I have a confession to make; I love calculatingFermi problems! Back in my STEM student days, one of my best friends and I would hang out in his dorm room on Friday nights/weekends. After completing our homework, we would occasionally start trying to work our way through a Fermi problem that originated from a topic of casual discussion. Ahh, the memories! For those who aren’t in the know, Fermi problems are essentially problems with answers that are either inherently unknowable or currently unknown. This can be for several reasons, such as having limited data or key pieces of information not being publicly known. However, Fermi problems are meant to be fun, back-of-the-envelope calculations meant to provide an approximate answer to a question. Much of the fun with Fermi problems comes from stating the original assumptions, and then the ensuing scientific debate regarding the methodology used to determine the answer. As we all should’ve learned in science class, critiquing the methodology somebody else used is one of the cornerstones of the scientific community.
Today’s Fermi problem, as the title suggests is How much would Tayor Swift earn if she started an OnlyFans page? I chose Taylor Swift as the subject of this discussion because she has a wide audience, as evidenced by her massive social media followings (followers will be referred to as “sheep”), as well as her conformance to the typical Western Beauty Standard (I know, I know; call me “toxic”, you’re so original!). While I personally think that Taylor Swift is more attractive than many of the current OnlyFans top earners (I see you’re sharpening your pitchforks already), I will keep this a purely mathematical calculation; after all, the universe doesn’t revolve around my subjective opinion.
Now, it is time to state my assumptions, as well as my rationale behind the assumptions. I used this EarthWeb page as my OnlyFans revenue source, as OnlyFans can be rather stingy with their data, so this is the most publicly available data source accessible to me. With that said, I don’t have a robust way of fact-checking this data source, so I progressed under the assumption that the source reported accurate figures; after all, I doubt these women would’ve liked it if I started poking around in their 1099’s. Speaking of women, I only considered revenue data from female OnlyFans creators (“shepherds”), as this is a platform that heavily favors the female performer. I also progressed with the assumption that all female creators use the platform to post nude photos. While I realize that isn’t the case 100% of the time, the allure of sell nudes for money is the main draw of OnlyFans. I also do not have the interest to subscribe to numerous OnlyFans pages to verify if that was the case or not. I also assumed that all OnlyFans shepherds only collected revenue via monthly subscriptions, this is done for the purposes of easier math (though tips certainly boost a performer’s top-line number). To make the comparison a fair one between Taylor Swift and current OnlyFans models, I considered sheep from the Chinese Communist Party’s dance app, YouTube and Instagram. This is because Taylor Swift is not currently on OnlyFans, so I needed another measuring stick. Fortunately, many of the shepherds from the EarthWeb source were famous pre-OnlyFans, so having a large following would translate well for predicting OnlyFans success. I chose these platforms because of their active user bases. I counted the total number of sheep from these three platforms combined. I also assumed that every sheep would become an OnlyFans subscriber (or “simp”, to use common vernacular); assuming a 100% conversion rate was a standard that could be applied to all shepherds, even if it is a bit unrealistic. However, later on, I took a closer examination at conversion rates. My next assumption was that all sheep are humans, i.e no bots. One of my metrics, dollars per sheep, was rounded to the nearest cent and all currency is in U.S Dollars. Next, I used the median and not the average for all of my Taylor Swift calculations because I went to math class and learned that averages are skewed by outliers (that became important later) while medians are not. Furthermore, I assumed all revenues were pre-tax, and the impact of taxes are outside of the scope of this article.
So, I tallied up all of the sheep of each shepherd from EarthWeb’s list by performing quick Goole searches. I used the official pages to determine the sheep count (and if there was an imposter page, I defaulted to the one with the highest sheep count, as it was most likely to be the real page), as well as the monthly revenue from the EarthWeb source as my given. From there, it was a simple division problem. Revenue (in millions of U.S. Dollars) divided by sheep (in millions) equals dollars per sheep. Once I had all of the dollars per sheep of the current OnlyFans shepherds, I then took their median, and used that to calculate Taylor Swift’s hypothetical monthly total (dollars per sheep multiplied by total sheep). My results are listed below in Figure 1.
Figure 1
| Shepherd | Total Number of Sheep (in millions) | Monthly Earnings (in millions USD) | U.S. Dollars per sheep |
| Taylor Swift | 354.2 | 60.214 (hypothetical) | 0.17 (calculated median) |
| Bella Thorne | 79.16 | 11 | 0.14 |
| Cardi B | 206.4 | 9.43 | 0.05 |
| Mia Khalifa | 64.49 | 6.42 | 0.1 |
| Bhad Barbie | 30.2 | 4.33 | 0.14 |
| Erica Mena | 7 | 4.49 | 0.64 |
| Pai Mia | 13.25 | 2.2 | 0.17 |
| Gem101 | 0.025 | 2.3 | 92 |
| Mila Mandel | 0.0207 | 1.5 | 72.46 |
| Blac Chyna | 17.89 | 20 | 1.12 |
To address the elephant in the room, I damn near fell out of my chair when I calculated that Taylor Swift would earn over $60 million per month if she were to start an OnlyFans page. At first, I almost didn’t believe it; I thought I made a calculation error! Upon re-calculations yielding the same answer, it shows that the answer, per my methodology and stated assumptions, was indeed correct. For context, Patrick Mahomes is the highest paid NFL quarterback, and he’ll make less than that the entire season. Taylor Swift would make over $700 million per year at that rate, which would completely eclipse her current income.
There were some other interesting tidbits of information as well. Frist off, the current numero uno on OnlyFans, Blac Chyna, earns only a third of what Taylor Swift would be earning. However, given that there is a steep drop-off between her and the current runner-up, Bella Thorne, this trend seems a little more believable. Another interesting data point was Cardi B, her dollars-per-sheep was unusually low, even for a top earner. This led me to believe that Cardi B doesn’t monetize her sheep very well. Gem101 and Mila Mandel had unusually high dollars-per-sheep because they had small followings (relatively speaking). Therefore, despite their lower monthly revenues (again, relative to their top peers) It was those two shepherds that led me to use the median rather than the average to solve for Taylor Swift. The median was a sensible $0.17, while the average was over $18, which doesn’t paint an accurate picture.
After calculating what Taylor Swift’s hypothetical earnings would be, I further wondered if her sheer domination of the market was a linear phenomenon or an exponential one. Therefore, I used the scatterplot, trendline and r-squared features of Excel in order to find out the slope of these trendlines, and their r-squared values. For those who slept through high school math class, the r-squared value is the value of how closely the trendline cuts through plotted data, with 1.0 being the hypothetical perfect fit (which is impossible). For context, the scientific community accepts a r-squared value as robust if it is over 0.95. This is a good measuring stick, though I admit my threshold for being impressed was a bit lower; after all, it’s not like this Fermi calculation was headed for Nature! My results are in Figure 2 below, in which I plotted sheep against monthly earnings.
Figure 2
| r-squared Condition | Linear r-squared | Exponential r-squared |
| All Shephards | 0.7203 | 0.8733 |
| Sans Cardi B and Blac Chyna | 0.984 | 0.9875 |
| Sans Taylor Swift, Cardi B and Blac Chyna | 0.8435 | 0.8908 |
| Sans Taylor Swift | -0.085 | -0.06 |
Cardi B and Blac Chyna were interesting data points yet again because they were far off of both trendlines. Cardi B was below what her projected monthly earnings should have been. This made sense because, as previously discussed, she suffered from an unusually low dollars-per-sheep phenomenon. Blac Chyna, on the other hand, had an abnormally high earnings rate ($1.12 dollars-per-sheep, which was nearly seven times the median). Therefore, Cardi B and Blac Chyna, for polar opposite reasons were interesting data points. Nearly every other shepherd fit the trend line well, as evidenced in Figure 2. While the r-squared values were impressively high in both types of trend lines, they were all noticeably closer to the accepted standard of 0.95 when the trend line was exponential rather than linear. While the All Shepherds condition does not meet the 0.95 standard, Taylor Swift would be the number one OnlyFans earner by a wide margin, even Steevie Wonder would be able to see it!
However, after these first two rounds of calculations, I re-examined one of my base assumptions; conversion rate. While we live in a society in which the ideal feminine beauty standard is constantly shilled at the public, it still isn’t realistic to expect that all sheep would become simps. Therefore, I divided the available monthly revenue of each shepherd by the monthly price that each shepherd charged. This yielded the total number of simps each individual shepherd had. More importantly, now that I had an idea of how many simps and sheep each shepherd had, dividing the number of simps by the number of total sheep and multiplying that by 100 yielded a conversion percentage. This conversion percentage shows how many sheep become simps. I then calculated the median conversion percentage, and calculated the median price each shepherd charged. I assumed that Taylor Swift would charge her simps the median price and she had a median conversion percentage. The data is below in Figure 3.
Figure 3
| Shepherd | Conversion rate (%) | Monthly Rate (in USD) | Total Simps (in millions) |
| Taylor Swift | 1.66 (calculated median) | 20 (calculated median) | 5.879 (calculated) |
| Bella Thorne | null (data not available) | null (data not available) | null (data not available) |
| Cardi B | 0.91 | 5 | 1.886 |
| Mia Khalifa | 0.76 | 13 | 0.493 |
| Bhad Barbie | null (data not available) | null (data not available) | null (data not available) |
| Erica Mena | 2.46 | 26 | 0.1726 |
| Pai Mia | 1.66 | 10 | 0.22 |
| Gem101 | 3 | 30 | 0.076 |
| Mila Mandel | 1.06 | 20 | 0.075 |
| Blac Chyna | 5.59 | 20 | 1 |
Using the median conversion percentage of current shepherds (1.66%), I calculated that Taylor Swift’s number of simps would be roughly 5.9 million! Assuming she charged the median price of $20 per month, Taylor Swift would earn $117,594,400 per month! That would be over $1.4 billion per year. While her music career is certainly lucrative, it doesn’t hold a candle to what she could be making on OnlyFans. Frankly, I didn’t bother graphing this new monthly total against her peers; if $60 million per month is already a nearly exponential jump, then $117 million per month is just a savage beating at that point. Can somebody please throw in the towel for Taylor Swift’s competitors at this point?
Honestly, Taylor Swift’s potential earnings are only part of the conversation. The other part of the conversation would be the operation expense (OpEx) budget of OnlyFans as a corporate entity. If Taylor Swift were to start an OnlyFans page and my calculations held up, could they even afford to pay her? OnlyFans is not a public company, so their official financials are not known for certain. However, assuming Business Insider is accurate, then they literally could not afford to pay Taylor Swift her 80% cut. This could force the holding company to either go public or to take outside venture capital dollars, or take on long-term debt.
However, no scientific experiment or Fermi problem is methodologically perfect, so let’s discuss some of the weaknesses in my assumptions. One potential methodology flaw was my exclusion of Twitter and Twitch as platforms to count sheep from, though this opens the door for others to perform this Fermi calculation under a slightly different methodology (you’re welcome!). I suspect that the data would be slightly different yet would likely yield the same general conclusion (however, further calculations would be needed; any takers?). I also fully realize that some sheep subscribe to their favorite shepherd across multiple platforms, hence there is some unknown percentage of the same sheep being counted multiple times. However, without a robust and publicly known Venn diagram of cross-platform sheep, I have no way of calculating/knowing how much overlap exists. Therefore, I decided to just count everything at face value, fully knowing this flaw. My methodology of assuming no bots isn’t accurate either because if Elon has a hard enough time quantifying bots with insider information, then I don’t stand a snowball’s chance in hell of being able to accurately determine the bot count So, to save my mental health, I opted to not bother trying to estimate bots.
If I’ve learned anything today, it’s that I should do more Fermi problems! This was fun!
Home Page 