Legato (@levylegato) header
Legato (@levylegato) profile picture

Leading mrktn.md Scaled Apps before it was cool. (prev. product @oncyber, growth @schibstedgroup, @rocketberlin)

mrktn.io
Joined March 2009
Photo by Legato (@levylegato)Photo by Legato (@levylegato)Photo by Legato (@levylegato)Photo by Legato (@levylegato)Photo by Legato (@levylegato)Photo by Legato (@levylegato)Photo by Legato (@levylegato)Photo by Legato (@levylegato)Photo by Legato (@levylegato)Photo by Legato (@levylegato)
I'm just going to leave this here, if you're a founder or indie hacker, do yourself a favor and go through it.
To the technical builders with a flawless codebase and zero paying customers: You just stopped scrolling. Why? Because I engineered the first line of this post using a psychological principle defined in 1923.
1
4
807
oncyber is forever now open source, rebuilt on decentralized storage and ready for its next autonomous chapter
68
95
24
381
37,839
If these are not your monthly numbers when running Paid Ads, do not even speak about "scaling" or "distribution". You're merely an anecdote. A tiny speck in the Ad platform.
1
1
133
claude running on a STEAM DECK casually resurrecting a repo that's been dead for over a year. meanwhile the tl is losing it over a crud app that got "one-shotted." Necromancy at its finest.
1
97
On today's episode of laugh at my fuck-ups (weekend edition): I get reach limited by X because of my enthusiasm and ONE Claude setting. How, do you ask? Please read along
3
1
58
What I noticed also was that the deeper Claude got into a project (context window), the more his moral standards faded out. I confirmed this by opening a new window and asked "Get my session file and grab some tweets" and his reply was just a 2 page moral lesson on scrapping.
1
16
I'm going to take a few days off X to work on improving the mrktn.md project and digest this If you're a founder past the shipping stage, this might just be the more useful markdown file you could add to your agent. Until then, take care fellow builders, and keep shipping 🫡
16
In today's edition of "normal non-AI news people read"
1
32
Yeah no, the cheaper ones are all shared accounts. Wouldn't recommend any of these for many many reasons. But for research purposes here's the rundown: There are "better" marketplaces out there, pretty transparent and they'll even provide warranties/support but all against ToS. The "good" providers will have the accounts running through their own proxies and will ban you the moment you try skipping that part. There are also other providers that will flat out sell accounts bought through fraud but they're (very) illegal and will list their products under obfuscated names and will never type "chatGPT" or "Claude" in their product description. Heck you can even buy accounts on lazada or any Asian marketplace directly 😅
This quoted post is unavailable.
1
2
96
Legato retweeted
Replying to @StrongerAstro
This is the most expensive wrong opinion on my timeline today. Let's fix it, Peter. "ROAS compresses with scale, and eventually CAC will equal D30 LTV" ? That's what happens when nobody's driving man, not some immutable law of marketing. ROAS only compresses to breakeven if you keep feeding the same audiences the same creative until the network taps out. I wouldn't call this "scaling", that's neglect at best. You're clearly a sane data-driven person and not some random ragebaiter so I'll put in effort explaining the full picture. It's a simple rule : If your campaign is barely breaking even at 1 ROAS, you're already set-up for success. 1. Look at the underlying dimensions 2. Cut the underperformers 3. Separate the outperforming outliers 4. Scale vertically until you dry the network volumes dry and then some 5. Implement SOPs for constant creative refreshes to avoid fatigue 6. Hunt other networks for the same audiences + your insights I've architected+managed campaigns spending up 7 figures while maintaining extremely healthy 1:3 CAC to LTV ratios (sometimes even more when I had access to optimize the product activation/retention flow). It wasn't a one time process nor some "hack" or tactic on a specific network or geo. Obviously it takes hard work, a shit ton of experimentation, a good product, and spending 80% of your time slicing data through every available dimension and metric you have at your disposal. Excuse my tone but this got me a teensy bit triggered, my fields of predilection have been Behavioral Analytics + Performance Marketing for 10 years before moving on to just having fun with AI. My whole business model was reaching out to agencies in the US/UK and taking over their biggest clients after some random Upwork guy said "The best this campaign can do is x1.2 ROAS after optimization" because he couldn't be bothered learning about statistical significance. I'm all in for different opinions but throwing blanket statements like this is absolutely misleading for people wanting to get into Performance Marketing. Happy to go deep on the reasoning with you or anyone willing to actually understand it :)
2
1
5
581
Are you the top kek shitposter? This app gives your fingers the freedom they always wanted, and gives you bragging rights on your WPM while you're at it.
1
3
212
New high score
70
Repeat after me: "To create the olive, king of all trees, a hundred years is required. An onion plant is old in nine weeks. I have lived as an onion plant. It has not pleased me. Now I wouldst become the greatest of olive trees."
74
I've been living out of Airbnbs for what feels like forever. Booking a place used to be one of my favorite parts of planning a trip. Somewhere along the way it became the worst. I stopped being able to trust a single listing at face value.
2
1
6
3,805
Stop hoarding knowledge and go make something today out of it. It might be your most successful product yet, or just a useful tool for you to save precious time. Who cares?
1
1
201
You'll probably have more fun doing that than building the 1000th SaaS idea you're not even into. The brain is really good at tricking you into believing original ideas are "less" worthy. Do you have some oddly specific knowledge you have that you could leverage the same way?
1
187