[Journey] From $0 to… Hopefully Something – My OFM + AI Experiment

Will keep following this journey
Thanks!


Day 208 Update

Another strong day. Revenue keeps moving up, but today was mostly about improving decision speed without letting the system become careless.

Revenue
• Total: $5,792
• Subscriptions: $1,222
• PPV: $3,668
• Tips: $902

Subscribers
• New subs: 371
• Rebill rate: ~96%

Traffic Sources
• Instagram: ~12%
• Reddit: ~74%
• TikTok: ~9%
• Other: ~5%

Content Posted
• IG Reels: 4
• TikToks: 3
• Reddit posts: 43
• IG Stories: 51

Operations

Chatter handled most conversations. I focused mainly on reviewing response timing, buyer intent signals, and where the system is still moving too slowly with users who are already close to buying.

Observations

Biggest takeaway today:

Fast decisions beat perfect decisions when the user is warm.

This does not mean rushing everything.

But when a buyer is already showing intent, waiting too long or overthinking the reply can kill momentum.

Some users need warming.

Some users need patience.

But some users just need the right offer placed cleanly at the right moment.

Another thing I noticed:

Not all “interested” users are equal.

Some users ask questions because they are close to buying.

Some ask questions because they want free attention.

The difference is usually in their behavior, not their words.

Do they open paid messages?
Do they come back after offers?
Do they react to specific content?
Do they spend small before bigger?
Do they move forward or just keep chatting?

These signals matter more than what they say.

On the monetization side:

PPV is still carrying the biggest part of growth.</b>

The reason is not just better offers.

It is better timing, better filtering, and less attention wasted on users who were never serious.

That makes the good buyers easier to spot and handle properly.

Reddit continues to dominate buyer quality. The intent is still the strongest there, but I am keeping the scaling controlled because this channel needs to be protected, not abused.

Another interesting point:

The best workflow is one that removes hesitation without removing judgment.</b>

I do not want the chatter guessing everything from scratch.

But I also do not want the chat to become robotic.

The goal is simple rules that help the chatter move faster while still making the conversation feel human.

New Focus / Tests

• Improve speed on warm buyers
• Reduce hesitation in high-intent chats
• Keep filtering low-value users faster
• Protect Reddit traffic quality
• Continue improving buyer notes without overcomplicating them

The system keeps growing, but the main focus now is speed with control. Move faster on the right users, slower on risky decisions, and stop wasting energy on noise.
 
Do you use reddit for searching for models and is it better then twitter?
 
Do you use reddit for searching for models and is it better then twitter?
I use Reddit more for traffic/buyer intent than for finding models.
For model sourcing, Twitter/X can be better because creators are already promoting themselves there and you can see how they present, how consistent they are, what kind of audience they have, etc. But it is also very noisy. A lot of time-wasters, fake interest, people who disappear, or creators who already have bad habits.
Reddit can work for finding models, but I don’t think it is the best main source. It is better for understanding niches, communities, angles, and traffic behavior. For actual outreach, I prefer sources where the creator is already clearly active and easy to qualify.

The most important thing is not the platform though. It is filtering. You need someone consistent, verified, realistic, and willing to work. A model from Twitter can be terrible and a model from Reddit can be great, or the opposite. The source matters less than the screening.
 
Day 209 Update

Another strong day. Revenue keeps moving up, but today was mostly about improving how we separate real buyers from users who only create activity.

Revenue
• Total: $5,906
• Subscriptions: $1,246
• PPV: $3,742
• Tips: $918

Subscribers
• New subs: 378
• Rebill rate: ~96%

Traffic Sources
• Instagram: ~12%
• Reddit: ~74%
• TikTok: ~9%
• Other: ~5%

Content Posted
• IG Reels: 3
• TikToks: 3
• Reddit posts: 43
• IG Stories: 52

Operations

Chatter handled most conversations. I focused mainly on reviewing low-intent chats, response speed for warm users, and where we are still spending too much energy on people who are unlikely to buy.

Observations

Biggest takeaway today:

Activity can lie to you.

A busy inbox does not always mean the business is moving forward.

Some users reply fast.
Some ask a lot of questions.
Some keep the conversation alive for hours.

But if they never open paid messages, never take small steps, and never move closer to buying, then they are not really valuable.

They are just noise with good response time.

Another thing I noticed:

Real intent usually shows through action, not conversation length.

The best buyers are not always the ones who chat the most.

Some spend quickly.
Some come back after seeing the right offer.
Some open paid messages without needing much warming.
Some say very little but still buy.

That is why we are paying more attention to behavior instead of just how “interested” someone sounds.

On the monetization side:

PPV continues to grow because the system is getting better at timing and filtering.

The offer does not need to be forced on everyone.

It needs to reach the right user at the right moment.

That sounds simple, but it makes a big difference when volume gets higher.

Reddit continues to dominate buyer quality. The traffic still has the strongest intent, but I am trying to keep the channel controlled because sloppy scaling can turn a good source into a problem fast.

Another interesting point:

The chatter’s attention is now one of the most important assets.

If that attention goes to weak users, revenue suffers.

If it goes to warm buyers, repeat buyers, and users showing clear intent, the whole day improves.

So the goal is not just better chatting.

It is better allocation of attention.

New Focus / Tests

• Separate real intent from fake activity
• Reduce time spent on long low-value chats
• Improve speed with warm buyers
• Keep Reddit posting controlled and clean
• Continue improving buyer behavior notes

The system keeps growing, but the next level is clearly about reading users faster and protecting attention from fake busyness.
 
Hi Shijin, thanks for documenting the journey in so much detail.

I’m working with one model and my biggest bottleneck is IG/TikTok short-form content. For editing, do you mainly recommend learning mobile editing like CapCut/in-app tools, or is desktop editing enough if posting is still done manually from the phone?

Also, when choosing audios, what is your actual process? Do you pick trending sounds from similar creators, use TikTok/IG trend pages, test several audios per clip, or follow another rule?

For Reddit, do you think verification in larger subreddits is necessary, or can smaller/mid-size niche subs work well enough if the fit is strong? Roughly how many Reddit accounts would you run per model now, and have you ever tested more than 10 per model? Did that help, or did it just create more risk/footprint?

A high-level answer is completely fine. Thank you.
 
Hi Shijin, thanks for documenting the journey in so much detail.

I’m working with one model and my biggest bottleneck is IG/TikTok short-form content. For editing, do you mainly recommend learning mobile editing like CapCut/in-app tools, or is desktop editing enough if posting is still done manually from the phone?

Also, when choosing audios, what is your actual process? Do you pick trending sounds from similar creators, use TikTok/IG trend pages, test several audios per clip, or follow another rule?

For Reddit, do you think verification in larger subreddits is necessary, or can smaller/mid-size niche subs work well enough if the fit is strong? Roughly how many Reddit accounts would you run per model now, and have you ever tested more than 10 per model? Did that help, or did it just create more risk/footprint?

A high-level answer is completely fine. Thank you.
No problem mate.
For editing, I wouldn’t overcomplicate it in the beginning. Mobile editing with CapCut/in-app tools is enough for most short-form content because the content needs to feel native, not like a polished YouTube ad. Desktop editing is fine if you are faster there, but the final result still needs to look like something that belongs on IG/TikTok. Clean cuts, good pacing, captions if needed, and not over-editing usually matters more than the software.
For audios, I mostly look at what similar creators are using, what is showing up repeatedly in the niche, and what fits the model’s vibe. I don’t just pick any viral sound because a huge trend can bring the wrong audience. Better to use sounds that already work in the same type of content. Sometimes I test a few angles with similar clips, but I try not to obsess over audio. Hook, visual, and profile fit matter more.
For Reddit, verification in bigger subs can help, but it’s not mandatory to start. Smaller and mid-size niche subs can work very well if the fit is strong. Sometimes they convert better because the audience is more specific and the post doesn’t get buried instantly.

About accounts, I wouldn’t rush into running tons per model. A few clean accounts with clear purpose and good history is better than 10+ messy accounts that create risk and management headache. I have tested going wider, but it usually creates more footprint, more tracking issues, and lower quality unless the system is already very organized.
High level: start simple, track what actually converts, don’t chase every trend, and don’t scale accounts faster than you can manage quality.
 
Day 210 Update

Another strong day and another nice milestone. Revenue crossed $6k/day, but today was mostly about staying calm and not letting the number change how the system is handled.

Revenue
• Total: $6,024
• Subscriptions: $1,271
• PPV: $3,818
• Tips: $935

Subscribers
• New subs: 385
• Rebill rate: ~96%

Traffic Sources
• Instagram: ~11%
• Reddit: ~74%
• TikTok: ~10%
• Other: ~5%

Content Posted
• IG Reels: 4
• TikToks: 3
• Reddit posts: 44
• IG Stories: 52

Operations

Chatter handled most conversations. I focused mainly on reviewing high-intent users, checking Reddit traffic quality, and making sure we are not letting the strong channel become a single point of failure.

Observations

Biggest takeaway today:

Crossing a milestone does not mean the system is suddenly safer.

$6k/day looks nice, but it also makes the weak points more expensive.

A bad decision costs more.
A missed buyer costs more.
A sloppy traffic channel costs more.
A tired chatter costs more.

The higher the revenue gets, the more important boring discipline becomes.

Another thing I noticed:

Reddit is still the strongest channel, but it needs to be protected.

The buyer intent is very good.

Users from Reddit still convert better, spend better, and come back more often.

But that also creates risk if I rely on it too much.

So I am keeping IG/TikTok alive even if the immediate buyer quality is lower, because backup channels matter before you need them.

On the monetization side:

Repeat buyers are still carrying the strongest part of the growth.

Fresh subs are important, but the real daily stability is coming from users who already trust the page and respond well to clean timing.

At this level, the backend matters more than ever.

Better notes.
Better filtering.
Better offer timing.
Less wasted attention.

These small things are what keep compounding.

Another interesting point:

The main job now is not chasing more chaos.</b>

It is easy to get excited and think the answer is more accounts, more posts, more models, more chatters, more everything.

But more only works if the base is clean.

Messy scaling can make the numbers go up short term while quietly damaging the operation.

New Focus / Tests

• Protect Reddit traffic quality
• Keep IG/TikTok testing active
• Improve handling of warm buyers
• Reduce wasted chatter time
• Prepare backend for higher daily volume

The system keeps growing, but the lesson today is simple: big numbers are nice, but clean execution is what keeps them from falling apart.
 
Hey mate, for Reddit, how are you handling your accounts/ what's the setup you're using? I read your "How we survive with cloud phones post", we're not doing NSFW content but losing a lot of Reddit accounts recently, wonder if you've had a similar experience and still running a similar setup or doing something else entirely with the new Reddit moderation changes. Would appreciate a high level answer and thank you so much for all the posts you've done and everything you've contributed to the community! I feel insanely selfish asking this question, but a reply would make my life 30 times easier right now ahaha
 
Hey man, I have some knowledge that could help you out—can we exchange info? Send me a message. I'm looking for people in the OFM niche who are currently making money.
 
If I'm reading this right, you now have 3 models and you've posted 44 times on reddit. That's roughly 14–15 posts per model per day, and they're bringing in about 75% of your traffic. That's amazing!
 
Hey mate, for Reddit, how are you handling your accounts/ what's the setup you're using? I read your "How we survive with cloud phones post", we're not doing NSFW content but losing a lot of Reddit accounts recently, wonder if you've had a similar experience and still running a similar setup or doing something else entirely with the new Reddit moderation changes. Would appreciate a high level answer and thank you so much for all the posts you've done and everything you've contributed to the community! I feel insanely selfish asking this question, but a reply would make my life 30 times easier right now ahaha
Appreciate the kind words mate
High level, I’m much more boring with Reddit now than people expect. I don’t try to survive by constantly playing setup tricks. The main thing is keeping the accounts looking like real accounts with normal behavior, clear purpose, and not pushing them too hard.

Reddit has definitely become stricter, and I’ve seen accounts get hit for things that probably would’ve survived before. But most problems still come from footprint + behavior: too much posting, too many similar actions, weak account history, bad subreddit fit, same style repeated everywhere, or accounts looking like they only exist to promote.
I don’t want to go deep into device/proxy/cloud phone stuff because that can turn into a rabbit hole fast, but my current mindset is simple: fewer cleaner accounts > tons of fragile accounts. Build them slower, post where the content actually fits, don’t abuse one pattern, and don’t rely on Reddit as the only channel.

If you’re losing a lot of accounts even without NSFW, I’d look less at “what setup trick fixes this” and more at what the accounts are doing before they die. Usually Reddit is reacting to patterns, not just one random post.

If I'm reading this right, you now have 3 models and you've posted 44 times on reddit. That's roughly 14–15 posts per model per day, and they're bringing in about 75% of your traffic. That's amazing!
Yeah roughly, but I wouldn’t look at it as perfectly even per model.
Some models/accounts carry more traffic than others depending on subreddit fit, content angle, and how strong the account history is. So 44 posts doesn’t mean each model gets exactly the same amount of posts or same results.

Reddit is definitely carrying the biggest part of quality traffic right now, but the important part is not just post count. It’s matching the right content to the right subreddit and not forcing volume where the fit is weak.
I’d rather have 10 good Reddit posts that bring buyers than 50 posts that only create noise or risk.
 
Day 211 Update

Another strong day. Revenue keeps moving up, but today was mostly about reviewing channel dependency and making sure the backend can handle more volume without getting sloppy.

Revenue
• Total: $6,143
• Subscriptions: $1,297
• PPV: $3,892
• Tips: $954

Subscribers
• New subs: 392
• Rebill rate: ~96%

Traffic Sources
• Instagram: ~11%
• Reddit: ~74%
• TikTok: ~10%
• Other: ~5%

Content Posted
• IG Reels: 4
• TikToks: 3
• Reddit posts: 44
• IG Stories: 53

Operations

Chatter handled most conversations. I focused mainly on reviewing Reddit performance, checking which traffic is actually turning into buyers, and making sure we are not confusing volume with quality.

Observations

Biggest takeaway today:

A strong channel can become a weakness if you depend on it too much.

Reddit is still producing the best traffic by far.

Better intent.
Better repeat behavior.
Better conversations.
Better buyer quality.

But that does not mean I can ignore everything else.

If one channel is carrying most of the business, that channel needs to be protected and supported by backups.

Another thing I noticed:

Post count does not tell the full story.

It is easy to look at Reddit and think more posts = more traffic = more money.

But that is not really how it works.

Some posts bring profile visits but no buyers.
Some bring fewer clicks but much better spenders.
Some subreddits look active but convert badly.
Some smaller niche subs quietly outperform bigger ones.

So the goal is not just posting more.

The goal is posting where the traffic actually has intent.

On the monetization side:

PPV continues to grow because the backend is getting cleaner.

The offers are not dramatically different.

The improvement is coming from better timing, better user notes, better filtering, and less wasted attention on users who were never serious.

Small backend improvements are still producing big revenue movement.

Reddit continues to dominate buyer quality, but I am keeping IG/TikTok active because depending on one platform too much is dangerous. They are not as strong right now, but they still matter for discovery and backup traffic.

Another interesting point:

The higher the numbers get, the less room there is for lazy execution.

At lower numbers, mistakes are annoying.

At higher numbers, mistakes become expensive.

A slow reply.
A bad note.
A weak follow-up.
A sloppy traffic pattern.

All of it matters more now.

New Focus / Tests

• Protect Reddit traffic quality
• Keep testing IG/TikTok as backup channels
• Improve subreddit quality tracking
• Reduce wasted chatter attention
• Keep repeat buyer handling clean and personal

The system keeps growing, but the lesson today is simple: when something works, protect it instead of abusing it.
 
Day 212 Update

Another strong day. Revenue keeps moving up, but today was mostly about checking whether the current systems are still simple enough to survive more volume.

Revenue
• Total: $6,265
• Subscriptions: $1,322
• PPV: $3,970
• Tips: $973

Subscribers
• New subs: 399
• Rebill rate: ~96%

Traffic Sources
• Instagram: ~11%
• Reddit: ~74%
• TikTok: ~10%
• Other: ~5%

Content Posted
• IG Reels: 4
• TikToks: 3
• Reddit posts: 45
• IG Stories: 53

Operations

Chatter handled most conversations. I focused mainly on reviewing buyer routing, Reddit traffic quality, and where the system still slows down when too many similar decisions happen at once.

Observations

Biggest takeaway today:

Simple systems scale better than clever ones.

This keeps coming back.

Every time I try to make a workflow too smart, it becomes harder to follow under pressure.

The best rules are still the ones that can be used quickly:

Who is warm?
Who is wasting time?
Who needs a soft touch?
Who is ready for an offer?
Who should be ignored?

If the chatter has to think too much on every reply, the system slows down.

Another thing I noticed:

Buyer routing is becoming more important than buyer volume.

Getting more users is not the hard part anymore.

The hard part is making sure the right users get the right attention fast enough.

A good buyer handled slowly can become a lost buyer.

A bad user handled too much can steal time from five good ones.

On the monetization side:

PPV is still the strongest engine, but only because filtering is getting sharper.

We are not just sending more offers.

We are trying to send fewer bad offers and better-timed good ones.

That difference matters a lot now.

Reddit continues to dominate buyer quality. The intent is still strong, but I am watching the footprint carefully because scaling Reddit too aggressively can create problems later.

Another interesting point:

The business feels less chaotic when bad users are removed faster.

Not every user needs to be convinced.

Some users need warming.
Some need patience.
Some need a clear offer.

But some just need to be filtered out before they drain the backend.

That one shift alone makes the whole operation cleaner.

New Focus / Tests

• Improve buyer routing
• Keep chatter rules simple
• Reduce attention wasted on low-value users
• Protect Reddit traffic quality
• Continue improving repeat buyer timing

The system keeps growing, but the main lesson today is simple: scaling is not about handling more chaos. It is about removing chaos before it spreads.
 
Hey Shijin, I've been following your journey and it's really impressive. Motivated me to start OFM myself, your posts have been super valuable!
My main issue right now is finding models. How do you do cold outreach? and is that the best way?
Thank you!
 
Hey Shijin, I've been following your journey and it's really impressive. Motivated me to start OFM myself, your posts have been super valuable!
My main issue right now is finding models. How do you do cold outreach? and is that the best way?
Thank you!
You said you have been following his journey. This implies that you have read the thread. If you have read the thread you will see that there is a ton of information here around this topic. Take some time to read the thread. It's well worth it.
 
Hey Shijin, I've been following your journey and it's really impressive. Motivated me to start OFM myself, your posts have been super valuable!
My main issue right now is finding models. How do you do cold outreach? and is that the best way?
Thank you!
Appreciate it mate, glad the thread helped
For model acquisition, cold outreach works, but only if it’s targeted. Random mass DM blasting is a waste of time and makes you look like every other fake agency account.
I usually look for girls who already show signs of being able to create content: active IG/TikTok/X, comfortable on camera, decent presentation, consistent posting, but not fully monetizing properly yet. That’s a much better lead than someone who only “looks good” but has no consistency.
The opener should be simple and personal. No long agency pitch, no huge promises, no “we can make you rich” nonsense. Just show that you actually looked at her profile, mention a specific angle, and ask if she’s open to growing/monetizing better.
Cold outreach is one way, but not the only way. Referrals, creators already in the space, small influencers, TikTok/IG discovery, and even paid lead forms can work. The real skill is filtering. A bad model will waste more time than a bad traffic source.
You said you have been following his journey. This implies that you have read the thread. If you have read the thread you will see that there is a ton of information here around this topic. Take some time to read the thread. It's well worth it.
Yeah fair point, there is a lot already buried in the thread about sourcing, outreach, filtering, and what kind of models I look for.
But I also get that the thread is long now, so I don’t mind giving a short answer again when someone asks. Sometimes people miss parts or read out of order.
Main idea is still the same: don’t mass DM random girls, target creators who already show consistency and comfort on camera, keep the opener natural, and filter hard before investing too much time. Model quality matters more than just getting replies.
 
Day 213 Update

Another strong day. Revenue keeps moving up, but today was mostly about reviewing the difference between growth that feels exciting and growth that is actually controllable.

Revenue
• Total: $6,387
• Subscriptions: $1,348
• PPV: $4,047
• Tips: $992

Subscribers
• New subs: 406
• Rebill rate: ~96%

Traffic Sources
• Instagram: ~10%
• Reddit: ~75%
• TikTok: ~10%
• Other: ~5%

Content Posted
• IG Reels: 4
• TikToks: 3
• Reddit posts: 45
• IG Stories: 54

Operations

Chatter handled most conversations. I focused mainly on reviewing buyer routing, subreddit performance, and checking where the system still depends too much on manual judgment.

Observations

Biggest takeaway today:

Controlled growth is better than exciting chaos.

It is easy to get addicted to seeing the numbers go up.

More posts.
More traffic.
More buyers.
More chats.

But if the backend cannot handle the extra volume cleanly, growth can turn into a mess very fast.

More revenue is only good if the system becomes stronger with it, not weaker.

Another thing I noticed:

The best users are usually obvious earlier than we think.

A lot of the time, the signs are already there.

They open paid messages.
They come back after offers.
They ask specific questions.
They react to certain content.
They do not just talk forever without moving.

The goal is to spot those signals faster and stop treating every user like they need the same long warming process.

On the monetization side:

PPV crossed $4k today, which is a nice sign that the back-end is still improving.

This is not coming from one big trick.

It is mostly coming from better filtering, cleaner timing, stronger repeat buyer handling, and less energy wasted on users who only want free attention.

Reddit continues to dominate buyer quality. It is clearly the strongest channel right now, but I am still keeping the focus on controlled scaling because Reddit can become fragile if pushed carelessly.

Another interesting point:

The system becomes easier when the wrong users are removed faster.

A lot of stress comes from trying to convince people who were never serious.

When low-value chats get cut earlier, the chatter has more energy for users who actually matter.

That alone improves the whole day.

New Focus / Tests

• Improve early buyer intent detection
• Keep Reddit scaling controlled
• Reduce manual routing decisions
• Protect chatter attention from low-value users
• Keep testing IG/TikTok as backup channels

The system keeps growing, but the main lesson today is simple: growth is only useful if the backend can stay clean while handling it.
 
Appreciate it mate, glad the thread helped
For model acquisition, cold outreach works, but only if it’s targeted. Random mass DM blasting is a waste of time and makes you look like every other fake agency account.
I usually look for girls who already show signs of being able to create content: active IG/TikTok/X, comfortable on camera, decent presentation, consistent posting, but not fully monetizing properly yet. That’s a much better lead than someone who only “looks good” but has no consistency.
The opener should be simple and personal. No long agency pitch, no huge promises, no “we can make you rich” nonsense. Just show that you actually looked at her profile, mention a specific angle, and ask if she’s open to growing/monetizing better.
Cold outreach is one way, but not the only way. Referrals, creators already in the space, small influencers, TikTok/IG discovery, and even paid lead forms can work. The real skill is filtering. A bad model will waste more time than a bad traffic source.

Yeah fair point, there is a lot already buried in the thread about sourcing, outreach, filtering, and what kind of models I look for.
But I also get that the thread is long now, so I don’t mind giving a short answer again when someone asks. Sometimes people miss parts or read out of order.
Main idea is still the same: don’t mass DM random girls, target creators who already show consistency and comfort on camera, keep the opener natural, and filter hard before investing too much time. Model quality matters more than just getting replies.
Thank you so much for the reply! what makes these girls consider replying to my newly created Instagram page?

You said you have been following his journey. This implies that you have read the thread. If you have read the thread you will see that there is a ton of information here around this topic. Take some time to read the thread. It's well worth it.
I have read the thread, and I have learnt a lot but that part still confuses me somehow which is why I thought I'd ask again.
 
I have read the first week and Im impressed. Its also easy to follow and very interesting to see what your thoughts are. Following.
 
so 213 days of work just made you profit of $6,387 ? not so worth it a lot i guess , better to focus on others fields on same time .
 
Back
Top