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

[shijin]

How do you know this?: "TikTok: ~28%".
On OF model's TikTok accounts at most I see "my IG is better" and link to IG. But I never see direct link to OF or linktree.

That % is from the traffic stats inside the OF side, not TikTok analytics
So it means roughly 28% of the incoming traffic was attributed to TikTok
And yeah, I don't rely on a direct OF link in every TikTok bio either. A lot of the traffic goes TikTok → IG/profile → link. It depends on the account/content tbh

best of luck from the onlyfans api team

appreciate it bro

 

[Singularity1]

That % is from the traffic stats inside the OF side, not TikTok analytics
So it means roughly 28% of the incoming traffic was attributed to TikTok
And yeah, I don't rely on a direct OF link in every TikTok bio either. A lot of the traffic goes TikTok → IG/profile → link. It depends on the account/content tbh

appreciate it bro

I asked ChatGPT too, but we didn't managed to figure out how does this work.
How would onlyfans know that the customer came from TikTok? I am doing a little research into this OFM business, if it's worth going into, and in the last few days I saw around 100 tiktok accounts of OF models. I never saw a direct link to OF or bio link tool. At most I see is that they put their IG account in the bio. But in this case the customer goes from tiktok to instagram, and from here to onlyfans (usually by bio link tool in stories). So onlyfans would see they came from instagram. They have no way of knowing that in reality they came from tiktok via instagram.

 

[Virandell]

Hi poppin in for 1 more question , Do I actually need a proxy at all, or can I just run raw mobile data? (I know you said earlier proxy is not necessary but i would want best results from start and I am ready to invest) Is proxy better paired with mobile data or Wi-Fi?
The main issue is I have 1 existing social media account that’s doing really well, so I absolutely cannot risk cross-linking IP/hardware signals or losing it. Want to make sure I isolate the new setup completely without burning my main page. Appreciate any thoughts!
Also I am from UK so I would like target US+UK audience

 

[shijin]

I asked ChatGPT too, but we didn't managed to figure out how does this work.
How would onlyfans know that the customer came from TikTok? I am doing a little research into this OFM business, if it's worth going into, and in the last few days I saw around 100 tiktok accounts of OF models. I never saw a direct link to OF or bio link tool. At most I see is that they put their IG account in the bio. But in this case the customer goes from tiktok to instagram, and from here to onlyfans (usually by bio link tool in stories). So onlyfans would see they came from instagram. They have no way of knowing that in reality they came from tiktok via instagram.

Yeah, you're right actually
If the path is TikTok → IG → OF, OF can't magically know the original TikTok source
The % I mentioned is based on the tracking/attribution setup we use, not just what OF sees from a normal click. So direct attribution and actual source can be two different things
If there's no tracking in between, you'd basically just see IG as the referrer

Hi poppin in for 1 more question , Do I actually need a proxy at all, or can I just run raw mobile data? (I know you said earlier proxy is not necessary but i would want best results from start and I am ready to invest) Is proxy better paired with mobile data or Wi-Fi?
The main issue is I have 1 existing social media account that’s doing really well, so I absolutely cannot risk cross-linking IP/hardware signals or losing it. Want to make sure I isolate the new setup completely without burning my main page. Appreciate any thoughts!
Also I am from UK so I would like target US+UK audience

If that existing account is important, I wouldn't overcomplicate it
I'd keep the main account completely separate and use a different setup for the new one
For a fresh account, normal mobile data is usually enough imo. Proxies aren't some magic boost. I wouldn't mix mobile data + WiFi + proxies just hoping for better results
For US + UK traffic, I'd focus more on the content and posting times than trying to force the location through IP stuff. Keep the setup clean and consistent

 

[shijin]

Day 261 Update
$4,417 today
Pretty much same story as yesterday. Not bad, not amazing either. We're definitely in a different phase now compared to the crazy growth weeks
Revenue
• Total: $4,417
• Subscriptions: $1,104
• PPV: $2,614
• Tips: $699

Subscribers
• New subs: 294
• Rebill rate: ~92%

Traffic Sources
• Instagram: ~36%
• Reddit: ~29%
• TikTok: ~29%
• Other: ~6%

Content Posted
• IG Reels: 3
• TikToks: 3
• Reddit posts: 30
• IG Stories: 42

Operations
Nothing major changed today
Chatter handled most of the conversations and I spent more time checking the numbers and going through some older content
One thing that stood out is how different traffic behaves once it gets inside the funnel
Some sources bring loads of clicks but barely move revenue
Others look pretty small on the dashboard and somehow bring buyers who spend way more
So I'm trying to stop looking at traffic as just numbers
A million views doesn't mean much if the people behind them don't care
Also noticed a few old content formats are starting to work again after sitting dead for a while
Didn't expect that tbh
Might recycle a few of them with new hooks instead of constantly trying to invent completely new stuff
Another thing we're still working on is retention
The first purchase is easy to obsess over because it's the obvious number, but the real difference comes from what happens after
If someone buys once and disappears, there's only so much value there
If they stick around and buy again without being chased constantly, that's a much better signal
We're also being more careful with the team workload
There are still too many little tasks that don't really need a human touching them
Slowly removing those
Nothing exciting, but it adds up
Traffic is still pretty split now
Reddit has less volume than before, TikTok is giving more reach, and IG is still bringing a steady mix
I actually prefer having it like this rather than depending on one source for basically everything
Still below the old peak, but I'm not trying to force a comeback overnight
Right now I'd rather figure out why we're doing $4-5k consistently than randomly hit $7k once and then fall back again

New Focus / Tests

• Rework a few old content formats
• Keep tracking repeat buyers
• Cut pointless manual tasks
• Test new hooks instead of reinventing everything
• Keep diversifying traffic
• Watch costs closely
Slow phase, but there's still plenty to learn from it.
260+ days in and somehow I'm still finding new stuff to fix

 

[Virandell]

Was you looking into hybrid ai or are you doing it already ?

 

[BotScale_AI]

That's really impressive. If I have the chance, I'd love to chat with you.
Does your CRM process involve any human intervention, or does it rely entirely on AI?

 

[shijin]

Was you looking into hybrid ai or are you doing it already ?

Yeah, already doing a bit of hybrid AI
Nothing crazy tho. AI helps with drafts, sorting chats, ideas etc, while humans handle the important conversations. Still works better for us than going 100% AI imo

That's really impressive. If I have the chance, I'd love to chat with you.
Does your CRM process involve any human intervention, or does it rely entirely on AI?

Yeah, there’s still human input
AI handles a lot of the boring CRM stuff, but I don't let it run everything on autopilot. Humans still check things and make the calls when it actually matters

 

[shijin]

Day 262 Update
Another pretty normal day tbh
Revenue moved a little, but nothing crazy
At this point I’m more interested in whether the system stays healthy when a day is slower
Revenue
• Total: $4,352
• Subscriptions: $1,087
• PPV: $2,576
• Tips: $689

Subscribers
• New subs: 287
• Rebill rate: ~92%

Traffic Sources
• Instagram: ~36%
• Reddit: ~29%
• TikTok: ~29%
• Other: ~6%

Content Posted
• IG Reels: 4
• TikToks: 3
• Reddit posts: 30
• IG Stories: 42

Operations
Chatter handled most of the conversations again. I spent more time looking at the stuff that usually gets ignored, like slow follow-ups, dead conversations that keep getting revived, and where the team is spending time without much coming back
One thing that stood out today is that a “busy” day can still be a bad day
There were plenty of messages coming in, but some of them were basically noise. If the team reacts to everything with the same urgency, the important users get buried
So I’m pushing more toward simple prioritization. Who actually bought, who is close to buying, who keeps coming back, and who is just talking.
Also still watching traffic pretty closely. TikTok has been taking a bigger share lately, but the quality is different from IG and Reddit. More views doesn’t automatically mean more money, which sounds obvious but is easy to forget when looking at the numbers lol
Retention is still doing a lot of the heavy lifting. The users who already know the creator and have bought before are way easier to work with than constantly trying to squeeze more out of cold traffic

Another small thing I changed today was cutting some pointless follow-ups. If someone has ignored everything for weeks, sending another “hey just checking in” message usually isn’t some secret revenue hack. Sometimes you just leave it alone

I’m also trying to keep the CRM useful without turning it into some giant complicated machine. More tags and more automation sounds good until nobody actually uses half of it

New Focus / Tests
• Keep cleaning up low-value CRM work
• Track traffic quality, not just traffic volume
• Improve repeat buyer handling
• Keep the team focused on users with real signals
• Avoid adding complexity just for the sake of automation

Still a bit lower than the peak days, but honestly I’m learning more from this phase than I expected. The boring stuff is where a lot of the leaks are hiding

 

[shijin]

Day 263 Update
Today was a bit better
Nothing insane, but the numbers moved in the right direction and the operation felt cleaner too

Revenue
• Total: $4,427
• Subscriptions: $1,105
• PPV: $2,621
• Tips: $701

Subscribers
• New subs: 293
• Rebill rate: ~92%

Traffic Sources
• Instagram: ~36%
• Reddit: ~28%
• TikTok: ~30%
• Other: ~6%

Content Posted
• IG Reels: 3
• TikToks: 3
• Reddit posts: 30
• IG Stories: 43

Operations
Chatter handled most of the conversations
I spent a good part of the day checking how the team handles users after the first purchase
That’s becoming more important now. Getting someone to buy once is one thing, but what happens after that is where a lot of the money seems to be
I noticed a few cases where the first purchase went fine, then the follow-up was just too slow. Nothing technically wrong, just bad timing. By the time we came back to them, the momentum was gone
So we’re trying to make the post-purchase flow a little tighter without making it feel automated or pushy
Another thing I’m watching is the difference between active users and valuable users
Some people can send 50 messages and buy nothing. Others barely say anything and spend within minutes. If you only look at chat activity, you can easily prioritize the wrong people
Traffic is still mixed. TikTok is growing, Reddit is a little softer than before, and IG is pretty stable. I’m not trying to force one platform back to its old numbers though. Better to let each source do its job and keep the overall funnel healthy
Also spent some time cleaning old CRM tags. There was way too much junk in there from older tests. Half the tags weren't even useful anymore
Small thing, but these little cleanups actually make the team faster

New Focus / Tests
• Tighten post-purchase follow-ups
• Separate active users from actual buying signals
• Keep testing TikTok traffic quality
• Clean old CRM data and unused tags
• Keep retention as the main focus

Nothing flashy today. Just another one of those days where a few small fixes add up.

 

[shijin]

Day 264 Update
Pretty steady day overall
Revenue dipped a little, but honestly nothing that changes the bigger picture
I’m still mostly focused on making the current level easier to maintain
Revenue
• Total: $4,361
• Subscriptions: $1,089
• PPV: $2,584
• Tips: $688
Subscribers
• New subs: 288
• Rebill rate: ~92%

Traffic Sources
• Instagram: ~37%
• Reddit: ~28%
• TikTok: ~29%
• Other: ~6%

Content Posted
• IG Reels: 4
• TikToks: 2
• Reddit posts: 31
• IG Stories: 42

Operations
Chatter handled most conversations again. I spent more time looking at the slower parts of the workflow instead of trying to squeeze more activity out of the team
One thing I keep seeing is that small delays matter more than I thought
A user can be interested, then get a reply way later and suddenly the whole thing feels cold. Nothing dramatic happened, we just missed the moment
So we’re testing slightly better timing around active buyers and recent purchasers, while leaving the colder users alone more often
Also noticed that some old conversations are still getting too much attention. Just because someone replied once doesn’t mean they’re worth 30 more messages
The better signal seems to be actual behavior. Did they buy? Did they come back? Did they respond when there was a clear offer? That tells us way more than message count
Traffic was pretty balanced today. TikTok is still bringing more volume than before, but the quality varies a lot depending on the content. Some posts bring tons of views and basically nothing else
That’s another reminder that raw reach can be pretty misleading
On the retention side, repeat buyers continue to be the easiest part of the funnel to work with. They already know the creator, so there’s less explaining and less friction
New Focus / Tests
• Improve response timing around high-intent users
• Cut unnecessary follow-ups
• Watch actual buying behavior instead of chat activity
• Keep testing TikTok without chasing vanity metrics
• Keep the retention side simple
Still in that slower phase, but I’m kinda starting to like it. There’s less pressure to make every single day bigger, and more room to actually fix stuff

 

[shijin]

Day 265 Update
Another pretty steady day
Revenue came back up a bit, but the main thing I noticed today was how much easier it is to spot problems now compared to a few months ago

Revenue
• Total: $4,438
• Subscriptions: $1,108
• PPV: $2,627
• Tips: $703
Subscribers
• New subs: 291
• Rebill rate: ~92%
Traffic Sources
• Instagram: ~37%
• Reddit: ~28%
• TikTok: ~29%
• Other: ~6%

Content Posted
• IG Reels: 3
• TikToks: 3
• Reddit posts: 31
• IG Stories: 43

Operations
Chatter handled most conversations again. I mostly checked where time was being wasted and whether the team was actually following the priorities we set
One thing that stood out was how often small decisions slow everything down
Which user gets answered first. Whether to follow up now or later. Whether a conversation is worth continuing
None of these are huge on their own, but when you have hundreds of conversations, those little decisions add up fast
So I’m trying to make more of them obvious instead of relying on someone’s mood or judgment every time
Another thing I noticed is that good retention is kinda boring
It’s not one crazy message or some magic offer. It’s just consistently being there, having good content, replying properly, and not annoying people with constant follow-ups
The users who stick around seem to respond better to that than anything overly aggressive
Traffic was pretty similar again. TikTok is holding around the same level, while Reddit has been a little softer than it was during the peak. Not worried about it yet. I’d rather have three decent sources than depend on one platform carrying the whole business
Also spent some time looking at older content that still brings traffic. A few pieces I almost forgot about are still sending people into the funnel, which is pretty funny
Sometimes the stuff you think is dead just keeps working quietly

New Focus / Tests
• Turn common team decisions into simple rules
• Keep improving retention without overworking users
• Review older content for hidden traffic
• Keep diversifying acquisition
• Watch margins instead of obsessing over daily revenue

Still not back at the old peak, but the business feels a lot more durable now. That matters more to me than having one huge day

 

[johnsmith11bb]

did u use aged accounts or all fresh
totally how many acc u manage per platform
do u use any tools to connect chatters

 

[shijin]

did u use aged accounts or all fresh
totally how many acc u manage per platform
do u use any tools to connect chatters

Mostly a mix, not all fresh. I don’t run a crazy number of accounts either, just a few per platform and keep them manageable. For chatters we use a CRM/shared inbox setup so everyone can see the same context, nothing super fancy tbh

 

[johnsmith11bb]

Mostly a mix, not all fresh. I don’t run a crazy number of accounts either, just a few per platform and keep them manageable. For chatters we use a CRM/shared inbox setup so everyone can see the same context, nothing super fancy tbh

so u pay for not tools external
sounds cool

 

[shijin]

so u pay for not tools external
sounds cool

Yeah, I pay for some external tools for sure lol. Nothing crazy tho, mostly stuff that saves time with CRM, analytics and team workflow. The expensive part is usually the people, not the tools tbh.

 

[shijin]

Day 266 Update

Pretty normal day again
Revenue is still hovering around the same range, which honestly is starting to feel like a win after the drop from the peak

Revenue
• Total: $4,402
• Subscriptions: $1,099
• PPV: $2,608
• Tips: $695

Subscribers
• New subs: 289
• Rebill rate: ~92%

Traffic Sources
• Instagram: ~37%
• Reddit: ~27%
• TikTok: ~30%
• Other: ~6%

Content Posted
• IG Reels: 4
• TikToks: 2
• Reddit posts: 30
• IG Stories: 43

Operations
Chatter handled most of the conversations. I spent more time checking the team workflow today and noticed something kinda funny
We have way more systems now than we had earlier, but sometimes the simplest stuff still gets missed
A follow-up sitting too long. Someone not checking the right tag. A buyer getting treated like a normal lead because the context wasn't clear
Nothing major, but those little mistakes are exactly what I’m trying to remove now
I also started looking more closely at how much time we spend managing users who probably aren't going to buy
It’s easy to think every reply is valuable because there’s activity happening. But activity and revenue are obviously not the same thing
So the team is getting a bit more freedom to leave low-signal conversations alone instead of trying to revive everything
On the traffic side, TikTok had another decent day for volume. Still watching the quality though. Some traffic looks great in analytics until you actually see what happens after the click
Reddit is softer than the old peak, but the buyers coming through are still pretty solid. IG remains more stable and useful for keeping the funnel filled
Retention is still doing most of the heavy lifting. Repeat buyers don't need much hand-holding, which makes them way easier for the team to manage

New Focus / Tests
• Clean up small workflow mistakes
• Reduce time spent on low-signal chats
• Keep checking TikTok traffic quality
• Improve buyer context inside the CRM
• Keep retention stable without adding more complexity

Nothing crazy today. Just slowly making the whole thing less messy, which tbh is probably the most useful work right now.

 

[shijin]

Day 267 Update
Pretty steady day again
No big spike, no disaster either lol
At this point I’m actually more interested in how stable the business stays than chasing another random peak

Revenue
• Total: $4,347
• Subscriptions: $1,086
• PPV: $2,574
• Tips: $687

Subscribers
• New subs: 285
• Rebill rate: ~92%

Traffic Sources
• Instagram: ~38%
• Reddit: ~27%
• TikTok: ~29%
• Other: ~6%

Content Posted
• IG Reels: 3
• TikToks: 3
• Reddit posts: 31
• IG Stories: 44

Operations
Chatter handled most of the conversations again. I spent more time looking at the handoff between shifts today
One thing I noticed is that even good systems get messy when information isn't passed properly. One person knows the user is a repeat buyer, next person sees an old tag and treats them like a fresh lead. Small stuff, but it changes the whole conversation
So we’re cleaning that part up a bit. Nothing fancy, just making the important context easier to see
Also still cutting down on pointless reactivation. We used to spend way too much time trying to bring every dead conversation back to life. Some users are just done. No need to chase them forever
The interesting part is that the team actually seems less stressed since we stopped doing that. Less random work, more attention on people who are actually showing buying signals
Traffic was pretty normal. TikTok is still bringing decent volume, but Reddit users continue to be stronger when it comes to actual spending. IG is somewhere in the middle and keeps giving us a steady flow
Retention is probably the biggest thing keeping the current numbers where they are. New traffic goes up and down, but repeat buyers smooth out a lot of that volatility
I’m also trying not to add another tool every time we find a small problem. That was a mistake earlier. Sometimes the fix is just changing the workflow, not buying another piece of software
New Focus / Tests
• Improve handoffs between chatter shifts
• Make repeat buyer context easier to see
• Stop wasting time on dead conversations
• Keep comparing traffic quality across platforms
• Fix workflow problems before adding more tools

Still in the boring phase, but honestly this is probably the part that makes the business actually durable

 

[johnsmith11bb]

can u plz write a guide if someone would do it from 0 and in this time
how could he start with low cost
and what steps to folllow and what issues most likely gonna face

 

[Yupwork]

Just checking, cool man, as always said I respect the Consistency