Marketing School - Digital Marketing and Online Marketing Tips

An internet marketer cut his team from a couple hundred people to under 100 and credits AI and OpenClaw with the work. Eric Siu and Neil Patel explain why AI cannot handle marketing work alone.Then Neil shares why he walked away from buying a European agency that is cutting staff to run on AI. Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/Want to recruit great marketers? Find them here: https://marketingschool.io/hire Eric and Neil start with a well-known internet marketer who cut his team from a couple hundred people to under 100 and says AI and OpenClaw now do 20 to 30 times more work. Eric says that claim hurts the marketer's credibility, and that AI should elevate capabilities rather than eliminate jobs. Neil describes a call with a European e-commerce agency he decided not to acquire, because it planned to shrink to eight to ten people, said two or three people could run its ad spend, and had lost three of its four biggest clients. Eric lists where AI output still needs a human editor, including his YouTube channel, ad creative, AEO and SEO content, and X posts. They also cover AI metrics that do not match business results, and how to handle client expectations and feedback. Human judgment still decides marketing results. Key takeaways◾Cutting staff and crediting AI for it can hurt a marketer's credibility◾AI output still needs human taste and an editor in the loop◾Optimize for business results and client value, not AI-suggested views Chapters00:00 Internet marketer says AI replaced his staff00:58 Does cutting jobs hurt credibility?01:46 OpenClaw is unreliable03:16 Marketers who talk about AI without knowing it03:45 Neil's call about acquiring an agency04:33 The agency plans to shrink to eight or ten people06:22 Why the agency is losing clients08:11 Up or out, and selling a company09:08 AI cannot handle marketing work alone10:20 How Eric's view of AI has shifted11:37 AI feedback creates a bias toward views12:58 Views versus revenue14:13 Clients pay for results, not effort15:34 Optimizing for revenue, not brand16:39 Two types of consultants17:52 Why feedback is a gift 𝗔𝗕𝗢𝗨𝗧 𝗧𝗛𝗘 𝗖𝗛𝗔𝗡𝗡𝗘𝗟Welcome to Marketing School, one of the top business podcasts with over 61 million downloads. Each episode delivers actionable marketing tips and strategies from two entrepreneurs who truly practice what they preach. The show is hosted by Eric Siu, founder of Leveling Up and Single Grain, and Neil Patel, co-founder of Neil Patel Digital and recognized by Forbes as a Top 10 Marketer. 🎙️ Learn More About the Hosts Eric Siu – Leveling Up: https://www.youtube.com/@LevelingUpOfficial Neil Patel: https://www.youtube.com/@neilpatel 📩 Free Resources Ubersuggest: https://www.ubersuggest.com/ Answer The Public: https://www.answerthepublic.com/ ✅ Subscribe for Daily Marketing Insights!

An internet marketer cut his team from a couple hundred people to under 100 and credits AI and OpenClaw. Eric Siu and Neil Patel explain why AI cannot handle marketing work alone, and why clients pay for results, not effort. They cover a $17K a month content agency, Gary Vee on the monetization of random content, and how Meta Muse recommended a client's product, which is why getting recommended by AI is the new AEO. They also get into paid marketplaces inside AI platforms, Zuckerberg's all-in data center bet, marketing to robots and agents, Forrester's prediction that agents will influence 75% of B2B buying, and Marc Andreessen on the AI-native gap. They close on hiring people who are AI native and why the biggest AI ROI is getting recommended. Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/Want to recruit great marketers? Find them here: https://marketingschool.io/hire Key takeaways◾AI cannot handle marketing work alone, human judgment still matters◾Getting recommended by AI is the new AEO and the biggest ROI◾Hire for AI fluency, and spend on AI where it makes money Chapters00:00 Internet marketer says AI replaced his staff01:46 OpenClaw is unreliable03:45 Neil's call about acquiring an agency06:22 Why the agency is losing clients08:47 AI cannot handle marketing work alone11:21 AI feedback creates a bias toward views13:57 Clients pay for results, not effort16:23 Two types of consultants18:38 A $17K a month helper asks for more21:56 What the helper's team takes home24:32 Under $100K a year for years25:59 Equity upside at fast-growing companies28:16 An agency that kept growing for decades31:25 No right or wrong way to build a business32:24 Gary Vee: the monetization of random content35:11 How algorithms find interested viewers38:25 Being filmed at Web Summit40:58 A bakery's authentic viral post41:53 Meta Muse recommends a client's product43:38 A water softener search story45:30 The cost of not optimizing for new channels47:59 Paid marketplaces inside AI platforms50:07 Get into new marketplaces and plugins early51:24 Zuckerberg's all-in data center bet54:41 Marketing to robots and agents56:46 AI build-out as a share of GDP58:33 Forrester: agents will influence 75% of B2B buying1:00:06 Do you still need an assistant?1:02:00 What makes a good assistant1:05:48 Marc Andreessen on the AI-native gap1:09:03 Hiring people who are AI native1:10:57 Eric wants both skills in one hire1:13:36 Did Eric ever want to replace all humans?1:16:52 The biggest AI ROI is getting recommended1:18:13 What a low AI spend can still mean 𝗔𝗕𝗢𝗨𝗧 𝗧𝗛𝗘 𝗖𝗛𝗔𝗡𝗡𝗘𝗟Welcome to Marketing School, one of the top business podcasts with over 61 million downloads. Each episode delivers actionable marketing tips and strategies from two entrepreneurs who truly practice what they preach. The show is hosted by Eric Siu, founder of Leveling Up and Single Grain, and Neil Patel, co-founder of Neil Patel Digital and recognized by Forbes as a Top 10 Marketer. 🎙️ Learn More About the Hosts Eric Siu – Leveling Up: https://www.youtube.com/@LevelingUpOfficial Neil Patel: https://www.youtube.com/@neilpatel 📩 Free Resources Ubersuggest: https://www.ubersuggest.com/ Answer The Public: https://www.answerthepublic.com/ ✅ Subscribe for Daily Marketing Insights!

Eric and Neil say the rule of 7 has grown toward the rule of 28, and that buyers touch a brand 8 times for a sub-$100 B2C buy and 23 times at $10K-plus. A gelato shop stranger who knew them from different places makes the case that you never know where people will find you. They cover why YouTube drives leads that nobody can attribute, then Meta's $14.9B deal for 49% of Scale AI and the 200-person AI team that Alexander Wang, the former Scale AI CEO, leads. Zuckerberg's line that talent density beats big teams for AI is tested against Meta's 70,000 staff. Neil explains his private equity mindset and his debt at SOFR plus 2%, and they close on the failures that come with big bets. Attention, talent and the cost of big bets. Key takeaways ◾Buyers touch a brand 8 times for a cheap B2C buy and 23 times at $10K-plus ◾Meta built a 200-person AI team out of 70,000 staff, which Zuckerberg calls talent density ◾Meta paid $14.9B for 49% of Scale AI, and Eric says the bet looks like it is paying off Chapters 00:00 Rule of 7 is now the rule of 28 00:44 Touches before buying: 8 cheap, 23 at $10K+ 01:34 A gelato shop stranger, and where people find you 03:53 YouTube drives leads you cannot attribute 04:27 Platforms stop sharing data 06:23 Meta's 200-person AI team, talent over size 07:36 Neil's private equity mindset 09:18 Zuckerberg asks peers what they would do 10:44 Neil's debt: SOFR plus 2% 13:14 Pain now for a bigger company later 14:17 Failures: rockets, timelines and the metaverse Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/

Eric and Neil open with Tesla's Roadster flying car, which they treat as a marketing stunt that keeps Tesla in the conversation. The AI fluency gap follows, with Karpathy's funnel, where most people barely touch an LLM, and agents that still misfire. Next is X Money and the newsletter debate, then a blog post test that a client took for human writing. The back half covers the rule of 28, Meta's 200-person AI team inside a 70,000 staff company, and why YouTube leads are hard to attribute. The show ends on Accenture's $29 monthly AI cap, Ramp's AI spend index and Bending Spoons' cuts at Evernote. Attention, trust and spend with a reason. Key takeaways ◾Accenture's $29 AI cap is smart, but proven high spenders should get exceptions ◾Meta built a 200-person AI team out of 70,000 staff, which Zuckerberg calls talent density ◾Tesla's flying Roadster is a stunt, and agents still need a human check Chapters 00:00 Tesla's flying car stunt 04:56 Is the Roadster $50,000 or $250,000? 09:48 Why Porsche lost its edge 12:22 Agents, not models 13:58 Agents still unreliable in marketing 17:19 Karpathy's AI-gap funnel 20:00 Agent slop and the cost of rework 21:00 Workflows as a service 22:08 X Money: 6% interest 26:34 The AI blog post test 32:08 Potato, SpaceX and GrokBot 32:41 Writing is thinking 36:34 Which channels are S-tier 37:55 Are newsletters dead? 41:24 Rule of 7 is now the rule of 28 45:17 YouTube drives leads you cannot attribute 47:47 Meta's 200-person AI team, talent over size 49:00 Neil's private equity mindset 52:08 Neil's debt: SOFR plus 2% 55:41 Failures: rockets, timelines and the metaverse 56:42 Accenture caps AI spend at $29 a month 1:00:14 Ramp's AI spend index: median is $11 1:02:09 Neil downgrades his ChatGPT plan 1:03:03 A 45-minute AI session that is worth it 1:04:49 Two-slice teams, not two-pizza teams 1:09:02 Bending Spoons: a private equity model 1:11:47 Evernote: 433 staff at peak, about 30 now Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/

Eric built thumbnail, editing and posting skills, and his long-form YouTube videos went from about 500 views to 15,000 to 25,000. Some of the team wouldn't use them, so they had to let people go, and the one person who stepped up is now fixing the Marketing School channel's trends, thumbnails and 30-second retention. Neil says AI optimizes for average, which still beats a below-average team, but you need humans in the loop. They cover Bain's 94% stat, why Neil would rather hire great marketers and teach them AI, and a senior SEO candidate who read straight off a Claude template. Then Neil's Zappos story, how Uber outspent Lyft, why AI rewards whoever raises billions, and Neil's Muse outreach to M&A advisors. An honest episode on AI skills and hiring. Key takeaways ◾Skills only work when someone spends real time tuning them ◾Hire people who stay curious and put their best foot forward ◾Small touches like Zappos' overnight shoes build loyalty Chapters 00:00 We got more views by firing 00:51 Skills took Eric's videos from 500 views to 25K 02:13 Cover the trend first, fix thumbnails and retention 02:56 Neil: AI optimizes for average 04:01 You still need humans in the loop 04:46 Bain's 94% and the coming job displacement 05:30 How we hire marketers now 06:36 A senior SEO candidate read off a Claude template 08:38 Lazy before the contract means worse after 09:46 The agency hire who said they're too slow 10:31 Hire people who never stop learning 11:12 Small touches build brands 12:38 Fundamentals move metrics more than campaigns 13:48 Uber won by subsidizing more than Lyft 15:08 Why a cheap Waymo rival can't win 16:10 AI rewards whoever raises billions 17:11 Muse is taking the market 17:58 Neil's Muse outreach lands a UK call 18:51 Red flags in the deals bankers pitch Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/

It's Q4, so Eric and Neil run through their business and marketing highlights and lowlights for the year. Eric's win was hiring AI-native people after learning he couldn't train everyone, and Neil's was the lowest churn in a long time. Neil explains bankruptcy clawbacks, where the lawyers always win, with FTX legal fees near $700M, and how tariffs and war cut client budgets. Eric's team now makes and manages bots, including one that shortlists recruits every day. On marketing, YouTube is bringing enterprise leads, Neil found out he did 12.7M views in 30 days, and Gavin Baker reports 80% of token spend has flipped to open-weight models. Eric covers outbound, wedges and Higgsfield's weekly launches, and Neil explains why he wants the chill life. A candid look back at the year. Key takeaways ◾Hire adaptable, AI-native people instead of trying to train everyone ◾Let bots handle evaluation work like recruiting shortlists ◾Test lots of ideas cheaply, then back the ones with traction Chapters 00:00 Q4 business highlights 00:44 Eric's lowlight: thinking he could train people 01:28 Neil's lowest churn in years 02:32 Bankruptcy clawbacks and legal drama 03:29 Lawyers always win: FTX's $700M in fees 04:43 Tariffs, war and budgets 06:20 Optimistic about 2027? 07:05 Payment plans for groceries 08:31 We make and manage bots 09:13 Recruiters who say they can't build it 10:00 Marketing highlights 10:47 Enterprise leads who found us on YouTube 11:32 Ad Age's performance marketing agency of the year 13:33 Neil's 12.7M views in 30 days 14:38 A slow speaking summer, two months at home 15:37 Neil only buys companies at a good deal 16:21 Best returns with the least stress 17:14 Outbound, wedges and 15-minute standups 18:39 Skillsdojo: pair engineers with marketers 19:43 Crank out 100 ideas, double down on winners 20:48 Gavin Baker: 80% of token spend flipped 21:46 Higgsfield ships a new feature every week 22:28 Neil does better in services than tech 23:42 Think like a capital allocator 25:08 Neil just wants the chill life Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/

Why versus pages are your secret SEO weapon. Noah Kagan shipped 19 SEO pages from a $20 Mac app and got six sales in three days, and Eric breaks down the two page types every product needs: versus pages and best-for guides. Neil adds two fixes, getting other sites to publish them and comparing the brands ChatGPT and Claude already recommend, and says comparison pages drive more revenue than listicles. Then Instinct: 10% daily growth, 40% of users adding a card, and why Neil would sell it to Apple. Sam Lessin says there is no moat in AI agents, Neil's Muse test turns hundreds of drafts into three sent emails, and Outrank's data shows sites with 400+ links got 115% more Google clicks. A practical episode on SEO and AI agents. Key takeaways ◾Build versus and alternative pages for the competitors AI already cites ◾AI agents have little loyalty, so the product has to keep earning it ◾Linkable assets still move Google clicks Chapters 00:00 Noah Kagan's 19 SEO pages and 6 sales 01:16 Neil: get other sites to publish your pages 02:38 Listicles are the new links 03:39 Comparison pages drive more revenue than listicles 04:02 Instinct: 10% a day, 40% add a card 05:21 Neil: Instinct should sell to Apple 06:37 Eric bets the founder sells 07:39 Apple could build this itself 08:24 Sam Lessin: there is no moat in AI agents 10:02 Neil's Muse test: hundreds of drafts, 3 sent 13:47 Every's personal agent comparison 14:24 Jev as a classifier for a fraction of a penny 16:36 Nobody is loyal to an AI product 17:37 Outrank: 400+ links, 115% more Google clicks 19:48 Neil: building something good? Ignore the hate 20:58 Linkable assets beat bought links Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/

AI labs are paying millions of dollars for company data through third parties, so should you sell yours? Neil says it depends on the business model, and Eric warns the labs will use that data to compete with you. Then a Bain study: only 6% of marketers say AI has had a significant impact on performance. Eric explains why that matches the top slice of any company, Neil shares the three places AI is actually winning for clients, and an a16z chart shows just 2.2% of households pay for AI. They explain why they stopped hiring clippers after YouTube's originality update, why niche videos bring qualified leads, and how Neil now picks topics from his sales pipeline. An honest episode on data and content. Key takeaways ◾Do not sell data that lets the labs compete with you ◾AI only pays when it is aimed at work that moves revenue ◾Original, niche content beats clipped reach Chapters 00:00 Should you sell your data to the AI labs? 02:06 Data is the moat as models converge 02:51 Eric's YouTube videos are driving leads 03:45 Bain: only 6% of marketers see AI impact 05:13 Neil asks: how many of you are growing? 06:10 Neil's top 3 AI wins in marketing 07:29 AI amplifies effort, and only 2.2% of homes pay 08:59 Why we stopped hiring clippers 09:58 YouTube's originality update for Shorts 11:54 Neil's team wanted 10 clip channels 12:58 Eric no longer needs an editor or a poster 13:32 Niche content brings qualified leads 15:27 Neil's channel: topics from the sales pipeline 16:52 Long form once a month, shorts twice Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/

Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/Want to recruit great marketers? Find them here: https://marketingschool.io/hire Eric shares how Cody Schneider works as an AI-native marketer, and Neil talks about his interview on Cole Gordon's show and why research is what separates a good interview. Then Neil says out loud that his personal brand is dying, judged by fewer photo requests at events, and Eric explains it with the four burners theory and the shift from social media to interest media. They talk about the content an experienced operator should make, why Neil keeps playing the game, and Eric's travel stats. The close covers why brand takes ten years or more, YouTube's coming thumbnail tests by audience segment, and Clay's CEO on why brand has to start with the founder, not the CMO. An honest episode about brand, time and attention. Key takeaways◾Interest media rewards the post, not the follower count◾A real brand takes ten years or more to build◾Brand starts with the founder; a CMO cannot install it Chapters00:00 The AI-native marketer: Cody Schneider01:36 Neil on Cole Gordon's podcast04:24 Research is what makes an interview good04:54 'My personal brand is dying'06:35 The four burners theory07:19 Social media became interest media09:03 What Neil's team does for him now12:49 Addicted to the game13:27 Eric's travel stats16:36 Which speaking gigs are worth it17:54 Brand takes 10+ years20:03 SEO has always evolved20:24 YouTube's big changes: thumbnails by segment22:11 Clay: brand starts with the founder

Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/Want to recruit great marketers? Find them here: https://marketingschool.io/hire Eric brings Peter Diamandis's take that Nvidia's greatest moat, CUDA, is leaking now that AI can port code to any chip, and why Jensen bought Mellanox years ago to own the interconnect instead. They debate the three moats that survive: speed of reinvention, capital velocity and brand. Neil keeps brand, drops capital after Harry Stebbings' three rounds in three weeks, and adds talent density, showing six of the top ten companies come from the Bay Area. Then Meta Muse: the $14B Scale bet Neil admits he would never have made, Eric's bot Doodoo picking his wardrobe, and Muse plus Jev sorting 116,000 followers for $3.30. It ends on the personality test that explains the show. A sharp episode on what still protects a business. Key takeaways◾Your moat lasts until the next one; speed of reinvention is the durable edge◾AI shoppers still buy brands, because people tell them what status looks like◾Neil waits for proof, then bets big; Zuckerberg bet first and won Chapters00:00 Nvidia's CUDA moat is leaking02:29 Neil: brand is a moat, capital isn't03:35 Harry Stebbings: three rounds in three weeks05:07 Six of the top 10 companies come from the Bay07:02 The three moats that survive09:13 Neil: AI shoppers still care about brand11:46 Meta Muse and Scale's $14B bet12:16 Doodoo picks Eric's wardrobe13:04 Muse + Jev sort 116K followers for $3.3014:29 Neil's influential followers15:43 Why Neil waits for proof, then bets big18:11 Why this podcast works: persist vs change19:03 The personality test: 99% vs 1%20:50 Empathy at work

Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/Want to recruit great marketers? Find them here: https://marketingschool.io/hire Neil explains why he despises coaches who sell how to get rich, and Eric separates them from the Bill Campbells and Matt Mocharys who coached Apple, Google, Coinbase and OpenAI without charging founders. Neil reads a cold pitch promising him an extra million a year and Eric dissects why it fails before the first reply. Then the two ideas Eric wants every operator to react to: supermemory killed its viral company brain to protect its core, because unshipping matters more than shipping when anyone can build anything, and Paul Jun of Ramp's brand as software, where a design file carries the brand into every asset so designers do the new work. Neil's take on ChatGPT agent ads, and BugBasher, the Devin agent Ramp gave a card and a phone that made $75 selling pest-control leads in four weeks. A sharp episode on focus, coaching and what AI agents can already do. Key takeaways◾If someone doesn't do the research before your check, it gets worse after it◾When you can ship anything, unshipping is the discipline that keeps you extraordinary◾Codify the brand as software so creative people spend their time on what a system can't predict Chapters00:00 Coaches who sell 'how to get rich'00:48 Two kinds of coaches: Bill Campbell vs the funnel02:03 People who are good at it, do it02:42 The Trillion Dollar Coach and Matt Mochary03:56 Neil reads the cold pitch05:14 What's wrong with that message06:00 No research before the check? It gets worse after07:03 The legendary coaches never charged founders08:25 Proven track record vs get-rich promises09:21 Unshipping beats shipping11:06 Shipping too much makes products harder11:41 supermemory's conflict of interest12:35 Content that wins on X and LinkedIn13:14 Brand as software: Paul Jun of Ramp15:25 Are you working like this? Design files and AI-pilled designers16:10 Boomerang employees prove the vision17:02 ChatGPT agent ads are coming17:35 BugBasher: the AI agent that made $7519:36 Agents that lose you money

Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/Want to recruit great marketers? Find them here: https://marketingschool.io/hire Eric opens with Bryan Johnson's team telling creators to forget the content calendar and iterate, and how he now rewrites and reposts flopped posts by voice on the walk to the gym. Neil's newest test asks viewers to comment for a free audit, and the pairing metric he wants on every social hire is leads, not views. Then enterprise specifics: a three-year contract that grew out of work in one region, why only two to three percent of NP Digital's enterprise revenue traces back to Neil's personal brand, and how ten-person search agencies quietly decide the biggest marketing contracts. Eric opens his Leveling Up channel to show the wide-TAM experiment, 8.9 million views on one short and interviews that got praise from entrepreneurs' spouses but no leads, before Neil explains the acquisition lesson that grows a bought agency's EBITDA 40 to 60 percent in a year and why he follows up with everyone. A grounded episode on measuring what actually pays. Key takeaways◾Pair views with leads for every social hire, or the team optimises for likes◾Wide-TAM content brings views and comments, not enterprise leads◾Buy the specialist that big brands already call; the follow-up is where the deals are Chapters00:00 F your content calendar01:36 Comment-to-lead: the 37-places reel03:34 Every metric needs a pairing metric04:17 NP Digital vs NP Accel: where SMB leads go04:53 A format library your agent keeps fresh06:09 A 3-year contract from one region's work07:56 Stop creating content for 90 days?08:22 6% of enterprise leads come from Neil's brand08:58 70% of SMB leads, 50% of SMB revenue10:17 Procurement, RFPs and the search agencies11:56 Dan Martell's formats and wider TAM12:52 Seminars and ascension funnels13:34 Eric's wide-TAM experiment on Leveling Up15:17 The shorts: 3.9M, 2.7M, 2.1M views17:02 Why Eric stopped: no leads17:37 Distribution beats content when you have leverage18:44 The acquisition lesson: 40-60% EBITDA growth21:21 The fortune is in the follow-up22:07 'How are things going?' scheduled for March 2027

Growth Newsletter: https://levelingup.beehiiv.com/subscribe Need marketing help? Visit: https://www.singlegrain.com/ and https://npdigital.com/Want to recruit great marketers? Find them here: https://marketingschool.io/hire It took 64 clicks to buy a Tesla online, 30 of them signing loan documents. Eric explains the five-step algorithm from the book of the same name (question, delete, simplify, speed up, automate) and how it 20x'd Tesla's website sales. Then the data: AI-generated ads are beating human creative on completion rate (92% vs 90% on connected TV), Shopify's Q2 shows $116 billion in GMV and a platform built for AI agents, and Neil's case that SEO and GEO are merging into one thing, brand. Eric on the new agents he is testing, Instinct and Meta Muse, and the week's grand finale: Rohan Nayak's post on how Pocket FM went from zero to $500 million in ARR with an AI ad machine that makes 17,500 ads a month. Key takeaways◾Question, delete, simplify, speed up, then automate; most people automate first and fail◾SEO or GEO, the long game is brand; you have to show up everywhere◾A 2% to 2.25% CTR lift cut Pocket FM's CAC by 30%; build the ad machine around that insight Chapters00:00 Tesla 20x'd website sales: 64 clicks to 301:35 Buying a car on a website02:55 AI ads beat human creative on completion04:19 Shopify Q2: $116B GMV and AI commerce05:38 SEO and GEO are merging into brand08:04 Search everywhere optimization, again10:49 A SaaS company diversifying off SEO11:46 Instinct and Meta Muse agents14:15 Pocket FM: $0 to $500M ARR with 17,500 ads a month17:11 Build your own ad machine