Stratechery
A paid daily tech and media analysis publication written by Ben Thompson.
Where Stratechery comes up1 episode or test
Ben Thompson built Stratechery as a one-person paid newsletter: $15 a month, 170 to 200 public pieces a year for about a decade, subscribers in over 100 countries, mostly annual with very low churn. He credits stories, daily volume and a recognizable site for beating equally smart imitators. Lesson: publish on a deadline, make the second piece as good as the first, and write what interests you.
- 1Publish free articlesCan reach a million new readers
- 2Win the second visitSecond article decides if they return
- 3Subscribe at $15/monthPaying for a steady stream
- 4Get daily paid updatesAssume knowledge, skip introductions
- 5Switch to annualMost subscribers annual, churn very low
From Ben Thompson interviews
What was said about Stratechery
1A one-person paid newsletter built on a niche
The internet lets a small niche pay a solo writer well. Thompson argues the internet rewards a few huge discovery platforms and countless small niche players, because zero marginal costs let a loyal niche support one writer. Stratechery has subscribers in over 100 countries.
Traditional media carries cost structures built for printing presses and delivery trucks, so it cannot shrink to serve a niche. Platforms that control discovery attract users, then supply, then more users, which leaves very few, very large winners.
Readers pay for steady delivery, so price the flow, not single pieces. Stratechery charges $15 a month for a steady stream of updates and an interview. Most subscribers are now annual, churn is very low, and the best cohort started monthly before switching to annual.
Thompson contrasts this with Dylan Patel, who charges $500 a year because he sells rare, high-value scoops.
Free articles find readers; paid updates serve the ones who stay. Free articles sit on the front page and can reach a million people who never read Stratechery, so they must explain more and open with a hook. Paid updates go to known subscribers, so they assume knowledge, skip introductions and launch from the nut of a Wall Street Journal or Bloomberg story.
Thompson says writing a lot is what makes this split possible: some pieces go free to find customers while others are sold to those who stay. A once-a-week writer has to trade acquisition against having something worth paying for.
Make your own pond instead of fighting the other fish. Thompson says defining the target audience is the most important thing online. Writing about the big five tech companies puts you against him and millions of free opinions on Twitter, while Neil Cybart owns Apple and Casey Newton owns social networks.
Online you compete with every writer in the world, so geography no longer protects anyone. "The key is making your own Pond."
2Winning the second visit
The second article a reader sees decides whether they come back. Thompson's thesis is that the most important article is the second one someone reads, because that is when they decide to expect quality. Stratechery grew by word of mouth, so each reader's second visit decided whether they returned.
"The most important article you write is the second article someone reads."
A distinctive look makes a returning visitor realize they've been there. Stratechery uses a custom font, an orange color and hand-drawn sketches so a reader arriving from a Twitter link notices they have visited before and builds the expectation that the site has good stuff.
Custom fonts were new on the web at the time, and the hard-to-pronounce name was memorable. He wanted readers to feel disappointed when something happened and he did not write about it.
3Daily deadlines and volume as strategy
A deadline every day beats waiting for inspiration. Thompson calls himself a procrastinator, so he publishes every single day, writing in the late afternoon once panic sets in. That yields about 170 to 200 public pieces a year, which he puts at 50 or 60 books over a decade.
After a weak piece he sleeps badly and says the day that stops is the day to quit. He credits volume and storytelling, not sharper insight, for outlasting early imitators who flamed out.
Readers reward the writer who shows up, so accept uneven quality. Thompson argues readers say they only want good work but reward daily presence. He accepts occasional weaker or overlong pieces, some around 5,000 words, as the cost of his model.
"I do think that volume or quantity is underrated."
Write about what interests you to avoid burnout. His one near burnout came in 2019 and 2020 from writing about antitrust, regulation and Congressional hearings, which he found soul-sucking. The hours were not the problem; the subject was.
"If I'm like trying to cater to my customers I think that's a recipe for Burnout."
Protect uninterrupted flow, even if it means a separate office. Thompson holds a piece's whole structure in his head like a programmer holds an application, so interruptions break it. While his kids were young he rented a separate workplace because they saw dad home and wanted to talk.
He sets alarms for everything and does no side conversations while working. He works at home again now that the kids are older.
4The framework machine behind daily analysis
- 1Feed in daily newsEach day's story enters the model
- 2Framework outputs conclusionHe writes down the steps
- 3Surprise reveals gapA missing piece in the model
- 4Gap becomes bigger articleNew piece updates the framework
From Ben Thompson interviews
Feed each day's news into one evolving model of the world. Thompson treats his framework as a machine: news goes in and a conclusion comes out, and he writes down the steps. When the output is surprising, a piece is missing, and those gaps become his bigger articles.
He says research-then-write could not produce thousands of pieces. He compares it to TSMC's library of standard chip blocks, where each piece adds the last 5% of special sauce.
Hold theories until a news event makes them timely. Purely theoretical pieces usually do not resonate, so Thompson keeps outlines, such as the relative risk of the hyperscalers from Meta to AWS, until a news trigger arrives.
Because insights are anchored in time and place, he finds a book or anthology hard: it would lock in views he may no longer hold.
Quote your own earlier work instead of rewriting it. Thompson treats Stratechery as a journal of his understanding, so he re-quotes old pieces, like his earlier point on Facebook's network constraints during its TikTok shift, and adds the correction.
People on Twitter mock him for quoting himself. He says it is how he writes thousands of pieces without writing a thousand new short stories.
Fill in the space between what companies say. Thompson says company statements amount to PowerPoints, and his value is stating what sits between the slides rather than repeating them or listing every possibility. He watches the first five minutes of a keynote because the chosen frame says more than the products.
He wants to equip readers rather than always be right, so a CEO with more information can use his structure and reach a different conclusion.
Guard a strong framework against confirmation bias. Thompson checks that a company's actual numbers and specs match its narrative and worries about looking only for evidence that fits his model. He accepts that overcompensating sometimes makes him wrong.
His first Stratechery article came from going back to the source: Christensen had the Mac versus Windows history backwards, since DOS and the IBM market predated the Mac. When he needs depth he reads a whole book that morning; one day on GPUs paid off years later with AI and Nvidia.
5How a Stratechery piece gets written
Open analysis with a story and spend your hours on the opening. Thompson's best pieces have a beginning, a path and a resolution. His Disney and Charter standoff piece opened with the first commercial satellite launch before moving into carriage disputes.
Much of a long article's time goes into finding the opening; once the first paragraphs work, the rest follows. His less interesting pieces are usually the ones where the opening never came, shipped anyway because they had to publish.
Plant a thread early and pay it off at the end. Strong updates set something up in part one and tie it up in parts two and three, as when a held-back Google capex point closed a Microsoft earnings piece. Readers should feel the 10 minutes were well spent.
To check endings, he pastes the introduction below the draft and asks how the conclusion ties back; if it does not, he adds guideposts or scraps the intro.
Start writing to find out what you think. When Thompson lacks an opening it is often because he does not yet know his position, so he writes until the gaps between ideas fill in, like Bezos on bullet slides.
He says writing also makes him a better podcaster because every claim must be correct and structured.
Read every piece aloud before publishing. Because his articles also go out as a podcast, Thompson reads them aloud first, catching stumbles, repeated words and typos. His typo rate dropped to about one a month.
Tight thinking up front means little other editing; he rarely rewrites after about 10 years on his subjects.
State views directly, then correct mistakes openly and fast. Thompson writes in an A because of B because of C pattern and says hedging makes analysis uninteresting. When wrong, he follows John Gruber's rule of correcting immediately and walks readers through the step he got wrong.
He calls himself highly disagreeable and says errors are inevitable when writing daily. "The best way to always be right is to immediately correct yourself when you're wrong."
How He Started the Paid Newsletter Trend — Ben Thompson
Thompson's one-person paid newsletter has subscribers in over 100 countries, charges $15 a month, and publishes 170 to 200 public pieces a year. Most subscribers are now annual, churn is very low, and the best cohort started monthly before switching to annual. A custom font, orange color and hand-drawn sketches make the site recognizable on a second visit.
The internet lets a small niche pay a solo writer well. Thompson argues the internet rewards a few huge discovery platforms and countless small niche players, because zero marginal costs let a loyal niche support one writer. Stratechery has subscribers in over 100 countries.
Traditional media carries cost structures built for printing presses and delivery trucks, so it cannot shrink to serve a niche. Platforms that control discovery attract users, then supply, then more users, which leaves very few, very large winners.
Readers pay for steady delivery, so price the flow, not single pieces. Stratechery charges $15 a month for a steady stream of updates and an interview. Most subscribers are now annual, churn is very low, and the best cohort started monthly before switching to annual.
Thompson contrasts this with Dylan Patel, who charges $500 a year because he sells rare, high-value scoops.
Free articles find readers; paid updates serve the ones who stay. Free articles sit on the front page and can reach a million people who never read Stratechery, so they must explain more and open with a hook. Paid updates go to known subscribers, so they assume knowledge, skip introductions and launch from the nut of a Wall Street Journal or Bloomberg story.
Thompson says writing a lot is what makes this split possible: some pieces go free to find customers while others are sold to those who stay. A once-a-week writer has to trade acquisition against having something worth paying for.
Make your own pond instead of fighting the other fish. Thompson says defining the target audience is the most important thing online. Writing about the big five tech companies puts you against him and millions of free opinions on Twitter, while Neil Cybart owns Apple and Casey Newton owns social networks.
Online you compete with every writer in the world, so geography no longer protects anyone. "The key is making your own Pond."
The second article a reader sees decides whether they come back. Thompson's thesis is that the most important article is the second one someone reads, because that is when they decide to expect quality. Stratechery grew by word of mouth, so each reader's second visit decided whether they returned.
"The most important article you write is the second article someone reads."
A distinctive look makes a returning visitor realize they've been there. Stratechery uses a custom font, an orange color and hand-drawn sketches so a reader arriving from a Twitter link notices they have visited before and builds the expectation that the site has good stuff.
Custom fonts were new on the web at the time, and the hard-to-pronounce name was memorable. He wanted readers to feel disappointed when something happened and he did not write about it.
A deadline every day beats waiting for inspiration. Thompson calls himself a procrastinator, so he publishes every single day, writing in the late afternoon once panic sets in. That yields about 170 to 200 public pieces a year, which he puts at 50 or 60 books over a decade.
After a weak piece he sleeps badly and says the day that stops is the day to quit. He credits volume and storytelling, not sharper insight, for outlasting early imitators who flamed out.
Readers reward the writer who shows up, so accept uneven quality. Thompson argues readers say they only want good work but reward daily presence. He accepts occasional weaker or overlong pieces, some around 5,000 words, as the cost of his model.
"I do think that volume or quantity is underrated."
Write about what interests you to avoid burnout. His one near burnout came in 2019 and 2020 from writing about antitrust, regulation and Congressional hearings, which he found soul-sucking. The hours were not the problem; the subject was.
"If I'm like trying to cater to my customers I think that's a recipe for Burnout."
Protect uninterrupted flow, even if it means a separate office. Thompson holds a piece's whole structure in his head like a programmer holds an application, so interruptions break it. While his kids were young he rented a separate workplace because they saw dad home and wanted to talk.
He sets alarms for everything and does no side conversations while working. He works at home again now that the kids are older.
Feed each day's news into one evolving model of the world. Thompson treats his framework as a machine: news goes in and a conclusion comes out, and he writes down the steps. When the output is surprising, a piece is missing, and those gaps become his bigger articles.
He says research-then-write could not produce thousands of pieces. He compares it to TSMC's library of standard chip blocks, where each piece adds the last 5% of special sauce.
Hold theories until a news event makes them timely. Purely theoretical pieces usually do not resonate, so Thompson keeps outlines, such as the relative risk of the hyperscalers from Meta to AWS, until a news trigger arrives.
Because insights are anchored in time and place, he finds a book or anthology hard: it would lock in views he may no longer hold.
Quote your own earlier work instead of rewriting it. Thompson treats Stratechery as a journal of his understanding, so he re-quotes old pieces, like his earlier point on Facebook's network constraints during its TikTok shift, and adds the correction.
People on Twitter mock him for quoting himself. He says it is how he writes thousands of pieces without writing a thousand new short stories.
Fill in the space between what companies say. Thompson says company statements amount to PowerPoints, and his value is stating what sits between the slides rather than repeating them or listing every possibility. He watches the first five minutes of a keynote because the chosen frame says more than the products.
He wants to equip readers rather than always be right, so a CEO with more information can use his structure and reach a different conclusion.
Guard a strong framework against confirmation bias. Thompson checks that a company's actual numbers and specs match its narrative and worries about looking only for evidence that fits his model. He accepts that overcompensating sometimes makes him wrong.
His first Stratechery article came from going back to the source: Christensen had the Mac versus Windows history backwards, since DOS and the IBM market predated the Mac. When he needs depth he reads a whole book that morning; one day on GPUs paid off years later with AI and Nvidia.
Open analysis with a story and spend your hours on the opening. Thompson's best pieces have a beginning, a path and a resolution. His Disney and Charter standoff piece opened with the first commercial satellite launch before moving into carriage disputes.
Much of a long article's time goes into finding the opening; once the first paragraphs work, the rest follows. His less interesting pieces are usually the ones where the opening never came, shipped anyway because they had to publish.
Plant a thread early and pay it off at the end. Strong updates set something up in part one and tie it up in parts two and three, as when a held-back Google capex point closed a Microsoft earnings piece. Readers should feel the 10 minutes were well spent.
To check endings, he pastes the introduction below the draft and asks how the conclusion ties back; if it does not, he adds guideposts or scraps the intro.
Start writing to find out what you think. When Thompson lacks an opening it is often because he does not yet know his position, so he writes until the gaps between ideas fill in, like Bezos on bullet slides.
He says writing also makes him a better podcaster because every claim must be correct and structured.
Read every piece aloud before publishing. Because his articles also go out as a podcast, Thompson reads them aloud first, catching stumbles, repeated words and typos. His typo rate dropped to about one a month.
Tight thinking up front means little other editing; he rarely rewrites after about 10 years on his subjects.
State views directly, then correct mistakes openly and fast. Thompson writes in an A because of B because of C pattern and says hedging makes analysis uninteresting. When wrong, he follows John Gruber's rule of correcting immediately and walks readers through the step he got wrong.
He calls himself highly disagreeable and says errors are inevitable when writing daily. "The best way to always be right is to immediately correct yourself when you're wrong."
More in News & Media
News & MediaSubstack
A platform where writers publish newsletters and charge readers for subscriptions.
22 sources · 36 tipsAppAdvice
An app discovery and news site that features app deals and promotions.
16 sources · 26 tips
News & MediaThe New York Times
A major newspaper that rebuilt a digital bundle with games, cooking and Wirecutter.
10 sources · 29 tips
News & MediaMarginal Revolution
Tyler Cowen's economics blog, written daily for over two decades.
3 sources · 18 tips
News & MediaTechCrunch
A technology news publication known for startup and funding coverage.
8 sources · 14 tips
News & MediaThe New Yorker
A weekly magazine of essays, reporting, fiction and humor.
7 sources · 15 tips
News & MediaZinio
A digital newsstand app offering over 6,000 magazines on web, iOS and Android.
4 sources · 13 tips
News & MediaMark Manson
A self-help author and blogger known for The Subtle Art of Not Giving a F*ck.
2 sources · 19 tips