Mai-Lan Khong

We run content for an ads team, so I keep getting asked to point at the post that booked the call. The instruction underneath: find the winner, do more of that. In their world that is correct. In mine the instruction has nowhere to go. Ads rent a new audience every time you pay. The winner goes out to 10K strangers who have never seen it, so you scale it by running it again, wider. Repeating with variations is the whole point. Content doesn't work that way. You live in the same room. The same people see you post after post, so when you repeat your winner they have already heard it. You can't strip-mine a good post the way you strip-mine a good ad. That one thing, renting strangers versus living in the same room, changes the sequence: → The winner stops being a post and becomes a lesson. You never run it again. You take whatever made it good and pour it into a new idea next week, for the same people. → So the job is a new idea every week. I used to think that meant I was scattered 🙃 → Ads get narrower over time, down to the one thing that works. Content keeps getting wider, because you learn what kind of ideas land and then keep inventing new ones. The method is same: test, measure, learn. I drew the whole thing into a map 👇

@vaughngooding :D they are not fixed set of 25 questions but an interactive conversation - so the questions will be proposed at that moment when you share your interest. Then the voice profile will be captured from that natural flow. I'm very proud of it haha, I have a version of 100 questions as well for deep dive. Will be sharing soon!

@mailankhong Can you share those 25 questions?

I have a content twin. It runs inside Claude Code, it writes like me, and I do not start from a blank page anymore 🙂 If you are still typing "write me a LinkedIn post" into an empty chat, you will keep producing the kind of thing people have read a hundred times, written by a hundred people who all sound the same. What turned it around was one file: 25 questions about how I actually talk. My beliefs, my humor, the words I reach for, the ones I would never use. Everything else stacks on top: → It learned who I write for and the few themes I keep coming back to. → Research pulls from five places at once, so I write about what my audience cares about. → Ideas arrive already ranked. Monday never starts empty. → 20+ hooks per idea, I keep the one that pulls. → It drafts in my voice, then I trim and drop in the real story. → It grades itself before I see anything. Five dimensions, out of 50. Below 38 goes back. → What faded gets dropped, so the next cycle starts sharper. Save it for the day you open Claude and feel the blank page panic.

you liked something of ours thirty days ago. you've forgotten. we have not. you'll reach out eventually and think you found us.

My first viral post passed 50K impressions. I sat there near midnight refreshing, watching the number climb, feeling like I had finally won something. The next morning my inbox was still empty. 🥲 That is all a loud post is, sometimes. A lot of strangers clapping in a room they will walk out of and won't remember. Months later a she-EO messaged me. "I've been reading you for a while, can I hire you to help with my LinkedIn profile." I went back to find which post had brought her. It wasn't the viral one. That was the day I understood these 3 things: 1. Reach is a crowd walking past a street singer. 2. Resonance is the one person who stops and writes the name down. 3. Revenue arrives months later, when she comes back and asks for you by that name. Almost all the energy goes into gathering the crowd, and the business is sitting with the one person who stopped. So before you chase the next viral post: how many of today's readers will remember you tomorrow?

@itsalexvacca context is everything!

Creating content at scale, especially with tools like Claude Code, comes down to building context more than good prompting. Give the system as much context as possible: - angles that work - posts that did well - posts that haven't done well - your best hooks and templates among many other things... and your AI content would be significantly better.

@itsalexvacca this is absolute gold! 🔥

Your AI writes like a robot because you talk to it like a human. It doesn't fill in what you left out, and it never will. Tell a friend you are thirsty and they hand you a glass of water. Say the same thing to AI and it takes the words at face value, because it has none of the context you carry around without noticing. You have to hand that context over on purpose. Two fixes I run on every draft. The first is to brief it like a new hire, with a role, a task, constraints and a style. "You are a content strategist for B2B SaaS, 300 words, no bullet points, write like you're explaining it to a colleague." That will always beat "write me a post." The second is to show it instead of describing it. Paste three things you have written yourself, point out what makes them sound like you, and ask it to match that pattern. AI is hopeless at following a description of your voice and surprisingly good at copying a real example.

still can't believe people pay thousands a month for lead lists

LinkedIn changed its algorithm in March. Eight things that worked in 2023 now actively cost you reach. The four that trip people up the most: 1. Length. The old sweet spot was short and punchy, and now it is the opposite. Long posts win, 1,250 to 3,000 characters, for +227% reach. 2. Links. They used to bury a post, and now they lift it. One to three links, +43%. More than three, +441%. As long as you are sending people somewhere worth the click. 3. Format. Carousels fell off a cliff, down to 6% of top posts. Infographics, video and quotes carry 79% between them now. 4. Replies. The strongest creators answer their comments 255% more than everyone else, because the conversation is what keeps a post alive. Reach is already down 68% since 2023. Running the old rules just gets you there faster. I put together an in-depth breakdown of everything I know about growing on the platform (coming from someone that has scaled multiple LinkedIn accounts that generated millions in recurring revenue):

Publish a LinkedIn post, glance at it once, close the app, and you just threw away 70% of its reach. Almost everyone does exactly that. LinkedIn tries every post out on a small crowd first. Get some traction in the opening 60 to 90 minutes and it opens the doors wider, get nothing and it lets the post slip away. That single window decides about 70% of the reach a post will ever see. The actions inside it carry different weight, too. A comment is worth 15x a like, a share 5x, a save 3x. Which means the moment you publish, the work has barely started. Reply to every comment in that first hour and you earn 50% more engagement across the whole life of the post.

Almost every call I've booked on LinkedIn started in someone else's comment section. Posting is what makes people warm to you, but the real conversations tend to start earlier than that, down in the comments, weeks before I have pitched anyone anything. There is a system underneath it all. Fifteen comments a day, spread across three tiers. 1. Your ICPs, the people you want as clients. Open one prospect's profile, look at who they reply to, and go find those people too. They will have seen your name around long before you reach out. 2. Your peers, same niche, competitors included. Their feed doubles as a live readout of what is landing this week. 3. And the big accounts, where one insightful comment in the first half hour can put you in front of thousands of exactly the right readers. How you comment is the part that decides everything. Also, don’t hesitate from disagreeing or sharing a contrarian viewpoint. A like or a supportive comment is generally forgettable and lost in a sea of other affirmative comments. But a disagreement or a no can end up earning you actual replies.

@itsalexvacca distribution ftw!💪

I've seen a lot of worse products beat better ones, and the reason is almost always the same. We've run outbound for 275+ B2B companies so far, across just about every kind of product you can think of, the polished ones and the ones barely holding together. Across all of them, it comes down to the same thing: distribution. Distribution is getting the right offer in front of the right people, with a message relevant enough that replying is the obvious move. It takes patience, and there's no clever way around it. None of that is new, but AI has pushed it to an extreme. Now that anyone can rebuild your product over a weekend, how hard it was to build stopped counting for much, and what you're left competing on is distribution. That happens to be the part most people are worst at. It also adds up over time. Someone can clone your software and still be starting from nothing, because they don't have the list you spent months cleaning up, the offers you already tested, the reply data, or the pipeline that's warm to you and nobody else. That takes a long time to build, and none of it comes across when they copy you. That's what we build and run for you at @Frontal_AI. You start with it already booking meetings, instead of spending your first year building it.

Every week I read B2B founders' LinkedIn posts, and I can place almost every one of them on a ladder they don't know they're standing on. Four rungs, and most are stuck on rung 2, dead sure they're on rung 4. Rung 1 is generic. "5 tips for better marketing." No story, no numbers, and you can smell the empty Claude tab it came out of. Rung 2 is the templated framework. A hook, a tidy list, a call to action at the bottom. It photographs like strategy, but change the name up top and it still works for anyone, which means it belongs to no one. Most people settle here and mistake the comfort for progress. Rung 3 is where a voice shows up. The framework is still there, only now there is a real client sitting inside it, a real number from the business, a way of phrasing things you would know even with the headshot cropped out. Rung 4 is authority. You are teaching from something you did last quarter, with an opinion you paid for by doing the work instead of reading about it. People start trusting you at rung 3. Rung 4 is where that trust starts turning into pipeline.