redsymbol

First-principles understanding is durable to many changes in culture, trends, and details of technology and implementation

@brett_fish So it sounds like this was not something that affected performance of e.g. Purchase conversions? I had an adset whose CPA exploded around then. According to this statement, this LPV issue would not have been the cause of that.

While it’s now resolved, here’s Meta’s official statement in the LPV bug, FWIW: "On January 9, 2026, we discovered a system issue that caused a significant decrease in reported Pixel Landing Page Views (LPVs) for some advertisers. As a result, some advertisers’ Meta Ads Reporting and Ads Manager dashboards may have shown lower than expected landing page views, even though the actual user activity did not decrease. The issue began on January 6, 2026, and advertisers should rely on their own reporting metrics for landing page views during the impact period. Following our investigation, we identified the root cause as a recent update. We reverted the update and developed a fix to restore accurate LPV tracking and reporting. The issue was resolved on January 13, 2026, at 2:56 PM PST, and all future reports, after the issue was resolved, will show accurate LPV data. We are currently exploring options for restoring the accuracy of LPV reports during the impact period and possibly backfilling the data. We recognize stability and reliability are critically important for the people and businesses who use our services around the world. Over the years, we've invested in a wide range of processes and technologies to ensure that our infrastructure is resilient and highly available. And we will continue to work on this and we’re committed to getting even better.”

@theericcarlson Dif this LPV issue affect assets optimizing for Purchase? I had an adset whose CPA more than doubled around this time, to the point I turned it off.

Something weird happened on Meta starting around January 5th and 6th. A bunch of accounts saw a major drop in the percent of link clicks that turned into Landing Page Views (LPVs). And I don’t mean a small drop. Normally you’ll see something like: 75% to 90%+ LPV rate But suddenly it was showing: 20% 30% 40% At first, a lot of people assumed it was iOS. But when you break it down by device… it didn’t look isolated. It was happening across the board. Not every account got hit. But enough did that it mattered. And yes, I think it impacted performance. Most people forget this though. Meta is not a controlled environment. It is not a static platform you figure out once. It is a living codebase. Constantly changing. Every day you’re dealing with: Marketplace competition Seasonality New placements and formats Measurement quirks Delivery changes And yes, bugs This is why I laugh when people talk about Meta like it’s a clean science experiment with stable variables. It’s more like trying to run a business on top of a product that’s being rebuilt mid-flight. Also, I’ve noticed this pattern for years. When engineers come back from holiday break, weird stuff spikes. Big dev pushes. More unexpected behavior. Good news: it looks like it’s already correcting. This morning, multiple accounts I checked were back to: 80% to 85%+ LPV rate So traffic quality should normalize quickly. But the bigger takeaway is simple: Meta is imperfect. And it changes constantly. So if performance randomly drops, don’t assume you suddenly got worse at ads overnight. Sometimes the platform is just having a week.

@daltonbrewer First-priciples thinking

HOW DOES ELON THINK OF THIS STUFF?!

@jacob_posel I like to think in terms of what I call "protein efficiency", which is the grams protein divided by cals. This ratio is independent of serving size, and is helpful for quickly evaluating different food choices

How I got shreddy during the great lock in of 2025: 1. Diet Protein is your best friend. Aim for 1-1.5g per lb of body weight. Some protein hacks: 1. Greek yogurt + protein powder 2. Cottage cheese + eggs 3. Low calorie protein bars Anything around 20g protein per 100 calories in excellent DON’T OVER EAT The goal is maximize protein to calorie ratio Typical day of eating for me: 9am: bowl of plain greek yogurt with protein powder 12pm - high protein lunch meal (usually 93/7 ground beef w/rice OR eggs w/cottage cheese) 5pm - pre-workout strawberry rx bar 6-7pm - high protein dinner (usually 93/7 ground beef/rice/kimchi) 8pm - another bowl of Greek yogurt I’M ALWAYS HUNGRY, and that’s the point Supplements: Creatine. I mix it in the yogurt. Use ChatGPT to count cals and protein 2. Workouts My split: push, pull, legs Sometimes I’ll do chest, back+shoulders, legs because I’m vain The two core elements to training most people miss: Intensity: Training should be HARD. Go close to failure. Don’t wait long between sets. Overload your muscle. Every set I grimace in pain. Form: Form is so important. I’ve watched thousands of hours of Cbum, Mike Thurston etc to learn. Record yourself, workout in front of a mirror. You’re leaving a lot of quality muscle development on the table if you don’t complete an exercise properly. The basics of form: controlled deep descent, proper squeeze during concentric phase, touch your muscle for mind muscle connection Ie if I’m doing a machine chest press, I twist my body 45° because I know I will get a better contraction of my chest if my elbow moves across my body rather than in front of it Provide maximum signal to your muscles to grow (form+intensity), give it nutrients to grow (diet) 3. Cardio We staying lean, cardio gives a good buffer Running is literally the worst form of cardio if you’re training for aesthetics. Massive hunger spikes and destroys leg mass. I still do it because I’m a sucker for a 5am run around golden gate park Best form of cardio is fast walking on a treadmill or the stairmaster Aim for 10k steps a day to keep the weight down 4. Aesthetics We want to look good. So prioritize: Chest (incline bench) Abs (hanging leg raises and cable crunches) Shoulders (lateral raises) Quads (quad extensions) Lats (lat pull downs) Everything else is important, but try progressively overload these exercises With the diet and protocol I described, you’ll gain muscle but will feel quite flat. On days you need to look big or take your shirt off, carb up for some extra juice. On days you need to look lean, drink tons of water and don’t eat. 5. Starting off If you’re JUST starting, aim for consistency over intensity until your muscles adapt. Workout so that you can workout the next day. At least for the first 3 months. Disclaimers: Prior to Jan 2025 I had been big before (~2021), so my bounce back was quicker due to muscle memory. I have elite genetics. My dad looks better than me pushing 70 and my mom could be mistaken as a bodybuilder whilst never working out. Your transformation probably won’t be this good, but if you follow these steps it will be optimal.

@codyplof Do you know why Meta has decided to do this? I'm having a hard time understanding it from their perspective.

Not to dunk on Olly but these are actually considered very similiar ads in Meta's system. We recently got an audit from our Meta reps and I was surpised to see what was considered different. It's not just visuals. It's the whole package; visual, copy, offer, and most importantly persona. If Meta doesn't think the changes fundamentally alter the experience or outcomes and thus the targeting, it saves compute and buckets them together. Meta VPs have told me that two ads with the same image and different copy will only enter the copy into the system, and vice versa on these above. So the smaller the change, the less Andromeda is even looking at and the less chance there is of entering new auctions. Also, just because an ad is deemed similiar does not mean there is zero point in making it. It just won't be used to explore new audiences, and will likely be entered into very similar actions as previous ones. That being said, I do fear creative teams are wasting tons of time and money iterating.

@AaronOrendorff For LinkedIn ads, I like to run all campaigns as max bid, then dial the bid down to $.01 on the weekends. Effectively disables delivery without resetting learning afaict

B2B paid f****** sucks on the weekend. I basically drop all budgets just enough to keep ads from turning off. Unless I’m doing something wrong?

That's a great list. I like to think of it in terms of income: - Level A: Your business is profitable enough that you can pay yourself what you could earn in a high-paying job - Level B: You can 10X that personal salary - Level C: You can pay yourself that 10X salary, but personally only work a few hours per week I wrote about it here: docs.google.com/document/d/1-r…

@danielfazio God bless you

Randomly remembered that I’ve been rich for a while now. Price tags have not registered as existing in my psyche for years I don’t even remember what it’s like to feel scarcity Thank you God

@wearetheselect Had no idea about these. Thanks for making this video!

My Meta rep just informed me they've been doing EVEN MORE additional generative AI tests on the back end without advertisers knowing Testing unproven new features on ads were paying for is never great Here's how you can turn them off

@Anvartis1ng But you can avoid that by simply not transmitting the Purchase event value. Right?

@Thedonofstocks From something I wrote for an wanna-be-entrepreneur friend: "It is actually possible to work 100x harder than normal for periods of time. You will probably not believe this until some situation forces you to actually do it"

@dtc_alchemist Congrats!

In Q4 2023 my ecom agency was stuck in a literal hamster wheel, signing 3-4 new clients each month but churning 3-4 old ones. We were so close to 100k months but no matter how hard I tried, we were stuck in the same position and couldn’t break past that barrier. I had two options: Double down on my client acquisition efforts, or take a step back to focus on client retention. I chose the latter. So in Q1 I stopped posting on Twitter which was our main source of clients, and decided to put more of my focus and energy into building the foundations. By Q3 I went down from 15 clients to 5 (combination of churn + letting go clients we didn't want to work with long term) On paper, I was the worse CEO ever because we "lost" 2 out of 3 of our clients. But it was a necessary sacrifice to make. Let me tell you why: My goal has always been to build the best ecom growth agency in the space. To be respected and well known, and have the best case studies and clients. To have BOTH our client acquisition and retention so dialled in, we’d be able to reach $1M/mo in revenue by end of 2026. If sacrificing short term revenue and my ego meant we can fix our foundations, and be in a much better position to achieve our long-term goals, it’s a sacrifice I’m willing to make. As the saying goes: Someone who can’t give up everything will never be able to change anything In the past 6-8 months I’ve worked relentlessly on: → Building systems and processes → Solidifying our organisational structure → Hiring A tier talent → Mapping out the client journey → Giving clients quick wins as soon as they get onboarded → Solidifying our project management and tracking → Being ultra selective with brands we choose to work with → Offering more services to take full control of the customer journey → Adding performance-based pricing for a shared incentive for growth → and more The Impact? Fewer Clients, Better Results, Higher Margins Here’s the thing: growth isn’t always about doing more. Sometimes it’s about doing better. By stripping everything back, we created a business that’s leaner, stronger, and positioned for sustainable growth. The short-term sacrifices hurt, but the long-term payoff will be worth it. We’re no longer chasing numbers—we’re building a legacy. And now, as we head into the next chapter, we’re ready to scale smarter, faster, and stronger than ever before. Because the best businesses aren’t built on quick wins. They’re built on a foundation that stands the test of time.

Highly recommended

@dougdanna 👋

I’m cleaning my following list of bot accounts, if you are active and a REAL person say something! I need to make room for ACTIVE people.

@engineers_feed It will take 14 x 21 minutes. Most of which will be snowball fighting.

⛄️ Three adults can build a snowman in 14 minutes. Three children can build a snowman in 21 minutes. How long will it take all six people working together to build one snowman?

@EricTilbury_RTB One of the great joys of a man's life

Is there anything more dangerous but fun/hilarious than annoying your wife on purpose?

@alz_zyd_ This absolutely works. And it will give you a deeper first-principles understanding of the topic than almost anyone you come into contact with

There is a brute-force learning method which is the following - Take a frontier paper in a field you want to learn but know nothing about - Try to read the paper - Pause every sentence you're confused and look for older papers/books/etc until you understand it

@dtc_alchemist Tough lesson. Thanks for sharing this story.

I once lost the opportunity to work with someone running 7+ brands, all doing 7 to 9 figures in revenue. It wasn’t just a trial project we lost—it could have been the key to scaling our agency exponentially. He ordered a trial batch of 4 video concepts for one of his brands. And while I believe the team did a great job with the creatives, we hit a major roadblock. We couldn’t get access to their ad account for reporting and analysis. It was strange—we’d never encountered that issue before. For some reason ads manager was just not on our side. Regardless, the client gave us access to his Motion account. But I thought “This isn’t how we usually do things”. It wasn’t part of our SOP to use the client’s Motion account instead of ours. I figured it was too much of a hassle to build custom reports in his system, so we tried again to connect to his ad account. But after multiple failed attempts, communication slowed down. A few weeks later, he quietly removed us from Slack. When I followed up, he told me he’d gone with another agency. Now looking back, who the hell cares what the SOP is? If I’d just spent an hour building those reports in his Motion account and asked my Creative Strategist to do the analysis—20 minutes, tops—things might’ve turned out differently. We would have probably signed a monthly deal with the client and gotten the opportunity to work with all seven of his brands, easily adding 50%+ to our agency’s MRR. And we’d have achieved it without having to chase for new clients. That one decision could have altered the entire trajectory of the business. This experience made me really question my decision-making as the CEO. Sometimes, being too rigid about processes costs you more than you realize. It’s not always about SOPs—it’s about finding solutions, being flexible, and meeting the client where they are. But mistakes are part of the journey. The best business owners reflect, learn, and adapt. Every loss makes us sharper and wiser. The key is to keep moving forward, no matter what.

@czue The latter may require a sales call process. That is easier than you think, and you can outsource it. I can tell you how to make that happen, dm me and we'll hop on a call

@czue Raise your prices, brother. Pegasus is a steal at 10X the current price, if not 100X. Not exaggerating Another option: remove some of the stronger features from Unlimited, and sell an enterprise version that keeps them. And charge $10-20k USD/year per production deployment

I've been a little down lately as it's been a slow month, sales-wise, but I got a Slack message from a Pegasus customer who is closing a multi-million dollar raise and told me that "SaaS Pegasus helped change the trajectory of my life" and now everything feels better.