Nobody talks about site speed as a growth lever…
A one-second delay in page load time can drop conversions by 7%. At scale, that's not a rounding error, but a significant chunk of revenue quietly walking out the door.
Your customers aren't patient, they have seventeen other tabs open and zero tolerance for a slow experience.
Fast sites win. Slow sites lose. Simple as that.
Check your load time today like it's a KPI, because quite frankly, it is.
The engineers I respect most have surprisingly vanilla workflows.
No elaborate AI orchestration or month-long sprint to find the perfect stack.
Too much setup replaces raw skill with delegation. And delegating to a tool you haven't benchmarked is just organized guessing.
My approach: try one thing, see if it beats the baseline, and take the next best step.
It’s simple, iterative, and honest about what's working.
The most sophisticated workflow isn't the best workflow. The one that actually makes you sharper is.
Every few months a new supplement, device, or protocol promises to change your life.
Most of it is noise.
Here's the filter I use for everything:
Does it replace the fundamentals, or does it enhance them?
If it replaces them → it's a distraction. Move on.
If it enhances them → it's a tool. Learn it and use it.
The fundamentals never change. Sleep, movement, nutrition, and stress management will always be foundational.
Anything that helps you do those things better is worth exploring.
Anything that lets you skip them entirely is selling you a shortcut that doesn't exist.
This is why I'm so bullish on peptides done right. They don't replace the work. They make the work more effective, and the results compound faster.
Same goes for AI.
Same goes for any coaching program.
Same goes for anything you put in your body.
Apply this filter before you buy the next thing promising overnight results.
Netflix knows what you want to watch before you do. Furniture retail still hands you a search bar.
Streaming figured it out, music figured it out, and even grocery delivery is starting to figure it out.
Furniture never did.
The recommendation engines that exist today in home retail are essentially popularity contests.
They show you what's trending, what's frequently bought together, and what other people with vaguely similar carts have looked at. But none of that is personalization.
Real personalization means knowing your room, your existing piece you inherited from grandma, and knowing that you lean transitional but hate anything that feels too formal.
This includes your ceiling height, your lighting situation, and your budget ceiling, and surfacing the three items that actually fit all of it, instead of the three thousand that technically match one filter.
But this begs the question…why hasn’t this been solved yet?
That's the thing about a problem that's been unsolved for long enough.
People stop expecting it to be solved.
That's exactly why we're solving it.
The bar is so low in this industry.
We’re changing that.
Nobody talks about the cost of keeping the wrong things.
The product line that’s draining ops.
The channel that barely breaks even.
The team member you’ve been managing around for eight months.
Cutting feels like losing.
But everything you hold onto that isn’t working is taking resources - time, money, focus - away from the things that are.
The best move you make this year might not be adding something.
It might be finally letting something (or someone) go.
Everyone’s trying to “figure out” Claude Code right now...
But they’re all doing it the WRONG way.
I’ve spent every hour of every day in the last 4 months learning claude code…
It isn’t just another tool you learn once and move on.
Because once you move past the “this is cool” stage…
you need structure.
You need repeatable ways to:
→ find leads
→ clean and qualify data
→ turn that into something you can actually use for outbound
That’s where most people drop off.
So I put together a full walkthrough.
Exactly how I’m using it for GTM right now.
Think of it as the first step before you go deeper into this properly.
Inside this loom video:
1. how I approach lead generation
2. how I keep things from falling apart halfway
3. a few workflows you can copy and tweak
4. where Claude Code fits into your existing stack
I WISH I had this when I first opened it.
If you want it, comment “BUILD” and I’ll send it over.
(must be following)
We booked 107 meetings in 30 days using two channels and no paid ads.
We used outbound and content only.
The 8-step system behind it:
1. LinkedIn Content: 44 total posts as a team, 1.6M impressions, Claude Code drafts copy with Pinecone, Scripe for ideation and Figma for design
2. Optimized LinkedIn Profiles: featured section and CTA links consistent across all team members
3. Website Positioning: copy tailored to ICP with social proof, Tally and Cal(.)com for application routing, partial submissions routed to Clay then Slack
4. Track Website Visitors: 7,000 monthly visitors, Warmly tracks those who don't convert, accounts pushed to Clay
5. Track LinkedIn Engagement: Teamfluence for profile visitors, Jungler for post engagement, periodic connection exports to Clay
6. AI Account Research: accounts auto-qualified in Clay, Apollo for decision makers, Findymail validates 90% of emails, BetterContact for phone numbers
7. Multi-Channel Outreach: daily Slack alerts with warm leads, HeyReach for LinkedIn, InstantlyAI for email, Nooks for cold calling
8. 107 meetings booked.
Outbound and content are the two channels we keep doubling down on.
PS
I wrote TWO different step-by-step guides...
1. A full outbound masterclass
2. A full LinkedIn content masterclass
Comment “107” and I’ll send you both.
If you aren’t making $ on your ads
Have you ever LOOKED at your funnel/ads/sales process like a prospect?
Like actually look at it with a fresh pair of eyes
Would YOU buy from you with that funnel?
Would YOU click on those ads?
Or is it just slop
We've overseen the GTM motions of 250+ B2B companies.
In this document are the NINE strategies worth running in 2026.
And we ranked each by 2 things:
1. difficulty
2. time horizon.
The full high-res PDF includes:
- Each motion rated 1-5 stars by difficulty
- Each mapped to short, mid, or long time horizon
- The one-line description of how each motion actually runs
- Which team profile fits which motion
- The two motions we'd start with at every stage of company
This is based on ACTUAL data we’ve receieved from our clients across the enterprise and mid-market sector.
Comment “MOTIONS” and I'll send you the full PDF.
(must be following)
Just got off a call with a friend of mine who took an earplug brand from $10M to $250M in under 2 years.
An absolute killer with the way he thinks.
The brand was built around parties and festivals.
Then COVID hit.
There were no more parties and they nearly went bankrupt.
Instead of panicking, he started mapping every possible use case for the product.
- Sleeping
- Parenting
- Commuting
- Restaurants
- People with ADHD
37 use cases in total.
Each one became its own angle, its own creative, with its own audience.
That move took them from $10M to $250M.
But what stood out to me was how he runs the operation day to day…
- Revenue was tracked per ad, per channel, and per use case
- Out of every 100 ads, minimum 20 have to work
- 6-7 agencies were running simultaneously, each with a clear ROI target
- Every agency knew exactly what return they needed to keep the budget
One thing he constantly presses on:
"I don't care if it's an image or a video. I just want to know if the message works."
One of the sharpest operators I've ever seen in the game.
There are 17-year-olds right now scaling health supplement brands to $100k+/mo…
While 40-year-old DTC founders with 20 years of experience are stuck flat.
They didn’t win with more talent, longer hours, or bigger budgets.
They grew up with AI as the default and reduced the entire Meta ads learning curve into weeks instead of years.
I've generated over $500M+ in DTC revenue, and I’m watching this gap widen every single month.
Old way:
- Years of trial & error + wasted spend
- Full creative teams + $15k/month retainers
- 8-12 tests per month max
AI-native way (2026):
- Claude Opus 4.8 + Meta MCP = instant performance audits & next-test recommendations
- Higgsfield MCP = unlimited creatives & videos at zero marginal cost
- 40+ angles tested monthly
- Day-3 iteration instead of waiting weeks
Creative velocity is now the only edge that matters.
The algorithm rewards operators who flood it with fresh, diverse creatives before fatigue hits.
If you’re a 7-9 figure health brand still running the old playbook…
You’re getting lapped by kids with laptops.
DM me “AI” and I’ll install the full AI-powered Meta workflow for your supplement funnels (creative → testing → scaling).
Let’s close the gap.
I set up my columns to follow the buyer’s full funnel, top to bottom.
Every single step is visible. When something breaks, I spot it instantly.
Here’s the exact setup I use for health supplement brands:
→ CTR
→ Landing page views
→ Add to carts
→ Cost per add to cart
→ Add to cart rate
→ Checkouts initiated
→ Cost per checkout
→ Checkout rate
→ Checkout to purchase rate
→ CVR
→ Purchases
→ ROAS
Most people only watch Purchases & ROAS.
Big mistake.
Cost per add to cart = your earliest signal. Especially on $100+ AOV products, where purchases can take days. It tells you which creatives are actually pulling people into the funnel.
Cost per checkout = the hidden gold. High CPA but low cost per checkout? The ad is doing its job — something on the offer page is just leaking.
Build this column preset once. Never fly blind again.
Bookmark this and thank me later!
how to make scroll-stopping human UGC:
first, understand the moment that matters MOST is the first 2 seconds.
a viewer's thumb is already moving when your video loads.
so you've got about 2 seconds to freeze it before they're gone.
everything below is built around winning that one window.
1. lead with the most interesting thing
no intro, no "hey guys," no slow windup.
open on the boldest line or the most visual moment you have.
2. make it look native
- film on an iphone with window light.
- add snapchat-style text overlays.
- don't edit it too much.
you want it to look like a real person posting.
3. put a face in frame one
the brain locks onto human faces before it processes anything else.
an actual person looking into the camera in the opening frame is one of the strongest scroll-stops there is.
an AI avatar can't pull this off, the eyes always give it away.
4. keep the imperfection
a slightly awkward, casual, real take beats a perfect one every time. don't over-direct your creator.
the tiny imperfections are exactly what signal "real human," which is what holds attention.
5. open a loop
give them a reason to keep watching past the hook. a question, a "wait for it," a problem they recognize instantly.
the open loop carries them through to the end, which is the watch-time signal the algorithm rewards.
6. caption everything
most people watch on mute.
if your hook only works with the sound on, half your audience never even gets it.
do these and the video stops the scroll on its own.
7. USE AN ACTUAL HUMAN
AI won't be able to replicate this, no matter how many times you regenerate or how good you are at prompting
and it's inexpensive to run real humans too compared to AI UGC
you can literally get it for $1 per video nowadays on platforms like HourlyUGC
then just have a filipino edit it for you and claude find winning scripts.
fully automated at a low and scalable price + volume.
- Start with one AI tool
- Use voice dictate to prompt faster
- Build your personal intelligence blueprint
- Create 100 posts in one session
- Use agent mode for outreach
- Let AI find your speaking gigs
- Build an asset that pays you back
- Teach what you learn
- Show your work publicly
- Pick your spot and get in
This is the path.
I just built a full team of LinkedIn revenue agents.
(trained from 19,000+ inbound leads & $5m in revenue)
Each sub-agent has its own context window, loads only the knowledge it needs, and ships drop-in-ready work for its step.
Every "AI agent for LinkedIn" I see tries to run all 5 steps of a revenue system in one giant system prompt.
Why these break:
- The context window blows up halfway through the week
- You can't run ONE step without reloading the entire system
- Knowledge relevant to one step is noise for another
- Handoff between steps is implicit, you have to remember which prompt feeds which
None of this gets fixed by a bigger agent.
The fix is splitting into 5 smaller ones, one per LinkedIn surface that actually drives revenue.
The roster:
- The Content Agent: owns the feed. Mines sources, writes posts in my voice, schedules the week, AI-isms QA'd inline.
- The DM Agent: owns the inbox. Inbound LM triggers, outbound to engagers, the reply playbook for the 4 reply types you actually get.
- The Profile Agent: owns the static surfaces. Headline, banner brief, featured section, About-as-qualification-gate, pinned post rotation.
- The Lead Magnet Agent: owns the assets. Picks the topic, ships the LM, builds the funnel, updates the library so the rest of the team can route to it.
- The Ads Agent: owns paid distribution. Graduates the winning post into a Thought Leader Ad
Plus an orchestrator claude.md that routes the week's work between them with explicit handoff contracts.
Claude Code's native sub-agent fan-out runs them in parallel.
The pack:
- Orchestrator claude.md
- 5 sub-agent specs
- 9 knowledge files
- Install for Claude Code local, Claude Projects
Want them?
1. Comment "TEAM"
2. Follow me
And I'll DM you the pack.
Hard work does not beat talent. Game selection beats both. Pick the wrong game and a 12-hour day still finishes last. Sick people understand this in their bones. They watched healthy friends "work hard" and got told they were lazy for the same effort.
Fibro Support opens with a sister who called the writer a faker for four years. Four years of medical records, specialist notes, research papers handed over family dinners. Every dinner ended the same. "You just need to push through it."
This is the same mechanism every chronic illness operator is sleeping on right now. Your buyer is not fighting the disease. They are fighting the people around them who decided the disease was a personality flaw. The villain in your hook should be the brother, the sister, the husband, the boss. Not the symptom.
The ad spends the first 800 words on social betrayal. Then pivots to the body experience. Skin that hurts to touch. Brain fog that drops thoughts mid-sentence. The two days in bed paying for one good day.
The product, an Earthline Naturals fibro supplement, enters as the first thing that took her side.
One of 8,400+ advertorials indexed at https://t.co/Y5bWx951ID. The social-villain hook outperforms the symptom hook by 30%+ in our top decile.
Name the person, not the pain.
Link to this ad: https://t.co/ziUil4zLkF
We track 70M+ winning ads at https://t.co/Y5bWx951ID