Andrei Shemet (@jammer3k) header
Andrei Shemet (@jammer3k) profile picture

founder @ninja_reach // make anything go viral

Joined January 2018
Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)Photo by Andrei Shemet (@jammer3k)
hi, i’m andrei, founder of @ninja_reach 2011: left professional football at 21 because of injuries with no plan B 2012: a former teammate got me a job selling used pipes, and 6 months in i knew a 9-5 wasn't for me 2013: my girlfriend was working at a video studio and pulled me in to organize shoots and sell projects, and for the first time i actually loved what i was selling 2014: we opened our own little studio on a borrowed computer, and we made ads for local businesses in our city 2015: i launched shema and it grew to $500K a month in sales 2019: we applied to y combinator and got invited to interview, which felt massive at the time 2020: shema got valued at $3M, we went out to raise money, but we were too ambitious and too picky about investors, and nothing closed 2021: by the end of it, we were four people in an office with one month of runway left but trying to sell shema had taught us how to make content that gets people to buy, so when someone in our network asked us to help with their meta ads content in january, we said yes, and that’s how UGC Ninja started 2022: in february we signed our first real client, a $3k contract, with 6 people on the team, zero creators in our network and a 4 week deadline end of 2022: by december we had 150 clients, 30k creators worldwide, 12 people on the team and an average contract of $10K+ 2023: became an official creative partner for TikTok 2024: became an official creative partner for Google 2025: made 14,455 videos for 562 brands, our ads did 5B+ impressions, and we spent 10 months building our first organic campaigns, content that grows brands without paying for ads 2026: bought virality, a clipping company, for $1M, and grew to 500K+ creators and 1,500+ clients across gaming, apps, dating, lifestyle and more, with an average deal of $40K+ today: 500K+ creators, 1,500+ partners and 80B+ organic views shema didn’t work out, but the skill we built trying to sell it turned us into one of the best UGC agencies out there so if you’re running an app, software, a game or an ecom brand and want millions of people seeing it, shoot me a DM
11
8
43
5,440
we had a brand whose Meta ads were off for a while because of a problem with their ad account during that time, their Amazon sales sat at around $10K a day then we turned Meta back on, spending $5K a day and their Amazon sales jumped to around $40K a day a lot of people see an ad on Facebook or Instagram, like the product, and then go buy it on Amazon instead, because they trust the returns and already have an account there so on Meta, it looks like the ad barely made anything but if you look at Amazon at the same time, the ad was actually driving a huge amount of sales this is why we tell every brand with an Amazon store to look at their blended numbers, Meta and Amazon together if you only judge your ads by what Meta reports, you might turn off the ads that are actually doing the most work for your business so before you cut spend because Meta says an ad isn't profitable, check what happened on Amazon that same week
4
4
123
hiring used to mean one person with one set of skills now the candidates i meet often bring a set of small services with them - a script that sorts their inbox - a tool that drafts their weekly report - a helper that checks their code before they send it over each one makes their work faster, and each one is a small team member they carry around so when you hire this person, you’re getting their output plus the assistants that multiply it the practical move is to ask what they’ve built, then look at what those tools would replace on your team some of it will be useful on day one and some of it will show you where your own process is slow
4
9
360
the brands that scale fastest on Meta usually have one thing in common, people buy from them again and again vitamins that show up every 28 days, or a cat fountain that needs new water filters every month the first order might only break even after paying for the ad, but then the refill comes around at the same price, with no ad cost this time so these brands can afford to pay more for each customer, and outbid brands that need all their money from the first order but you need to know what a customer is actually worth before you push the easiest way is to look at everyone who first bought from you 3 to 6 months ago, add up what they've spent since, take off product and shipping, and divide by how many of them there were that's your real number, and if you're too new for that, guess low and update it as orders come in then plan for the cash, because this is where brands with great numbers still run out of money say you pay $55 for a customer and their first order leaves you $40, you're down $15 each until the refills come in at 3,000 new customers a month, that's $45k out of pocket every month before it pays back so scale in steps, let last month's refills help pay for this month's customers, and use payment terms that give you more time, so the bills land after the money does once you know what a customer is worth and how you'll cover the gap, you can push a lot harder than you think
6
1
11
441
if you're spending 6 figs a month on ads, creative is probably costing you more than you think a lot of agencies charge somewhere between $10k and $20k a month for creative, paid upfront for that, you usually get around 30 to 35 minutes of new video a month and you pay it every month, whether those ads work or not the other option is building it in house a good creative strategist alone starts from $4k a month, and that's before you've paid editors, creator fees, and someone to actually launch and test the ads so one hire turns into three or four, and now you're managing all of them on top of running your brand and in both cases, the people making the ads usually aren't the ones running them, so a lot of good ads never get tested properly, they get turned off after one bad day or never get enough budget to show what they can do the brands we work with are already spending, usually around $150K a month, and their numbers already work what they're missing is enough new ads to keep pushing, so every time they try to scale, costs creep up and things stall so we come in on top of what's already working we look at your account, see which ads and angles are already winning, then start adding new ones, new angles, new creators and ads for people who've never heard of you, so there's always something fresh for the budget to go behind we make the ads, we launch them, we test them and we scale the ones that work and instead of a retainer, we take 10% of the ad spend we manage, with nothing upfront while we're testing, we keep it capped at $500 to $700 a day until the new ads hit the target you give us, and only then do we push so if the ads don't work, we don't make money either now, the reason we only work with brands spending around $150k a month or more behind every brand there's a dedicated team, a project manager, editors, our creators and a media buyer running everything every day, plus new videos going out every week at $150k a month, our 10% is $15k, which is roughly what it takes to pay that team properly and keep new ads coming at the pace a growing account needs at $30k a month, 10% would be $3k, less than a single creative strategist costs on their own at that level we'd either have to cut corners on the ads or lose money on every brand, and that doesn't help anyone it has to work for the brand too, so we look for products with enough margin that paying 10% on top still leaves plenty of profit and we're looking to work together for a year or more, not one good month so if you're already spending around $150K a month and your growth has slowed down, you're getting a full team making, running and scaling your ads, paid as a share of your spend instead of a retainer
7
2
16
654
this month several candidates walked into our interviews with tools they built for their own work i’ve sped up a lot with ai over the past few months, and it’s changing who we talk to someone with basic ai and programming skills can get a working tool running in a day, so the strong candidates now show up with their own helpers already in use so now we ask one thing, can you show me something you built for yourself the answers tell us three things - how they break down a messy problem - how they use ai in their daily work - whether they automate the repetitive parts of their job or wait for someone else to do it a resume tells you what someone did → a tool they built tells you how they think if you’re hiring, add this question to your next interview, ask to see it running, not just described
5
1
11
384
there are two problems we hear from almost every brand that books a call with us the first one is, our ads worked fine as we scaled from $50K to $150K a month, but now the numbers have plateaued and start to slip every time we push harder the second one is, we're still growing, but it's slowing down every month, and the new ads we make just aren't catching and a lot of the time, the brand thinks it's a media buying problem, so they hire a new media buyer, try a new campaign structure, and change up the bidding but from what we've seen, about 70% of the time, the problem is the ads themselves if your account was profitable for months and then stopped being profitable, even though the product and the offer is the same, then it's your ads that stopped working, not your product and when you look inside these accounts you can spot the same pattern almost all the ads are made for people who already know everything about the brand the ads usually are reviews, offers, comparisons or before and afters and there's almost nothing made for people who've never heard of them those ads get you to $150K a month, but they can't take you much further because there aren't enough people who already know you think about who actually sees those ads an ad that opens with 30% off or 4.8 stars only makes sense to someone who already wants the product, so Meta shows it to the same small group of people who are already close to buying it works great at first cause those people buy fast and cheap but that group runs out fast, once they've seen your offer 5 or 6 times, the ones who were going to buy already have so Meta keeps spending to show the same ads to the same people, your frequency goes up, your CPMs go up, and every new customer costs you a bit more than the last that's where top of funnel ads come in to save your account they don't open with the product at all, they open with a problem, like a woman who's stopped wearing short sleeves because of her arms, or a guy whose knees hurt every time he climbs a ladder and on Meta, the ad itself decides who sees it, so when an ad opens with that problem, Meta finds the people who have that problem, even if they've never heard of your brand and aren't shopping for anything some of them buy straight away but a lot of them don't, they watch your ad, think about it, check out your page, but now they know who you are and that's where it gets interesting, because now those people are warm so now, when Meta shows them your review ad or offer ad, it actually makes sense to them, and they buy for a lot less than a total stranger would we saw this on one account, the products we barely added any budget to got about 35% cheaper per customer, just because the top of funnel ads were pulling new people into the account so top of funnel ads aren't only bringing in their own sales, they're constantly refilling the group of people your offer ads can close without them, that group slowly empties out and your account stalls, and with them, it keeps getting topped up as fast as you spend that's why getting past $150K usually comes down to a steady flow of ads that reach people who've never heard of you, so the rest of your ads always have someone new to sell to
7
1
1
11
801
Andrei Shemet retweeted
I've built a new AI model that is not allowed to generate text. Instead, it produces a verifiable program that answers your prompt using the original reference. XERJ Reader-1 is a neurosymbolic model for working with data. It points to the exact words in your document, writes a small program over them, and exact code computes the answer: • 20-90x faster • 10-1,000x cheaper (output tokens free) • every answer is a program you can replay The shortest path to automating data work.
239
219
140
772
1,724,645
Meta changed a lot of things in ad accounts in the last few weeks, and a few of them are switched on by default if you’re about to spend big on Black Friday, these are worth checking first: on October 28 Meta turned on a feature called product browsing in existing ads, which lets people scroll through your products right inside the ad Meta says it got advertisers about 3% more conversions on average in their test, so it’s probably worth keeping on, but you should know it’s there on October 19 they added an option to pick an audience for each ad creative, and it’s on by default in Sales campaigns it can be useful, but it also means Meta might be showing different ads to different people in ways you didn’t set up on October 27 Ads Manager started showing results for each image and video separately, which is actually really helpful, because now you can see exactly which creative in an ad is doing the work and Meta’s AI creative tools like background generation, touch-ups and auto translations can now be turned off for each ad, which matters because brands have been finding out Meta’s AI changed their ads without them knowing the reason all of this matters is that ads are already expensive US Facebook CPMs were $21.45 in October, about 35% higher than last year, and they’ve been above last year’s Black Friday prices every month since July so if something in your account is quietly changing your ads or your audiences, Black Friday week is the worst time to find out spend 20 minutes this week going through your ad sets, check what’s switched on, and use the new creative breakdown to see which ads are actually bringing in sales before you put more money behind them
14
36
1,710
before you pay anyone for UGC views, work out what a view is actually worth to you because you can keep paying creators for months without knowing if the videos are even making you any profit here are the 4 numbers you should calculate first, your click rate, how many people who watch the video actually end up on your site, let’s say 0.5% then your conversion rate, how many of those visitors buy something, let’s say 2.5% then your first order value, let’s say $45 and last, how much of that you keep after product and shipping, let’s say 60% now say a creator’s videos get you 1 million views 0.5% of those people click, so 5,000 people land on your site 2.5% of them buy, so that’s 125 orders at $45 each that’s $5,625 in sales, and after product and shipping you keep $3,375 so with these numbers, the most you can pay is $3.38 per 1,000 views if you want to break even on the first order pay $1 per 1,000 and you keep about $2,375 on every million views now here’s where it gets interesting with UGC most creators charge per video, not per view, so you have to turn their price into a cost per 1,000 views yourself say a creator charges $200 for a video if it gets 50K views, you paid $4 per 1,000, which is over your limit, so you lost money on that video if it gets 200K views, you paid $1 per 1,000, and that same video made you money You're paying the same creator the same price but getting a completely different result that’s why you should judge creators on what you pay per 1,000 views and how many clicks they send, not on how good the video looks and watch how fast your limit moves if only 0.2% of viewers click instead of 0.5%, it drops to $1.35 if 30% of the views are bots, it drops another 30%, because bots don’t buy anything it moves the other way too, if your customers come back and order again, every view is worth more and you can afford to pay more for each one so track clicks for every creator, check the views are coming from real people, and rerun this every month with your own numbers
5
16
570
we took a DTC brand from $900K a month in ad spend to $6.7M a month in 12 months in that time we got them over 1 million new customers, which was about 2,700 new people buying from them every day some days we were spending over $750K and at that level the biggest bottleneck for scaling was how many ads we were putting out the people who see your ad but don’t buy will see it a few times, get bored of it, and never buy from that ad if someone’s seen the same ad around 4 times and still hasn’t bought, they probably won’t they need new ads, MOFU and BOFU ones that answer whatever question is stopping them from buying if you don’t have those ready, your cost per customer is gonna climb really fast those people who got ad fatigue from that ad convert, so Meta keeps spending to find new people then those new people need to see your ads even more times before they buy that’s why spend keeps going up when you don’t have a consistent machine pumping out top, middle and bottom of funnel ads everyday when an account can’t scale, it’s almost always this - not enough angles being tested - no middle and bottom of funnel ads to close people - and no system for putting out new ads consistently most of the work in scaling a DTC ad account is just making sure there’s always a next ad at minimum you should be testing 5 new ads every single day, we were doing almost 20 to 25 for this brand If you're starting off with 5, just make TOFU ads with new angles every day for the first week then from the second week, make 2 TOFU, 2 MOFU, and 1 BOFU ad each day roughly 50% of your ads should be TOFU, 30% MOFU and 20% BOFU once that’s running smoothly, try to get it up to 15 a day with the same split if you want to check your own account, look at two things: - how old your best ad is - and how many new ads you launched this week if your best ad is over a month old and you’re launching a few ads a week instead of a few a day your account doesn’t have enough new ads to keep growing, and your costs will keep creeping up most teams can’t put out 20 new ads a day on their own, which is the exact problem we solve for brands in 2 weeks we set up the whole system for your brand from the research to the creatives and the production to make sure new ads are going live every single day then our media buyers, who’ve managed 9 figures in ad spend, run the campaigns track every ad, what each customer costs, which hooks are working, what to scale and what to turn off and whatever works feeds straight into the next batch of ads, so the account keeps getting better instead of burning out we have over 20k active UGC creators making content for our DTC brands every single day and we also have an in-house team of 20 people that are pumping out enormous amount of AI creatives every day and we charge just 10% of the ad spend we manage, so the only way we make more is if your ads are working well enough to spend more If you want to scale your DTC brand with us, book a call with us ↓↓
22
1
1
37
1,450
Andrei Shemet retweeted
how did a video of a scared ball of dough get 90M views? the format is one ecom brands should be copying for their ads it starts with a pile of dry flour on a wooden counter then water jumps in, more ingredients follow, and the pile slowly turns into a dough ball with a worried little face a tiny oil character skates around a hot pan, the dough gets fried, and the last shot is a finished hotteok here’s why it works: -> every step of the recipe is a character, so the process is the story and nobody has to explain anything -> it ends on the real finished product, so the payoff is guaranteed -> there’s no host, no voiceover and nobody on camera, the same warm kitchen look runs through every scene the hook stops the scroll, the story holds attention, and the product shows up at the moment people want to see the result how to use this for your brand: -> turn your key ingredient, or the problem your product fixes, into a character -> show the full cycle from raw to finished in one video -> keep one visual style across all scenes -> test it organically first, then put the winners behind ad spend we build ai content systems like this for ecom brands DM me "UGC" and i’ll show you how to build one for your brand
9
8
2
110
5,882
Andrei Shemet retweeted
there’s a turmeric ad that doesn’t mention turmeric until 1:38 it’s from BioRoot Labs, and their name doesn’t show up until 2:12 it’s 4 minutes and 17 seconds long, it’s been the #1 ad out of 1.3K in its category for weeks, and it’s reached 2.4M people in the UK in 27 days and it looks nothing like a supplement ad it starts on a building site, where Bill is working up on the scaffolding and his knees start giving out next he’s at the doctor, finding out it’s bone on bone and he’ll need a new knee in 12 to 18 months for a builder that’s his whole job on the line, so you’re already invested before you even know what the ad is for when turmeric finally comes up, Bill’s saying he already tried it and it didn’t do a bloody thing a lot of people watching have tried turmeric too and felt nothing, so now he’s saying exactly what they think then his coworker tells him why, the turmeric you get at the shop barely gets absorbed, so it never actually makes it to your joints that’s the turn in the ad, because now the viewer has a reason to try turmeric again and right then BioRoot comes in as the clinically dosed one with no fillers the caption under the video opens the same way, saying you’ve heard turmeric helps with inflammation but you’ve probably never felt it do anything they don’t need clever targeting either, it’s shown to everyone in the UK aged 18 to 65, and the story filters the audience by itself if you’ve got bad knees you stay for the whole 4 minutes, and if you don’t you scroll past their page says they’ve sold over 200,000 this is the kind of ad we make at UGC Ninja, start with a person and their problem, and let the product show up as the answer
1
9
256
a supplement brand came to us at launch, spending only $262 a day on ads they had one question, could a single format carry the whole account if we pushed it hard enough so we went 100% video and no statics and carousels, literally nothing to fall back on the plan was to make a lot of videos, let the numbers pick which ones deserve money, and only raise the budget when the ads proved themselves we made 200+ videos in about 6 weeks, and every one was built around a new angle we didn’t re-cut the winners, because new angles reach new people, while a re-cut mostly shows the same people the same ad again every video had to pass a spend threshold before it got more money → 97 made it past $1K → 25 made it past $10K → 4 made it past $50K those 4 spent $75.4K, $65.0K, $60.8K and $50.1K the best one alone reached 1.32M people and got over 595K plays each time a video passed the threshold, that was proof it could handle more money, so the daily budget went up with it that’s how $262 a day turned into $20K a day by day 21 and this is usually where accounts fall apart you find a winner, put all the money behind it, stop testing, and the whole thing crashes when that ad gets burned out we kept testing the whole way through, even at peak spend 42% of the ads running were still brand new tests so when one winner slowed down, the next one was already coming up behind it 43 days in, the account was still holding at $20–22K a day, 83x where it started if you’re scaling, try to keep at least a third of your live ads as new tests that way when your best ad slows down, you already know which one replaces it
4
1
13
1,185
we’ve paid a lot of creators over the years, and the biggest money drain isn’t bad content, it’s fake views a creator sends you a screenshot of 2 million views, you pay them, and then nothing happens, you get no clicks, sales, or followers because half of it was bots, or it came from countries you don’t even sell in here’s what we check before paying anyone → ask for a screen recording of their analytics, not a screenshot, it’s a lot harder to fake a recording where they’re scrolling through the app live → look at where the audience is, if you sell in the US and 70% of their viewers are somewhere else, those views are worth almost nothing to you → check likes and comments against views, a video with a million views and 300 likes is weird, and comments that just say nice or wow over and over usually aren’t real people → watch for random spikes, an account that normally gets 5K views suddenly doing 800K on your post is worth a closer look and then the boring part that saves you the most start every new creator with a small paid test, check the real stats, and only then scale up keep upfront payments small until someone’s proven they’re legit, because some people take the money and disappear it works the other way too, there are brands that get the content and never pay the creators, which is why the good creators get careful about who they work with
6
18
863
before we plan any content for a brand, the first thing we ask is do you want views, installs or purchases because the content looks completely different for each one, and so does how you judge it if it’s views, we’re doing logo campaigns, your logo sitting on memes and viral videos like celebrity news, streamer clips, cat videos etc + clipping long videos into a ton of short ones that can do billions of views, it’s basically how Stake got their name everywhere but people are watching for the meme, not your product, so don’t expect sales from it what you’re paying for is getting your name in front of a huge number of people for cheap if it’s installs we do slideshows and faceless edits, screen recordings of the app doing its thing a hook up top and the app as the answer fewer views than a meme, but every single video shows what the app actually does so you look at installs per 1,000 views and give every creator their own link so you know who’s driving downloads and if it’s purchases, it’s real creators using the product, showing how it fits into their day and answering the stuff people wonder about before they buy that usually gets the fewest views of all three, but the people watching are the closest to buying, so you judge it on sales per creator, not views the thing we see all the time is brands wanting billions of views and a ton of sales from the same videos then they run a logo campaign and get upset it didn’t sell, or run product UGC and get upset it didn’t go viral and neither one actually failed, they were just measured on something they were never made to do if you want both, run them separately, a logo campaign so people start recognizing your name, and product videos for when they’re actually ready to buy so figure out what you actually want before you make anything, it changes everything about what you post and how you know if it’s working
7
13
1,058
a lot of brands are gonna lose money on black friday and won't realize it until january say your product sells for $100 and costs you $40 to make and ship, so you keep $60 give 30% off, and now you sell it for $70 but it still costs $40, so you keep $30 you just cut your profit per order in half, which means you need to sell twice as much just to make the same money, in the most expensive week of the year to run ads and it gets worse on paid ads, because your break-even ROAS goes up at full price you break even around 1.7x, at 30% off you need around 2.3x just to not lose money, while CPMs are at their highest so here’s how I’d actually run Q4 the offer: raise the order, protect the margin → a gift with purchase instead of a discount, something that feels worth $40 but costs you $8, the customer sees value, your margin barely moves → spend tiers, spend $10 0 get X, spend $150 get Y, pushes the average order up so the ad cost gets spread over more money → bundles, three products for less than buying them separately, still more profit per order than one discounted item → a limited holiday edition that’s gone when it’s gone, scarcity does the work a discount usually does make the offer so good people feel stupid saying no, and you can do that by adding value instead of cutting price the timeline: buy your customers early, cash in late nov 1 to 16 → ads are still cheaper, so spend on content that gets people watching, not buying, and build big warm audiences from everyone who watched most of a video same window → collect emails and phone numbers with early access, a VIP list, a holiday gift guide, every signup is someone you can sell to later for free nov 17 to 23 → open early access to your list first, they buy before the rush and you pay almost nothing in ads for those sales black friday to cyber monday → put most of your budget behind your warm audiences and proven creatives, and hit your list hard with email and SMS december → CPMs usually drop after cyber monday, so keep running, push gift shipping deadlines, and sell your new customers their second order the real profit is the second order if the first order breaks even, that’s fine, as long as you’ve got emails, upsells and a reason for them to come back in january that’s where Q4 actually pays off Kiyoko Beauty did this and made 43% of its whole store revenue from email during black friday last year and its owned-channel revenue grew 3.4x
9
2
1
15
1,563
We've scaled 8 ecom brands to 7 figures in the last 3 months If you want us to do the same for your brand this Q4, then book a call with us: ugc.ninja/andrew
1
164
Steal a Brainrot peaked at 25.4M people playing at the same time, and the way it got there is something any game can learn from it puts players against each other, and that creates reactions, someone losing their best item, a revenge raid, a last second save reactions are what people film, so every server was producing content without anyone asking then the devs scheduled big events on purpose, which gave every creator a reason to post on the same day and flooded everyone’s FYP at once and on august 23rd they did a fake admin war with Grow a Garden’s developer, so both games borrowed each other’s audience in one day, and Roblox hit a record 47.4M players online so it’s conflict that creates clips, events that sync the creators, and crossovers that bring in new audiences that’s the creator system we’ve been building for games at UGC Ninja, and @wizofugc broke down how to set it up in this article
1
1
16
1,509