Brian Wool 🍳

20M views on an adapted winner from a direct competitor. Good artists copy. Great artists steal. Can you spot the similarities in these two YouTube Ads???

@madsf88 Abraco, devocion, hi collar for an intense coffee experience

best coffee in nyc ?

@SidBendre Have you also been getting that weird "an image is taking up too much context" error forcing you to compact all the time?

Even Opus 4.7 on chat seems to be worse at connecting the dots, or processing implicit stuff, switching back to 4.6 since it seems to be better at analysing info

I wonder what the stats will look like at the end of week 1 🤔🤔🤔

Gruns x Olipop launched a YouTube ad test 48 hours ago 🤯 I pulled the data via Pathfinder. Same product, two creators, two scripts. → jennakatz31: 25.7K views → badgyalbren: 61.2K views (2.4x ahead) The winner is 7 seconds shorter and pulling 2.4x the views. The losing ad closes with "Thanks, Grüns" and lists features: gut health, energy, immunity, recovery, beauty, cognition. It reads like a supplement label. The winning ad closes with "Limited edition. Go get yours before they're gone" and sells sensation: "cute packets," "strawberry-shaped," "I always have to do my little dance." Definitely s signal, but DTC growth teams kill losers in 2-3 weeks. By day 2, the team has its winner. Cold audiences buy the vibe. They read the label after.

Do not complicate YouTube ad research. - Go to Pathfinder - Search for winners - Feel incredibly stupid when you see that the top performing YouTube ad is literally just a shitty rocket video - Cry in the corner I wish I was kidding. Over 600M views 😅

Wouldn't it be crazy if you could see every YouTube ad WisprFlow was testing and which ads became winners 🤔🤔🤔

@jaketheadnerd Want to run *good* YouTube Ads? See what works and doesn't work for everyone else usepathfinder.com

Want to run YouTube Ads? This is it. The framework for a profitable/reliable funnel. Baseline holds the funnel. Multipliers bend the curve. Simplicity > Complexity. Ie. Simple for a reason. You don't need MMM from Day1 to show what's working (but add it). Questions? DM me

@signulll chegg

which product in the market today has effectively lost product market fit after once actually having it?

when you’re stretching in bed & you feel your calf start to tighten

@WillManidis Half the bit was that they’re already successful and connected, so they were just having fun with it

i genuinely think everyone who has “studied” TBPN fundamentally misunderstands what made the show special and are destined to nuke capital chasing it.

@Jibran_05 pretty unfair imo but I guess thats social media. Adapt to survive

@Jibran_05 Go to a few of those "day in the life of a harvard student" or calm lofi nyc lifestyle content creators on Youtube, go to their content library, sort chronologically, and you'll see it. It's like over the course of one month, their accounts got nuked

the influencer cinematic video -> boosted ad method companies like - notion - cursor - adobe - anthropic spam this method. find an influencer with 500k-2 million followers. pay them anywhere from $2-5k for a video. the video inevitably flops (no one wants to watch their favorite influencer shill a product) then, they boost it as a paid ad. i doubt the ad converts, but everyone starts associating your brand with "the aesthetic notion-esque company" prediction: this will get quickly saturated, these influencers will start charging more, and viewers will stop giving af there can only be so many "notion-esque" companies, and spamming this method eventually will become a new type of slop

@spgreenwood @JackKuhr Is the vulnerability worth the cost savings? Absolutely

@JackKuhr Feels vibecoded / vulnerable to exposing private data

So ridiculous, yall must know. For the last month I’ve been using Empower ride hailing app in NYC. (They charge drivers a monthly fee instead of taking 20%, I believe). Cars always available. And always much, much cheaper than Uber. Tonight was particularly insane. LaGuardia airport to home: - Uber: $125 (criminal) - Empower: $30 IDK guys it feels like $5 Via days. Don’t know which good soldier built this app, or why. But it feels like they’ve punched a hole in the absurd NYC ride hailing pricing schemes that we all had come to just accept.

@codyschneider Or you get @BoraMutluoglu to cook for you

How to Build an AI Agent for Influencer Outreach Outreach is the grunt work of influencer marketing. You find creators, vet them, personalize every email, follow up multiple times, and hope they reply. Most brands burn out after week two. What if an AI agent could do 80% of this for you—on autopilot? Here's how to build one. The Problem With Manual Outreach Traditional influencer outreach is broken: Slow. One person can manage ~50 emails a day before quality drops. Generic. "Hey love your content!" gets ignored 97% of the time. Inconsistent. Follow-ups get forgotten. Relationships fall through the cracks. The influencer marketing industry is projected to hit $32.55 billion by 2025. Brands still doing outreach manually are getting left behind. The solution isn't another tool. It's an agent. AI Tools vs. AI Agents AI Tools Requires prompts for every task Helps you do the work AI Agents Executes multi-step workflows autonomously Does the work for you Your outreach agent should handle discovery, personalization, outreach, follow-ups, and tracking—without you touching the keyboard. The 3-Step Framework Step 1: Content-Matched Discovery Don't filter by follower count. That's vanity metrics. Your agent should analyze: Video content & hooks – Find creators whose aesthetic matches your brand Engagement quality – Comment sentiment, not just likes Audience demographics – Does their crowd align with your target customer? Pull this data from APIs like SociaVault or platform APIs directly. Feed your agent real-time data, not stale databases. Step 2: Hyper-Personalized Outreach This is where most outreach fails. Generic templates = generic results. Your agent should: Pull the creator's last 10 posts Extract their hashtags, themes, content style Generate a personalized pitch referencing their actual work Instead of: "Hey! Love your content. Want to collab?" Your agent sends: "Hey [Name], been following your [specific video]—the way you broke down [specific topic] was exactly what I needed. Our product [specific benefit] would be a natural fit for your audience." That gets responses. Step 3: Automated Follow-Ups (With Human Oversight) Outreach takes 5-7 touches to get a response. Your agent should: Schedule follow-ups based on response or silence Manage multiple outreach threads in a CRM Escalate to humans for contracts and negotiations Human-in-the-Loop (HITL) matters. Let the AI handle logistics. Humans sign off on contracts, creative briefs, and high-stakes decisions. What Your Agent Can Do ✅ Discover creators based on content, not follower count ✅ Analyze engagement quality and audience fit ✅ Generate personalized outreach emails ✅ Send initial outreach + automated follow-ups ✅ Track replies and update your CRM ✅ Flag opportunities needing human attention One person with this agent can do the work of an entire outreach team. The Bottom Line The brands winning at influencer marketing aren't working harder. They're working smarter. An AI outreach agent removes the busywork so you can focus on creative and relational parts that actually close deals. Start small. Build discovery first, then outreach. Layer in follow-ups and CRM integration as you go. The future of influencer marketing isn't more tools. It's agents that actually get stuff done.

@SimonasDip @Polymarket Big @tobinwtang energy

so @Polymarket is literally changing how UGC is done they're making short-form more like longform a video, that has nothing to do with polymarket has a quick ad of it in the middle it's genius > video is made for entertainment only > viewer is hooked and wants to see the end result > has to watch the ad to get to the result it's exactly what YouTubers been doing ngl this might be the new meta for broad products hats of to whoever came up with this p.s. the DM format itself is from elsewhere, not orignal

@trq212 @hnshah Cowork + freetaxusa >>>>

there's a turbotax connector in Claude Code now, so glad I procrastinated on taxes 😭

@Nate_Google_ People treating YouTube different from meta will never make sense to me. Test. Test. Test. Top performing accounts are always testing new creative

currently doing an audit on a brand doing around $3m/m frustrating man… Google again is just 10% of rev well known guru running the account as well they can’t crack YouTube ads, shopping and search is all over the place, not even testing advertorials or different PDP styles the BASICS not being done and this account has been like this for over a year time wasted, but soon to be gained back FAST

Google's Ads Transparency Center is useless on purpose. It shows you every YouTube ad a brand runs. But all the important details? - View counts? - Spend? - Which ads are scaling and which got killed after a week? They buried all of it. You're staring at 200 ads with zero way to tell the $50K winner from the $500 flop. Google knows. They just won't show you. So I built Pathfinder. Search any brand. See every YouTube ad they're running. - Real view counts. - Views per day. - Ad lifespan. - Estimated spend. - Which ones are still live and which ones got axed. Pulled Perplexity's full library just now. 83 unique videos. Top performer: a Brazilian UGC clip with 122M views, averaging 964K views/day for 127 days. Bottom of the list: An unlisted shorts ad with just 45 views in 11 days before they killed it. Same brand. Same budget team. Google treats those two ads identically. Pathfinder shows you the difference in two seconds.