splitTwo totally different questions

Ask "will TechCrunch cover us" and you're asking an editorial question: is this interesting to a human, today, compared to everything else pitched today.

Ask "does AI understand what we do" and you're asking a completely different question: does enough structured, factual, freshly dated material about this company exist anywhere on the internet for a model to retrieve.

Those two questions used to be roughly the same question, because for twenty years, "does information about your company exist online" mostly meant "did a journalist write about you." Earned media wasn't just the best way to be known. It was close to the only way.

It isn't anymore, and the research is unusually blunt about it. Semrush's work on ChatGPT's citation behaviour found that around 90% of the pages it cites rank 21st or lower in classic Google search - pages the old system of authority had already buried. The thing that used to gatekeep visibility, a search ranking or an editor's inbox, doesn't gatekeep it the same way now. Ahrefs, analysing roughly 17 million AI citations, found the cited material runs meaningfully fresher than what organic search ranks, with about half of it published or updated inside the previous 13 weeks.

Freshness and structure are starting to outweigh scale and authority. That is a genuinely new rule, and most companies are still playing by the old one - stockpiling quarterly recaps, guarding a press page, treating publication as something that happens to them rather than something they do.

The analogyThe newspaper front page vs. the public records office

Here's the whole idea in one picture.

TechCrunch is a newspaper front page. Finite space. An editor decides who's on it. Getting on is a genuine achievement, and it should be - scarcity is what makes the placement valuable in the first place. But by design, almost everyone gets left off, and the ones who make it get one placement, on one day, that starts fading the moment the next day's front page gets made.

A newsroom is a public records office. Nobody has to decide you're interesting enough to have a file. You get one because you exist and you're doing things. It's a structured, always-on public record of your launches, your people, your product, built and published on a cadence rather than pitched story by story. The file just sits there, gets added to every week, and gets deeper instead of staler. Nobody's competing for a slot on it, because there isn't a slot. There's just your record, next to everyone else's record, all separately compounding.

The table to stealMoments vs. memory

Put this in your next pitch deck, your next board update, whatever you need it for.

The important row is the third one. A press hit is a moment. A public record is memory. Moments are great. They're just not the same asset, and companies have been pricing them as if they were.

The patternWhy this isn't a media story, it's an infrastructure story

Every time the way people find information has shifted, the businesses that treated "getting covered" as their whole strategy got caught flat-footed, and the ones who understood the shift as infrastructure rather than a media relations tactic compounded quietly in the background until it was obvious.

Blogs did it to print. A company blog could out-publish a print quarterly and rank for a thousand more search terms, for less money, forever, and the companies that got this early treated their blog as a permanent asset instead of a marketing expense. Most of their competitors kept buying full-page ads and wondering why the ad stopped working the day the budget ran out.

SEO did it to blogs next. Suddenly it wasn't enough to publish; you had to publish structured, keyword-mapped, internally linked content, or you were invisible on page one no matter how good the writing was. A whole industry of companies who'd never once thought of themselves as publishers had to become publishers, fast, or watch competitors who did it earlier eat the organic traffic they used to take for granted.

Both times, the pattern was identical. The shift looked, at first, like a media trend - something for the marketing team to keep an eye on. It turned out to be an infrastructure decision, and the companies that treated it that way, building a durable asset instead of chasing a tactic, were the ones still standing when the next shift came.

AI answer engines are the same shift again, one layer up. It's no longer only about getting in front of readers. It's about getting read by the thing that reads on behalf of readers. The winners of the next ten years won't be the companies with the best press page. They'll be the ones with the deepest, freshest, most interconnected public record - the ones an AI system can cite with confidence instead of hedging, guessing, or reaching for a competitor because that's the only name it has enough material on.

That's not a bet on TechCrunch getting worse. TechCrunch, if anything, might get more valuable as a scarce prize. It's a bet that scarcity was never the point. Coverage was always just a proxy for "does information about this company exist that a system can trust." Now there's a direct way to answer that question instead of a proxy for it.

The practical partWhat to actually do about it

Enter the lottery. Keep pitching TechCrunch, The Information, whoever's on your beat. When it lands, it's real, it's earned, and - per that same citation research - it's some of the highest-trust material an AI system will ever pull from about you. Don't stop.

But stop treating it as your only public record. Build the file that exists independent of whether an editor said yes this month.

• Write it the week it happens. Launches, hires, customer wins, an executive's actual expertise - dated when it occurred, not batched into a quarterly recap nobody reads twice.
• Structure it for a parser, not just a skimmer. A machine should be able to work out who you are, what you sell, and who you serve without inference.
• Interconnect it. The fifth story should make the first one worth more instead of competing with it for attention.
• Keep the cadence. Freshness is the variable the research keeps pointing at. A record that stops updating starts decaying immediately.
• Measure retrieval, not impressions. The useful question is where an AI system actually cited you last month, not how many people saw a headline.

This is the specific gap a YesPress Newsroom is built to close. Not a blog, not a press page - a branded, always-on newsroom at your own address, where the raw material you already have (a Slack update, a customer call, a founder's notes) gets run through real editorial judgment and published as structured, citable stories, live within 24 hours, with a monthly report on where AI actually cited you. It doesn't replace the TechCrunch pitch. It's what your record looks like on every one of the days a reporter doesn't write back.

Ask ChatGPT or Gemini about CRM, or cloud infrastructure, or enterprise software, and notice how the same handful of names come up every time, across completely different prompts. Nobody planned that. It's just what happens when a public record is deep enough, fresh enough, and cited often enough that a model stops needing to guess.

The category leaders in an AI-native economy won't necessarily be the biggest companies. They'll be the best-documented ones. And right now, "best-documented" is still wide open in most categories, because almost nobody has bothered to compete for it yet. That's a strange kind of advantage: cheap to build, compounding by design, and available to a company that's never once made a TechCrunch front page.

Our founder, still refreshing her inbox, should keep refreshing it. But nineteen days of silence from one reporter shouldn't be the only thing standing between her company and being understood. It never had to be. While she waits, her newsroom keeps running anyway - publishing, compounding, getting read by the thing that's actually going to introduce her to her next customer.

— End —