Freelancing and productised services

The research was done by AI agents that open web pages. Some sites refuse them; where that happened we say so. The small numbers point to the source list at the end. Where a figure is our own sum or guess, it is marked “our estimate” and the basis is given.

The claim

These are examples of the claim, copied from our collection of 14 tweets on this scheme¹. View counts are as collected on 6 October 2026. Dates are worked out from each tweet’s ID number. The page judges the scheme, not the people who tweeted.

“I Quit My 9-5 & Went From $0-$10,000/Month in Just 90 Days With Freelancing” (first line; the tweet continues)

@Mike_Scully_, 14 September 2023, 123,571 views¹

“If I lost everything and needed to make $5k-$10k per month with AI starting today here’s what I would do: 1) Go to Fiverr and find manual tasks people are charging decent money for. I had one in mind - Keyword Research. I looked at 5-star reviews of Fiverr Pros and found a few profiles that are charing $100-$200+ for this job” (truncated; this is the first of seven steps, and the rest are described below)

@boringmarketer, 8 May 2025, 151,286 views¹⁷⁵

“If you’re unemployed, research TEACHING ENGLISH as a FOREIGN LANGUAGE. Learn how to teach English online, to foreigners. “Easy” way to make money online & create a job for yourself.”

@penuelist_, 30 August 2023, 381,729 views¹

“Starting your own business in Web3 is easier than you think. (And you can make $10k per month with only 5 clients)” (truncated)

@AlexIsBuilding, 7 February 2025, 5,460 views¹⁷⁶

Web3 is a name for the crypto industry. Fiverr Pro is a programme Fiverr runs for some of its sellers³.

The collection is small: 14 tweets with about 802,000 views in total (our sum)¹. Nine of the 14 state an income figure. None states an hourly rate, the hours worked, a platform fee, how many pitches it took to win a client, or how long the first client took, and none links to proof of earnings in its text¹. We read the full text of two of the tweets through a public copy of X⁷⁵⁷⁶ and the collected text of the rest. Images, videos and threads were not viewed, so some of this may be shown there.

Smaller accounts in the collection make related claims. One says a seller made $30,000 in a month selling ChatGPT services on Fiverr. One says 300 automated reports a month bring in $8,400. Two say a program read 5,137 open jobs and found three niches where clients pay and almost nobody bids¹.

What the scheme is

A freelancer sells work to clients one job at a time instead of being employed. The work can be writing, design, programming, marketing, bookkeeping, teaching, or anything else a business or person will pay for.

There are two main ways to find clients.

Marketplaces. Fiverr and Upwork are websites where clients and freelancers meet. On Fiverr the freelancer posts a fixed offer, called a gig, and waits for buyers. On Upwork clients post jobs and freelancers send proposals (bids). Tutoring sites such as Cambly and Preply work the same way for lessons. The marketplace handles payment and keeps a cut.

Direct clients. The freelancer finds clients through contacts, referrals, their own website or social media, and bills them directly. Many do this on a retainer: a fixed monthly fee for ongoing work.

A “productised service” is a service sold like a product: a fixed scope at a fixed price, such as “one SEO audit for $300”, instead of an hourly rate. SEO means search engine optimisation, the work of getting a website to rank in search results.

The tweets describe three versions:

1. The plain version. Sell a skill, get to $10,000 a month, in one case within 90 days¹. These tweets do not say where the clients come from.
2. The AI version, which comes in two forms.
   • Build your own tool. The largest account’s tweet says to find a task that Fiverr Pro sellers charge “$100-$200+” for, build a small app that automates it, “do some QA testing, let some folks try it for free”, sell the result as a service with “a proprietary custom workflow”, and later “turn it into a micro-saas”, meaning a small piece of subscription software⁷⁵.
   • Resell raw output. Two small accounts describe selling ChatGPT services on Fiverr, or hundreds of automated reports a month at a low price each¹.
3. Niche versions. Services for crypto projects (ghostwriting, managing paid promoters known as KOLs, “reply guys” who post replies for a project), and teaching English online¹.

Freelancing is real work that millions of people do. The question here is narrower: what the claimed incomes require, and who reaches them.

The arithmetic

An income from freelancing is four things multiplied together: a price, a volume (hours, orders or clients), a win rate (how many pitches it takes to get a client), and what is left after the platform’s fee.

Step 1: what the platform keeps

Fiverr pays the freelancer 80% of each order. Its help page says: “For every order you complete, you earn 80% of the purchase amount, including Gig Extras and tips.”² So Fiverr keeps 20% of the order from the seller. It also charges the buyer a fee on top. Taken together, Fiverr kept 28.0% of what buyers spent in the year to June 2026⁵.

The money is then held: “Funds are subject to a 14-day clearing period before they become available for withdrawal.”³ The same page says programmes such as Top Rated (Fiverr’s highest seller level) and Fiverr Pro “benefit from a shorter clearing period”³. Subscribers to Seller Plus, Fiverr’s paid plan for sellers at $25 or $49 a month, get “7-day payment clearance”⁹. A 48-hour early payout costs 1% and is open only to Top Rated sellers and members of the $49 Seller Plus Premium plan⁹¹⁰. Taking money out costs $0 to $3 depending on the method⁴.

Upwork charges the freelancer a fee that “ranges from 0% to 15% per contract”, fixed when the contract starts¹⁴. Upwork’s annual report says this variable fee replaced a flat 10% in May 2025 and depends on “platform-specific supply and demand factors”¹⁵. Many older guides still quote 10%.

What that means for the claimed figures (our sums):

These are before tax. None of the 14 tweets mentions a platform fee in the text we read¹.

Step 2: what it costs to look for work

On Upwork a freelancer normally spends Connects to bid. Connects are a token Upwork sells: “Connects cost $0.15 (USD) each and are sold in bundles.”¹⁷ The number needed “varies per job” and Upwork prints no typical figure¹⁷. Upwork also gives some Connects away, for example for earning a badge, and says “some freelancers may receive 10 free Connects each month”¹⁷.

The only per-bid figure we have is second-hand. A freelancer quoted by a tool vendor reported 102 proposals for roughly 1,000 Connects⁴¹. That is about 10 Connects, or $1.50, a proposal (our estimate from one person’s account).

How many bids does a job take? Upwork does not publish this. UpHunt, a company that sells an Upwork job tool, reported from 3,596,353 jobs it collected: “It takes about 11 proposals for one hire to happen, market wide, and that ratio swings from 8 to 22 depending on the category.”²² From a separate random sample of 44,863 postings it found that “Only 40% of postings ever hire anyone”²².

Three cautions. The 11 is measured only on jobs that ended in a hire, which UpHunt itself calls a limitation²². Proposals sent to the 60% of postings that hired nobody come on top, so a bidder who cannot tell the two kinds apart needs more. It is a ratio for the whole market, not one freelancer’s win rate. And it is a vendor’s own analysis that nobody independent has checked. A newcomer with no reviews will probably do worse than the market.

On those figures, 11 proposals cost about $16 in Connects (our estimate: 11 x $1.50)²²⁴¹. The cash is small. The scarce thing is proposals that win.

On Fiverr there is no fee to list a gig, but a new seller cannot buy visibility either. Fiverr Ads are open only to sellers at Level 1 and above⁸, and Level 1 requires 5 completed orders, 3 different clients and $400 earned⁷.

Other costs are small. ChatGPT costs $0, $8 or $20 a month depending on the plan²⁹. Seller Plus is optional⁹. So what a newcomer puts at risk is mostly time, not money.

Step 3: the price and the volume

Here is how many hours a month each income needs at different hourly rates (our sums; a month has about 730 hours in total, and a full-time job is about 160):

The first and third rows come from the International Labour Organization (ILO), a United Nations agency. It found: “On freelance platforms, when unpaid time is accounted for, the total average hourly earnings (paid and unpaid) drop to US$7.6, with 64 per cent earning below the average.”²⁴ Unpaid time is the searching, bidding and profile-building that nobody pays for. This benchmark is weak for today’s market. It rests on about 300 freelancers surveyed in 2019 and 2020, across the whole world²⁴. The middle freelancer in it earned less than the average: $7.20 for a paid hour, $5.30 counting unpaid time²⁴.

A newer figure is higher. SnipeWork, which sells an Upwork job-alert tool, tallied the hourly budgets clients stated in 24,421 Upwork job posts over seven days. The averages ran from $14 an hour for admin support to $58 for legal work, with most categories between $21 and $35⁷⁴. These are what clients offered in job posts, not what freelancers earned, and the tally is the vendor’s own.

The table shows the shape of the problem. At ordinary platform rates, $1,000 a month is a realistic part-time load. $10,000 a month needs about $62.50 an hour for 160 billed hours, or $100 an hour for 100. That is about six to nine times the 2019-20 survey average for paid hours ($11.20), and about two to three times what clients now typically offer in Upwork posts (our sums)²⁴⁷⁴.

The same thing by the job, for a Fiverr seller (our sums, using the 20% fee²):

The $28 row is the price implied by the tweet about 300 reports for $8,400 (our sum: $8,400 / 300)¹. The $140 row is the average order of a long-running, top-ranked Fiverr seller described below (our sum from that seller’s own figures)³⁹. The $150 row is the middle of the “$100-$200+” that the build-your-own-tool tweet says Fiverr Pro sellers charge for keyword research⁷⁵. We did not check that price, and we do not know whether one new seller can find 83 such buyers a month.

The “5 clients” tweet is about direct clients, so no platform fee applies: 5 clients at $2,000 a month each is $10,000¹⁷⁶.

Step 4: how much buyers actually spend

Fiverr’s results for the quarter to June 2026 say: “Annual spend per buyer as of June 30, 2026, reached $368”⁵. That is what an average buyer spends on the whole site in a year, across every seller they use, and it includes Fiverr’s fees. With Fiverr keeping 28.0%, about $265 of it reaches sellers (our sum)⁵.

A Fiverr seller who keeps $10,000 a month keeps $120,000 a year. That equals the sellers’ share of everything about 450 average buyers spend on Fiverr in a year. Keeping $1,000 a month equals about 45 (our estimates: $120,000 / $265 and $12,000 / $265)⁵. Real buyers are not average, and one seller never gets a buyer’s whole spend, so in practice it takes more buyers than that, or a few who spend far above average.

On Upwork the figure per active client was $5,230 in the year to June 2026¹⁶. Keeping $10,000 a month after a 10% to 15% fee means billing about $133,000 to $141,000 a year, which equals the whole yearly spend of about 25 to 27 average clients. Keeping $1,000 a month equals about 3 (our estimates)¹⁴¹⁶. Upwork’s figure also counts fees, Connects purchases, memberships and large corporate clients, so the amount reaching freelancers is lower and the true numbers are somewhat worse¹⁵.

One tweet in the collection, promoting a planning tool, says a $10,000 month comes from a few valuable contracts and not from hundreds of cheap sales¹. The sums agree with that structure. On Fiverr-sized budgets they need hundreds of buyers. On Upwork-sized budgets, or with direct clients on retainers, they need a few large ones.

Step 5: how many sellers there is room for

Fiverr’s annual report says that in 2025 “marketplace GMV was $1,073.0 million”⁶. GMV is the total buyers spent, fees included. Sellers received about 72% of it. If all of that went only to sellers keeping $120,000 a year, there would be room for about 6,400 of them (our estimate: $1,073 million x 0.72 / $120,000)⁵⁶. That is a ceiling, not a count, because most of the money is spread across many smaller sellers.

Fiverr says it offers “instant access to hundreds of thousands of freelancers”⁶. It no longer publishes a seller count. The last one we found was “approximately 255,000 active sellers” in its 2019 prospectus¹³. If there are 200,000 or more today, the average seller receives at most about $3,900 a year (our estimate: $1,073 million x 0.72 / 200,000)⁶.

Fiverr’s own ranking system gives a sense of scale. Top Rated requires $10,000 in total earnings, 40 completed orders and 20 different clients, plus a manual review⁷. The amount Fiverr asks for across a seller’s whole history at its top level is what the “$0-$10,000/Month in Just 90 Days” tweets claim as a monthly income¹. Those tweets do not say where the clients came from and do not mention Fiverr, so this is a comparison of size only.

Upwork’s Top Rated badge, which it says marks “the top 10% of talent on Upwork”, requires “$1,000+ in annual earnings” among other things¹⁸. That is a minimum for the badge, not a measure of what the top tenth earn.

What the arithmetic says

$1,000 a month is ordinary part-time work at ordinary rates, once a freelancer is winning jobs. On a marketplace, $10,000 a month needs an expert-level rate of roughly $60 to $100 an hour with a full book of work, or 80-odd jobs a month at $150. It cannot be reached through $28 jobs or tutoring at about $10 an hour without impossible volume.

Most of the $10,000 claims in the collection are about direct clients, not marketplaces¹. We found no data on what direct retainers pay, how many pitches they take, or how long the first one takes. So that route is untested here, not refuted. The “5 clients” tweet states its own conditions: do the work “for free for 2 people” first, then post about it, and “if you are above 2k followers and do this 5x a week, you will get leads”⁷⁶. No source we read shows $10,000 a month from nothing in 90 days as a typical result, by either route.

What people who tried it report

How many profiles ever get work

Neither Fiverr nor Upwork publishes how many of its freelancers earn anything, or what the middle freelancer earns⁶¹⁵. We found three counts, all old.

All platforms, 2020-21. The Oxford Internet Institute estimated 163 million freelancer profiles on online work platforms worldwide and found: “Approximately 19 million of them have obtained work through the platform at least once, and 5 million have completed at least 10 projects or earned at least $1000.”²³ That is about 12 in 100 profiles that ever got a job, and about 3 in 100 that reached ten jobs or $1,000 in total (our sums)²³. The study counts profiles, not people. Abandoned and duplicate accounts are in it, so it cannot tell “tried and did not get work” from “signed up and left”. For someone who bids steadily the odds are better than these numbers, by an amount nobody has measured.

Two platforms, September 2020. The ILO collected figures from the platforms’ own pages. Of 95,813 workers it found registered on Freelancer.com, 73% had earned something and 27% had earned more than $1,000. On Guru: “out of 1.05 million registered workers, only 0.5 per cent had earned any income and 0.1 per cent had earned more than US$1,000.”²⁴ These are lifetime totals. The ILO notes the Freelancer figure looks better partly because that site charges for inactive accounts, which clears out dormant ones²⁴. The Guru count in the ILO’s table is 1,048,575, exactly the largest number of rows a spreadsheet holds, so that data may be cut off; treat the Guru shares as rough²⁴.

Fiverr sellers, 2016. Researchers collected the public pages of 21,767 Fiverr sellers and found: “more than 40% sellers have only earned $5-$100 over the last nine months whereas the proportion of sellers with revenue > $50,000 is only about 0.4%”⁷³. Top Rated sellers in that data averaged $22,413 over the nine months⁷³. This covers sellers with listed gigs and sales, not empty profiles, so it is a fairer picture of people who actually tried. It is ten years old.

All three show the same shape: many earn a little, and a small group earns most of the money.

For comparison, a 2025 survey of 2,616 American adults by Bankrate found: “Median side hustle income earnings are currently at $200 per month”⁴³. The median is the middle person: half earn more, half less. The average was $885⁴³. That survey covers every kind of side work, not only freelancing.

Simple jobs are getting scarcer and cheaper

Fiverr’s results state: “Annual active buyers as of June 30, 2026, were 2.7 million, compared to 3.4 million as of June 30, 2025, a decline of 21.9% year over year.”⁵ Its annual report gives a reason: “the adoption of AI tools in the broader market has also created headwinds for simple and low-skilled services on our marketplace”⁶. In plain words, buyers who used to pay for small simple jobs now do many of them with AI tools themselves.

Upwork’s client count fell too: 785,000 active clients in 2025, down 6%, and 763,000 by June 2026¹⁵¹⁶. Upwork attributes the decline to “the evolving impact of AI on certain categories of freelance work and on new client acquisition and retention”¹⁶. Upwork’s own data adds that simple AI-made work is growing in volume while paying less per job: “generative AI and creative production work saw 90% year-over-year growth in contract starts while per-contract earnings declined 13%”³³.

Academic studies agree that the exposed kinds of work lost ground, but they disagree about who was hurt.

• A study in the journal Management Science found “a 21% decrease in the number of job posts for automation-prone jobs related to writing and coding compared with jobs requiring manual-intensive skills within eight months after the introduction of ChatGPT.”³⁴ We read only the summary of this paper.
• Researchers who studied one freelance platform found that freelancers in the kinds of work most exposed to AI saw “a 2% decline in the number of contracts and a 5% drop in earnings”³⁵. Those are small effects. They were “especially pronounced among experienced freelancers who offered higher-priced, higher-quality services”, and the authors say generative AI “appears to be ‘leveling the playing field’” between beginners and experts³⁵.
• An Oxford study of job posts found the opposite for beginners: “there was a reduction in demand for novice workers”³⁶.

So it is not settled whether AI helps or hurts a newcomer against established freelancers.

Individual reports

Individual accounts with numbers are scarce, and all of them are the person’s own statement.

• A long-running Fiverr seller. A seller wrote on Fiverr’s community forum of averaging “$24,000/year sales per year on average, 170+ orders completed on an average per year” for five years at the Top Rated level³⁹. That is about $2,000 a month in sales. The same post reports about $7,500 from about 60 sales in 2024, after Fiverr changed its level system and the account lost its rank, and a 2026 update says the seller now runs an agency off the platform³⁹. The post says screenshots are attached; we could not see them.
• Current Fiverr forum posts. In posts from late September and early October 2026, sellers with 1,500 to nearly 1,900 completed orders report fewer new buyers since about June. One writes: “Genuine new-buyer inquiries have dropped dramatically.”⁴⁰ On the same board another seller reports that seven months of this year are among their ten best ever, and credits repeat buyers and one large order⁴⁰. We read the board’s previews, not the full threads.
• An Upwork attempt. A tool vendor summarised a forum post by a freelancer who reported “submitting 102 proposals and spending roughly 1,000 Connects over six months without landing a genuine client.”⁴¹ We could not open the original post, so this is second-hand.
• A first year on Fiverr, 2014. The oldest detailed first-year report we found: “I’m up $7490 through 14 months, or a little over $500 a month on average.”⁴² The author runs a website about side income and had an existing audience. The report is twelve years old.
• Experienced developers, 2025. On Hacker News, a discussion site for programmers, in a thread titled “It is practically impossible to find a freelance project these days”, one commenter wrote: “Few people can make a good living from Upwork and Fiverr”⁴⁴. Several said the money is in direct clients and reputation. This is opinion, with no income figures.

Teaching English online

This is the most-viewed tweet’s suggestion¹. It is real paid work with a low ceiling on the platforms we read.

Cambly states: “You make $0.17/min ($10.2/hour) on Cambly and $0.20/min ($12/hour) on Cambly Kids.”²⁵ Forty paid hours a week at $10.20 is about $1,770 a month (our sum). Cambly also says: “We’re temporarily limiting the number of new tutors we welcome to Cambly to protect our current tutors’ schedules.”²⁵

Preply lets tutors set their own price and says: “On average, English tutors charge $15-25 per hour.”²⁶ It adds: “Most popular tutors on Preply earn up to $550 a week”²⁶. That is about $2,400 a month (our sum). A research firm’s older note, using 2021 and 2022 data, put average Preply tutor earnings at $18.30 an hour²⁸.

So the tweet’s modest wording (“create a job for yourself”) holds up. As a route to $5,000 or $10,000 a month, Cambly’s fixed rate makes it impossible there. On Preply it would need about three to five times the typical price with a full timetable (our estimate from the figures above)²⁶. We found no data on tutors who work independently.

What the rules allow now

Freelancing is lawful and the marketplaces exist for it. What the rules restrict is some of the shortcuts.

AI-made work

Fiverr allows AI tools in every category but not raw output. Its standards say: “Freelancers must not deliver generic, unmodified, or reused AI-generated output.”¹² The penalty is spelled out: “the order may be cancelled, the client may receive a full refund, and the freelancer’s account may be permanently suspended.”¹¹ This bears on reselling raw ChatGPT output. A seller who builds a tool and checks and edits what it produces is doing something different, and we saw nothing in these rules that bars it.

For Upwork we could not open the terms. A summary of a 2023 Upwork announcement says freelancers may use any tools, including AI, unless the client forbids it²¹. That is second-hand.

In the United States, a deliverable that is purely machine output may have no copyright protection. The US Copyright Office wrote in January 2025: “our conclusions turn on the centrality of human creativity to copyright”⁷¹. A freelancer who promises a client full ownership of unedited AI output may be promising something that does not exist. That last step is our inference, not the Office’s statement.

Paid replies, engagement and crypto promotion

Writing, design, promotion and community work for crypto projects are lawful when the work is genuine and any payment for promotion is disclosed. The rules below catch deceptive or undisclosed versions. We saw nothing to suggest that any account quoted on this page, or any agency it links to, has broken them.

• Fiverr bans deceptive versions on its site: “Services that create fake reviews or use deceptive methods to generate artificial traffic, engagement, or inflated performance signals are not allowed.”¹²
• A US Federal Trade Commission (FTC) rule from August 2024 bans selling “fake indicators of social media influence, such as followers or views generated by a bot or hijacked account”⁶¹. The FTC’s guidance on endorsements adds: “You could be liable if you play a role in creating or disseminating endorsements”⁶².
• In January 2026 X cut off apps that pay users for posting, a model known as InfoFi. According to the news site Decrypt, X’s head of product wrote: “We will no longer allow apps that reward users for posting on X”, citing “a tremendous amount of AI slop & reply spam”⁶⁵. The action was against those apps’ access to X’s data, not against agencies or freelancers.
• Promoting a crypto token for pay without saying so can break US securities law. In 2022 the Securities and Exchange Commission (SEC) announced that Kim Kardashian, who was paid $250,000 for one post, had agreed to pay $1.26 million without admitting or denying the SEC’s findings. The SEC’s enforcement director said promoters “must disclose the nature, source, and amount of compensation they received”⁶³. That order concerns only the person named in it.
• In the United Kingdom, promoting financial products without authorisation is a criminal offence. In February 2026 the Financial Conduct Authority announced that seven influencers who pleaded guilty to promoting an unauthorised foreign-exchange trading scheme had been sentenced: five were fined between £600 and £3,750, and two were discharged without a fine⁶⁴. That case concerns only the people named in it, and it was about currency trading, not crypto.

Being tied to one platform

• Taking clients off Fiverr to avoid the fee is banned: “Conducting business off-platform directly violates Fiverr’s Terms of Service.”¹² Clients a freelancer finds independently are not affected.
• A closed account delays your money by 90 days. Fiverr: “If your account was permanently disabled, we will send you an email with instructions and a one-time withdrawal link after a 90-day waiting period.”⁴
• Upwork can block permanently, and says: “Any new Upwork accounts you create will also be closed”²⁰.
• Payment is not guaranteed. Upwork’s protection for fixed-price jobs covers only work the client paid for in advance: “If a milestone isn’t funded in advance, we can’t help recover payment later.”¹⁹ A milestone is one agreed stage of a job.
• Preply’s first lesson is unpaid: “The commission for every trial lesson with a new student is 100%.” After that the cut “starts at 33% for new tutors” and falls towards 18% with hours taught²⁷.
• A market can close. In 2021 China ordered tutoring firms to stop “courses taught by foreign-based tutors to teenagers and children in the Chinese mainland”. One firm, VIPKid, had “more than 70,000 tutors in North America alone”⁶⁶.

Tax

A “$10,000 month” is sales, not take-home. In the United States the IRS says: “The self-employment tax rate is 15.3%”, owed on top of income tax⁶⁷. Payment platforms report sellers’ income to the IRS on a form called 1099-K, and the IRS says: “Whether or not you receive a Form 1099-K, you must still report any income on your tax return.”⁶⁸ In the United Kingdom, trading income above £1,000 a year must be declared⁷⁰, and platforms report sellers’ income to the tax authority⁶⁹.

Job offers that ask for money

People looking for online work are a target for offers that are not real jobs. The FTC, which calls these “job scams”, reported losses of more than $220 million in the first half of 2024⁶⁰. Its advice: “Never pay anyone to get paid.”⁶⁰ This is separate from genuine freelancing, but the two are easy to confuse when starting out.

Who makes money from it

The platforms

The platforms earn on every job, whoever wins it.

Fiverr charges both sides. Its “take rate”, the share of buyer spending it keeps, “was 28.0%” for the year to June 2026⁵. Fiverr also sells sellers optional extras, such as pay-per-click adverts and Seller Plus. The revenue line that includes these was $133.4 million in 2025, up 50.9% in a year⁶. That line also includes tools not aimed at freelancers, so it overstates what freelancers pay.

Upwork’s take rate was 18.7% in 2025 and 19.6% in the quarter to June 2026¹⁵¹⁶. Its annual report describes Connects as “virtual tokens that are required for talent to bid on projects”¹⁵. So Upwork can earn from a newcomer before the newcomer wins anything, though it also gives some Connects away¹⁷.

On both sites the platform’s share is rising while the number of buyers falls⁵¹⁶.

The accounts in our collection

We tallied the 14 tweets by what each one steers readers to. Six show nothing for sale in the tweet, and those six hold about 74% of the views (our sum). Three are the same post from one account. One mentions an email system in the profile. One sells a $19 download. Two promote software. One links to an article¹.

For the accounts large enough to name, this is what we saw:

• @Mike_Scully_ posted the “$0-$10,000/Month in Just 90 Days” tweet three times, in 2023 and 2024. Each offered a guide by direct message, asking readers to “Like & Comment “90” to receive it (Must follow)” or close to those words¹. Together the three have about 205,000 views, a quarter of the collection (our sum)¹. When we checked the profile on 10 October 2026, its link led to a sign-up page for a free training on a different subject, AI consulting. The page is titled “Free AI Training — Become a Paid AI Integrator in 2026” and asks for a name and email⁴⁹. It says: “I’ve personally coached over 300 entrepreneurs” and “This is free educational training. We do not guarantee income. Individual results vary based on effort and experience.”⁴⁹ The page shows no price. We did not sign up, and we did not see the guide the tweets offered.
• @boringmarketer’s profile links to a site offering a “Free account · a new walked funnel every week” in exchange for an email address⁵⁰⁷⁵. A funnel here means the chain of sales pages and emails a company uses to sell. No price is shown.
• @AlexIsBuilding’s profile says “I make tech brands go viral on X” and links to an agency site⁷⁶. The site says: “We takeover the timeline for the biggest brands in tech.” It describes campaigns run with “50+ trained content creators” and getting a client’s posts “quote retweeted by quality creators”⁵¹. It lists no prices.
• @penuelist_, the most-viewed tweet, showed nothing for sale¹.

So in this collection, the widest-reaching posts are advice with nothing sold in the tweet, and the account that repeated the income claim today links to an email sign-up for a free training on another topic.

Does Fiverr pay people to recruit sellers? Not on the plans we read. Its affiliate programme pays for bringing buyers: 25% of a new buyer’s first order plus 10% of their other orders for 12 months, and “We pay commissions for purchases only, not for clicks or registrations.”⁵²

Courses, communities and certificates

• Communities. A search for “freelancing” on Skool, a site for paid groups, returned 30 communities on its first page. When re-read on 10 October 2026, the page’s data showed a price for 12 of them and none for 18; earlier readings of the same page counted 13 or 14 as paid. Nine charged monthly, from $7 to $99⁵³. The largest, with 20,396 members in the page’s data, showed no price and is titled “Make $1k-$10k in 30 days”⁵³. We did not look inside any of them.
• Courses. One freelance copywriting course we looked at costs $997 and headlines “up to $6,000 a month… in just 8 weeks”. Its own disclaimer says: “Any financial numbers reference are illustrative of concepts only and should not be considered average earnings”⁵⁴.
• Teaching certificates. For teaching English, the first seller a newcomer meets is the certificate provider. One lists its 120-hour course at “Was: $549.00 Now: $219.60!”⁵⁵. Cambly, for its part, pays its fixed rate without requiring one²⁵.

What regulators have said about income claims

We found no regulator or court case about a freelancing course. Four nearby cases show the standard regulators apply to income claims. Each concerns only the company named in it. The three FTC cases were settled; no court ruled on the allegations. The FTC’s releases do not give the companies’ replies and we did not find them.

• The FTC alleged that Arise Virtual Solutions advertised remote customer-service work at “up to $18/hour” when 99.9% of those who joined earned less than that in base pay. Under a proposed settlement announced in July 2024, Arise would pay $7 million⁵⁶. We did not check the final order.
• In March 2025 the FTC alleged that Growth Cave took “approximately $50 million” selling a programme for building and marketing an online course business (a “digital education program”), and that most buyers “do not make a single sale and therefore do not earn any income”⁵⁷. In January 2026 the FTC announced court orders settling the case, under which the defendants are barred from selling business opportunities⁷².
• In April 2026 the FTC announced a proposed $1.5 million settlement with Publishing.com over income claims used to sell a course on self-publishing e-books and audiobooks, priced up to $1,995. The FTC alleged: “Most consumers who bought Publishing.com’s products and services, however, never achieved the income the company promised in its advertising.” The settlement was a proposal awaiting public comment when announced⁵⁸.
• The UK’s Advertising Standards Authority ruled in October 2025 on an advert for a £390 course that claimed £20,000 to £30,000 a month: “We therefore concluded that the ad misleadingly implied that the stated lifestyle and earnings were typical.”⁵⁹ The ruling records that the advertiser agreed to amend the ad⁵⁹.

The upside

Freelancing does pay, and for some people it pays very well.

The money is real. About $4.0 billion went through Upwork in 2025¹⁵ and about $1.07 billion through Fiverr⁶. Spending per client is rising on both: up 7% on Upwork in 2025¹⁵, and up 15.6% on Fiverr in the year to June 2026⁵. The totals are not rising, though. Upwork’s was roughly flat in 2025 (our sum: clients down 6%, spend per client up 7%)¹⁵. Fiverr’s fell 2.2% in 2025⁶ and was about 10% lower in the year to June 2026 (our sum: 2.676 million buyers x $368 against 3.425 million x $318)⁵. Spend per client rises partly because smaller buyers have left.

Among established skilled freelancers, the middle income matches a salary. A survey of 3,000 skilled American workers, paid for by Upwork, found that “those who earn exclusively through freelance work report a median income of $85,000, surpassing their full-time employee counterparts at $80,000”³². That is about $7,000 a month. It is self-reported and comes from a company that profits from freelancing. The survey also chose its participants “based on their roles, earning levels (per BLS data), and self-reported freelance work”³². So low earners were left out before the middle was taken, and the figure says nothing about freelancers in general.

Six figures is common in absolute numbers. A yearly American survey by MBO Partners says of independent workers: “Almost 6M of them now earn more than $100K per year.”³¹ Press coverage gives 5.8 million out of 74.9 million, up from 3 million in 2020³⁰. That is about 8 in 100 (our sum). It covers consultants, contractors and tradespeople of every kind, not newcomers on marketplaces.

The top of Fiverr. Fiverr gives awards to freelancers who pass $1 million in lifetime earnings on the site, with a higher tier above $5 million. These are confirmed by the platform, not self-reported. It does not say how many there are, and its named examples include “an Australian-based SEO entrepreneur who has grown his business to 100+ employees with Fiverr”⁴⁵. At that level it is a company, not one person.

Two press profiles. CNBC profiled a marketing freelancer with over $366,000 in Fiverr sales across seven years, who started at $5 a job and said: “I didn’t see a huge return in the first three years, and I was busy working other part-time jobs.”⁴⁶ It profiled a resume writer who reached $2 million in lifetime sales after starting in 2014, and who said traffic picked up “After a year or so” and that “In 2016, it really took off.”⁴⁷ Both figures are sales before Fiverr’s fee, both articles are from 2021, and neither says CNBC checked payment records. Both people started on Fiverr in 2014. Two sellers from one site twelve years ago cannot show a typical timeline, and they say nothing about freelancers with direct clients.

Service businesses with checked revenue. TrustMRR is a public list of businesses whose revenue is read from their accounts at Stripe, a payment processor. It shows agencies and productised services taking from $11,000 to $232,000 in the last 30 days. Examples: an on-demand video-editing service at $20,000 and a lead-finding service for tradespeople at $24,000⁴⁸. This is a list of winners who chose to be listed, it shows sales and not profit, and these are businesses with staff or contractors.

Skilled work with AI is the growing part. Fiverr’s annual report says AI “significantly expanded our customers’ needs for high-skilled and complex services”⁶, and its clients completing projects of $1,000 or more grew 13%⁵. Upwork’s figures point the same way. Its 2026 release says: “Freelancers performing AI work on the Upwork Marketplace earn 34% more per hour than those not incorporating AI”, that freelancers doing complex work with AI earned 45% more than a year before, and that professional services using AI “grew 72% in volume with earnings rising 22%”³³. Its 2025 release says spending on AI-related work on Upwork “grew 60% year-over-year in 2024”³². These are Upwork’s own platform numbers, with no method published beyond the releases. An Oxford study of job posts found that “the net demand for freelancing jobs has increased after the launch of ChatGPT”, while writing and translation “have seen demand drop by 20 to 50 per cent”³⁶.

A realistic good result (our estimate). No source measures what newcomers earn in their first year, or what share of them get established. For those who do, the evidence on this page suggests this: little for the first months; then some hundreds of dollars a month in the first year (the one first-year report we found averaged about $500 a month⁴²); around $2,000 a month in sales for an established seller at Fiverr’s top level (the one such report we found³⁹). Salary-level income is documented only for sellers with several years on a platform⁴⁶⁴⁷ and in a survey of established skilled freelancers³². We found no checked case of a newcomer reaching $10,000 a month in 90 days.

What it takes to compete

A skill businesses already pay for. The platforms’ own reports say demand is moving away from simple tasks and towards complex work⁶³³. Reselling raw AI output cheaply points at the part Fiverr says is under pressure, and Fiverr’s rules bar unmodified AI output in any case⁶¹². Knowing a field well and using AI to work faster points at the part the platforms say is growing³³. The studies disagree on whether AI narrows or widens the gap between beginners and experts³⁵³⁶, so this divide is the platforms’ account, not a settled finding.

The first reviews. A newcomer’s biggest hurdle is having no track record. In a controlled experiment on a freelance marketplace, a researcher “hired 952 randomly-selected workers” in an entry-level market and gave them public reviews. The result: “Both hiring workers and providing more detailed evaluations substantially improved workers’ subsequent employment outcomes.”³⁷ So capable beginners are held back by having no history. That fits the profiled seller who started at $5 a job⁴⁶. A related study found that beginners attached to an agency “have substantially higher job-finding probabilities and wages”³⁸. Both studies are old and predate AI tools.

Time before real money. The two documented sellers describe one to three years before real income, both starting on Fiverr in 2014⁴⁶⁴⁷. We found no data on how long new freelancers take today, or on people who start with an existing network. Someone who needs income within weeks is in a weaker position than someone who can start beside a job.

An audience or a network, for direct clients. The one tweet that explains how direct clients are found assumes 2,000 or more followers, unpaid work first and posting five times a week⁷⁶. The first-year Fiverr report came from someone with an existing audience⁴².

Fewer, larger clients. Every route to $10,000 a month in the arithmetic above runs through a high rate, a few clients paying about $2,000 a month each, or dozens of $150 jobs. The forum seller with a record year credits repeat buyers and one large order⁴⁰.

Not relying on one platform. The long-running Fiverr seller’s sales fell by about two thirds in the year a ranking change demoted the account³⁹. Clients found off-platform cannot be taken away by a ranking change.

What it costs. Very little cash: Connects at $0.15 each, an optional subscription, perhaps a ChatGPT plan at up to $20 a month⁹¹⁷²⁹. A teaching certificate costs a few hundred dollars where one is wanted⁵⁵. The real cost is unpaid time. The ILO found about 23 unpaid minutes for every paid hour on freelance platforms, a bit over a quarter of working time (our sum)²⁴.

How to tell early which side you are on. These are our suggestions, drawn from the figures above:

• Can you name the skill, and would a business pay for it without AI in the picture? If the whole offer is “I will run this through ChatGPT”, the buyer can do that too.
• After about 20 to 30 well-aimed proposals, has anyone replied? Across the market it takes well over 11 proposals per hire once dead postings are counted²². No replies at all after that many points to the offer or the profile.
• On Fiverr, have you reached Level 1 (5 orders, 3 clients, $400) within a few months⁷? Across all freelance sites, most registered profiles never reach ten jobs²³.
• Do clients come back? Repeat work is what separates the good outcomes above from a string of one-off small jobs.
• Is your price per job rising? If you are still at $28 jobs after six months, the sums in Step 3 do not reach a living wage in a high-cost country.

What we did not verify

• Earnings on the platforms. Neither Fiverr nor Upwork publishes how many freelancers earn anything, the middle freelancer’s earnings, or a current count of sellers. The 6,400 ceiling and the per-seller average are our estimates, and they assume sellers receive 72% of buyer spending. The counts of who earns what are from 2016 to 2021, before AI tools, and we do not know whether they are better or worse today.
• Direct clients. We found no data on what direct retainers pay, how many pitches they take, how long the first client takes, or how long new freelancers of any kind take to reach a living income. Most of the $10,000 claims are about this route.
• The tweets’ own claims. We checked none of the income figures in the 14 tweets: the 90-day claim, the $10,000 and $30,000 months, the 5,137-jobs analysis, or the price Fiverr Pro sellers charge for keyword research. Twelve tweets were read from our collection only. Images, videos and threads were not viewed.
• What sits behind the sign-up pages. We did not submit any email form, so we do not know what is offered after the free training or the free account. The $19 download in one small account’s tweet was not found. The profile links were recorded on 10 October 2026; two of the three were seen again in profile text that day⁷⁵⁷⁶, and @Mike_Scully_’s was not re-opened by the editor.
• Blocked sites. Fiverr, Upwork and Preply help pages refused direct requests; we read them through a reader service (r.jina.ai) or the help centre’s public data feed. The FTC and SEC press pages also opened only through the reader. Fiverr’s and Upwork’s legal terms did not open by any route, so the fee and AI rules come from help pages. Reddit was fully blocked. Fiverr category pages showed a CAPTCHA, so we have no measured gig prices. Upwork’s own guide to hourly rates returned an error.
• Upwork details. The typical number of Connects per proposal (our $1.50 rests on one second-hand account), the price of the Freelancer Plus subscription, and Upwork’s current wording on AI.
• Vendor figures. UpHunt’s figures (11 proposals per hire, 40% of postings hiring) and SnipeWork’s hourly budgets are tool vendors’ own analyses and were not independently checked.
• Papers read only in part. The Management Science paper (summary), the study by Hui and co-authors (through their Brookings article), the Oxford job-post study (through the university’s press release), the two studies on first reviews and agencies (abstracts), and the 2016 Fiverr study (through a web copy of the paper).
• Self-reported figures. The Fiverr forum seller’s figures, the CNBC-profiled sellers’ sales, the 2014 first-year report and the 102-proposal account are all people’s own statements. The MBO Partners report itself was not opened beyond its landing page. The Skool counts differ between readings.
• TrustMRR. We read the category pages but did not open individual listings to check team size, costs or how each was verified.
• Regulator cases. We did not check the final orders in the Arise and Publishing.com matters, and we did not find the companies’ public replies in any of the three FTC cases.
• Crypto services. We found no independent data on what freelancers earn from ghostwriting, promoter management or similar work for crypto projects, and no price list on any seller’s page. Whether $2,000 a month per client is typical there is unknown.
• Teaching. How many tutor applicants Cambly or Preply accept, how many hours are actually booked, the middle tutor’s earnings, and anything about independent tutors. Preply’s full commission table is an image we could not read. Whether a certificate or work permit is legally needed in particular countries was not researched.
• Costs not priced. Income tax, refunds, chargebacks (a buyer’s bank reversing a payment) and unpaid revision time. European Union tax and marketing rules were not checked.
• Newer figures. The FTC job-offer loss figures are for early 2024; we did not look for an update.

Sources

1. Does It Pay collection of 14 tweets on this scheme, collected 2026-10-06. Profile checks made 2026-10-10.
2. Fiverr Help, “Your earnings page”. https://help.fiverr.com/hc/en-us/articles/9234443621137-Your-earnings-page. Last edited 2026-06-16. Read at source on 2026-10-10 (through the help centre’s data feed; direct request refused).
3. Fiverr Help, “How Fiverr works for freelancers”. https://help.fiverr.com/hc/en-us/articles/34069565843985-How-Fiverr-works-for-freelancers. Last edited 2026-10-06. Read at source on 2026-10-10 (through the help centre’s data feed).
4. Fiverr Help, “Withdrawing your earnings & managing payout methods”. https://help.fiverr.com/hc/en-us/articles/360010530058-Withdrawing-your-earnings-managing-payout-methods. Last edited 2026-07-26. Read at source on 2026-10-10 (through r.jina.ai and the data feed).
5. Fiverr, second quarter 2026 results, Form 6-K exhibit 99.1. https://www.sec.gov/Archives/edgar/data/1762301/000117891326003624/exhibit_99-1.htm. 2026-07-29. Read at source on 2026-10-10.
6. Fiverr, Form 20-F annual report for 2025. https://www.sec.gov/Archives/edgar/data/1762301/000117891326000858/zk2634486.htm. 2026-03-12. Read at source on 2026-10-10.
7. Fiverr Help, “Understanding Fiverr’s freelancer levels”. https://help.fiverr.com/hc/en-us/articles/360010560118-Fiverr-s-level-system. Last edited 2025-12-17. Read at source on 2026-10-10 (through r.jina.ai).
8. Fiverr Help, “Promoting your Gigs with Fiverr Ads”. https://help.fiverr.com/hc/en-us/articles/360017729338-Promoting-your-Gigs-with-Fiverr-Ads. Last edited 2026-08-25. Read at source on 2026-10-10.
9. Fiverr Help, “Seller Plus Standard and Premium: Advanced tools for business growth”. https://help.fiverr.com/hc/en-us/articles/360017140717. Last edited 2026-10-01. Read at source on 2026-10-10 (through the help centre’s data feed).
10. Fiverr Help, “Early Payout”. https://help.fiverr.com/hc/en-us/articles/4402267122449-Early-Payout. Last edited 2026-08-19. Read at source on 2026-10-10 (through r.jina.ai and the data feed).
11. Fiverr Help, “Using AI on Fiverr: Guidelines for freelancers and clients”. https://help.fiverr.com/hc/en-us/articles/34998793899665-Using-AI-on-Fiverr-Guidelines-for-freelancers-and-clients. Last edited 2026-07-09. Read at source on 2026-10-10 (through r.jina.ai).
12. Fiverr Help, “Community Standards: Illegal and prohibited services”. https://help.fiverr.com/hc/en-us/articles/32243195699985-Community-Standards-Illegal-and-prohibited-services. Undated (not in the data feed). Read at source on 2026-10-10 (through r.jina.ai).
13. Fiverr, initial public offering prospectus. https://www.sec.gov/Archives/edgar/data/1762301/000104746919003639/a2239036z424b4.htm. 2019-06. Read at source on 2026-10-10.
14. Upwork Support, “Learn about the Freelancer Service Fee”. https://support.upwork.com/hc/en-us/articles/211062538-Freelancer-service-fees. Last edited 2026-01-22. Read at source on 2026-10-10 (through r.jina.ai; direct request refused).
15. Upwork, Form 10-K annual report for 2025. https://www.sec.gov/Archives/edgar/data/1627475/000162747526000012/upwk-20251231.htm. 2026-02-13. Read at source on 2026-10-10.
16. Upwork, Form 10-Q for the quarter ended 30 June 2026. https://www.sec.gov/Archives/edgar/data/1627475/000162747526000047/upwk-20260630.htm. 2026-08-10. Read at source on 2026-10-10.
17. Upwork Support, “Understanding and using Connects”. https://support.upwork.com/hc/en-us/articles/211062898-Understanding-and-using-Connects. Last edited 2026-10-06. Read at source on 2026-10-10 (through r.jina.ai and the help centre’s data feed).
18. Upwork Support, “How to become Top Rated on Upwork”. https://support.upwork.com/hc/en-us/articles/211068468-How-to-become-Top-Rated-on-Upwork. Last edited 2026-06-10. Read at source on 2026-10-10.
19. Upwork Support, “How Fixed-Price Payment Protection works for freelancers on Upwork”. https://support.upwork.com/hc/en-us/articles/211063748-How-Fixed-Price-Payment-Protection-works-for-freelancers-on-Upwork. Last edited 2025-11-13. Read at source on 2026-10-10 (through r.jina.ai).
20. Upwork Support, “What’s the Trust and Safety team at Upwork?”. https://support.upwork.com/hc/en-us/articles/211067618-Circumvention. Undated (not in the data feed). Read at source on 2026-10-10 (through r.jina.ai).
21. Upwork Community, “Trust and Safety AI Update, April 2023”. https://community.upwork.com/t5/Announcements/Trust-and-Safety-AI-Update-April-2023/ba-p/1294843. 2023-04. Secondary (seen as a search summary; the page did not open).
22. UpHunt, “How many proposals Upwork jobs get (2026)”. https://uphunt.io/blog/how-many-proposals-upwork-jobs-get-2026. 2026-09-20. Read at source on 2026-10-10.
23. Kässi, Lehdonvirta and Stephany, “How many online workers are there in the world?”, Oxford Internet Institute. https://arxiv.org/abs/2103.12648. 2021-04-19. Read at source on 2026-10-10.
24. International Labour Organization, “World Employment and Social Outlook 2021” (Tables 1.1 and 4.2 and section 4.2). https://www.ilo.org/sites/default/files/wcmsp5/groups/public/@dgreports/@dcomm/@publ/documents/publication/wcms_771749.pdf. 2021-02. Read at source on 2026-10-10.
25. Cambly, tutors page. https://www.cambly.com/en/tutors. Undated. Read at source on 2026-10-10 (through r.jina.ai).
26. Preply, “Teach” page. https://preply.com/en/teach. Undated. Read at source on 2026-10-10 (through r.jina.ai; direct request refused).
27. Preply Help, commission article. https://help.preply.com/en/articles/4171383. Undated. Read at source on 2026-10-10 (through r.jina.ai; the tier table is an image and was not read).
28. Sacra, “Preply”. https://sacra.com/c/preply/. Undated; figures refer to 2021 and 2022. Secondary.
29. OpenAI, ChatGPT pricing. https://openai.com/chatgpt/pricing/. Undated. Read at source on 2026-10-10.
30. Staffing Industry Analysts, “More independent workers go full time, plan to continue”. https://www.staffingindustry.com/news/global-daily-news/more-independent-workers-go-full-time-plan-to-continue. 2026-09-29. Secondary (a press report of the MBO Partners survey).
31. MBO Partners, “State of Independence” landing page. https://www.mbopartners.com/state-of-independence. 2026. Read at source on 2026-10-10 (landing page only, not the report).
32. Upwork, “Future Workforce Index” press release. https://investors.upwork.com/node/11771/pdf. 2025-04-23. Read at source on 2026-10-10 (through r.jina.ai).
33. Upwork, “Future Workforce Index 2026: How AI Is Redefining the Value of Work”, press release. https://investors.upwork.com/news-releases/news-release-details/upworks-future-workforce-index-2026-how-ai-redefining-value-work. 2026-07-14. Read at source on 2026-10-10 (through r.jina.ai).
34. Demirci, Hannane and Zhu, “Who Is AI Replacing? The Impact of Generative AI on Online Freelancing Platforms”, Management Science. https://doi.org/10.1287/mnsc.2024.05420. 2025-01-24. Read at source on 2026-10-10 (summary only, through OpenAlex, an index of research papers; publisher page blocked).
35. Brookings, “Is generative AI a job killer? Evidence from the freelance market”. https://www.brookings.edu/articles/is-generative-ai-a-job-killer-evidence-from-the-freelance-market/. 2025-07-08. Read at source on 2026-10-10.
36. Oxford Internet Institute, “The winners and losers of generative AI in the freelance job market”, press release. https://www.oii.ox.ac.uk/the-winners-and-losers-of-generative-ai-in-the-freelance-job-market/. 2025-01-29. Secondary (press release about the paper).
37. Pallais, “Inefficient Hiring in Entry-Level Labor Markets”, American Economic Review. https://www.aeaweb.org/articles?id=10.1257/aer.104.11.3565. 2014-11. Read at source on 2026-10-10 (summary only).
38. Stanton and Thomas, “Landing the First Job: The Value of Intermediaries in Online Hiring”, Review of Economic Studies. https://academic.oup.com/restud/article/83/2/810/2461214. 2016. Read at source on 2026-10-10 (summary only).
39. Fiverr Community forum, seller post on six years of earnings. https://community.fiverr.com/forums/topic/342192-after-6-years-given-to-fiverr-132k-in-earnings-been-a-top-rated-seller-for-straight-5-years-fiverr-is-officially-dead-for-me-as-a-seller-since-they-launched-the-new-level-system. 2024-12-26, edited 2026-05-11. Read at source on 2026-10-10 (through r.jina.ai).
40. Fiverr Community forum, “Your Fiverr Journey” board. https://community.fiverr.com/public/forum/boards/your-fiverr-journey-cbx. Posts dated 2026-09-24 to 2026-10-06. Read at source on 2026-10-10 (board previews, through r.jina.ai).
41. Snipework, “Upwork proposal views decline”. https://snipework.com/news/upwork-proposal-views-decline. 2026-07-29. Secondary.
42. Side Hustle Nation, “My first year on Fiverr”. https://www.sidehustlenation.com/first-year-on-fiverr/. 2014-10-09. Read at source on 2026-10-10 (through r.jina.ai).
43. Bankrate, side hustles survey. https://www.bankrate.com/loans/small-business/side-hustles-survey/. 2025-07-09. Read at source on 2026-10-10.
44. Hacker News, “It is practically impossible to find a freelance project these days”. https://news.ycombinator.com/item?id=43451205. 2025-03-23. Read at source on 2026-10-10.
45. Fiverr, press release on awards for $1 million and $5 million lifetime earners. https://fiverrinternationalltd.gcs-web.com/node/9001/pdf. 2024-07-02. Read at source on 2026-10-10 (through r.jina.ai).
46. CNBC Make It, profile of a marketing freelancer on Fiverr. https://www.cnbc.com/make-it/2021/05/05/founder-made-6-figures-from-side-hustle-turned-full-time-job-on-fiverr.html. 2021-05-05. Read at source on 2026-10-10 (through r.jina.ai).
47. CNBC, profile of a resume writer on Fiverr. https://www.cnbc.com/2021/02/05/teacher-made-7-figures-in-sales-from-resume-writing-fiverr-side-hustle.html. 2021-02-05. Read at source on 2026-10-10 (through r.jina.ai).
48. TrustMRR, marketing and content-creation categories. https://trustmrr.com/category/marketing. Live page, 2026-10-10. Read at source on 2026-10-10.
49. AI Integrator OS, free training sign-up page. https://ai-integrator-os.com/opt. Copyright 2026. Read at source on 2026-10-10.
50. Boring Funnels. https://boringfunnels.com/. Undated. Read at source on 2026-10-10 (through r.jina.ai).
51. We Takeover. https://www.wetakeover.xyz/. Undated. Read at source on 2026-10-10.
52. Fiverr Affiliates Help, “Commission plans”. https://help-partnerships.fiverr.com/hc/en-us/articles/12994521796113-Commission-plans. Updated about one month before reading. Read at source on 2026-10-10 (through r.jina.ai).
53. Skool, search for “freelancing”. https://www.skool.com/discovery?q=freelancing. Live page, 2026-10-10. Read at source on 2026-10-10 (first page of results and the data embedded in it).
54. Zero to $6K, course sales page. https://zeroto6k.teachable.com/p/zero-to-6k. Footer dated 2026. Read at source on 2026-10-10 (through r.jina.ai).
55. TEFL Org, online courses. https://www.tefl.org/courses/online/. Undated. Read at source on 2026-10-10 (through r.jina.ai).
56. US Federal Trade Commission, press release on Arise Virtual Solutions. https://www.ftc.gov/news-events/news/press-releases/2024/07/ftc-takes-action-against-gig-work-company-arise-virtual-solutions-deceiving-consumers-about-pay. 2024-07-02. Read at source on 2026-10-10 (through r.jina.ai; direct request refused).
57. US Federal Trade Commission, press release on the Growth Cave complaint. https://www.ftc.gov/news-events/news/press-releases/2025/03/ftc-takes-action-stop-sprawling-growth-cave-business-opportunity-credit-repair-scam. 2025-03-07. Read at source on 2026-10-10 (through r.jina.ai; direct request refused). The settlement is source 72.
58. US Federal Trade Commission, press release on Publishing.com. https://www.ftc.gov/news-events/news/press-releases/2026/04/publishingcom-pay-15-million-misleading-consumers-about-how-much-income-they-could-earn-using. 2026-04-13. Read at source on 2026-10-10 (through r.jina.ai; direct request refused).
59. UK Advertising Standards Authority, ruling on Self Made Girl Boss Ltd. https://www.asa.org.uk/rulings/self-made-girl-boss-ltd-a25-1293166-self-made-girl-boss-ltd.html. 2025-10-15. Read at source on 2026-10-10.
60. US Federal Trade Commission, “Paying to get paid: gamified job scams drive record losses”. https://www.ftc.gov/news-events/data-visualizations/data-spotlight/2024/12/paying-get-paid-gamified-job-scams-drive-record-losses. 2024-12-12. Read at source on 2026-10-10.
61. US Federal Trade Commission, final rule banning fake reviews and testimonials. https://www.ftc.gov/news-events/news/press-releases/2024/08/federal-trade-commission-announces-final-rule-banning-fake-reviews-testimonials. 2024-08-14. Read at source on 2026-10-10.
62. US Federal Trade Commission, “FTC’s Endorsement Guides: What People Are Asking”. https://www.ftc.gov/business-guidance/resources/ftcs-endorsement-guides-what-people-are-asking. Undated. Read at source on 2026-10-10.
63. US Securities and Exchange Commission, press release 2022-183. https://www.sec.gov/newsroom/press-releases/2022-183. 2022-10-03. Read at source on 2026-10-10 (through r.jina.ai).
64. UK Financial Conduct Authority, “Influencers fined for issuing unauthorised financial promotions”. https://www.fca.org.uk/news/press-releases/influencers-fined-issuing-unauthorised-financial-promotions. 2026-02-20. Read at source on 2026-10-10.
65. Decrypt, report on X banning apps that reward posting. https://decrypt.co/354736/elon-musks-x-bans-access-infofi-crypto-projects-ai-slop-backlash. 2026-01-15. Secondary (a news report quoting the X post).
66. Sixth Tone, report on China’s ban on foreign-based tutors. https://www.sixthtone.com/news/1008215. 2021-08-10. Read at source on 2026-10-10.
67. US Internal Revenue Service, “Self-employment tax”. https://www.irs.gov/businesses/small-businesses-self-employed/self-employment-tax-social-security-and-medicare-taxes. Last reviewed or updated 2026-06-27. Read at source on 2026-10-10.
68. US Internal Revenue Service, “Understanding your Form 1099-K”. https://www.irs.gov/businesses/understanding-your-form-1099-k. Updated 2026-06-28. Read at source on 2026-10-10.
69. GOV.UK, “Selling goods or services on a digital platform”. https://www.gov.uk/guidance/selling-goods-or-services-on-a-digital-platform. Updated 2025-09-22. Read at source on 2026-10-10.
70. GOV.UK, “Tax-free allowances on property and trading income”. https://www.gov.uk/guidance/tax-free-allowances-on-property-and-trading-income. Updated 2019-05-08. Read at source on 2026-10-10.
71. US Copyright Office, news release on its report on copyright and AI, part 2. https://www.copyright.gov/newsnet/2025/1060.html. 2025-01-29. Read at source on 2026-10-10.
72. US Federal Trade Commission, “FTC Secures Settlement Banning Growth Cave Defendants from Marketing and Selling Business Opportunities and Credit Repair Programs”. https://www.ftc.gov/news-events/news/press-releases/2026/01/ftc-secures-settlement-banning-growth-cave-defendants-marketing-selling-business-opportunities. 2026-01-27. Read at source on 2026-10-10 (through r.jina.ai; direct request refused).
73. Maity and co-authors, “A Large-scale Analysis of the Marketplace Characteristics in Fiverr”. https://arxiv.org/abs/1609.06004. 2016-09-20. Read at source on 2026-10-10 (full text through ar5iv, a web copy of arXiv papers).
74. SnipeWork, “Upwork rates by category”. https://snipework.com/market/rates. Live page, 2026-10-10. Read at source on 2026-10-10 (a vendor’s tally of budgets in job posts).
75. Full text and profile of the @boringmarketer tweet of 8 May 2025, through a public copy of X. https://api.fxtwitter.com/boringmarketer/status/1920590629786300801. 2025-05-08. Read at source on 2026-10-10 (X itself was not opened).
76. Full text and profile of the @AlexIsBuilding tweet of 7 February 2025, through a public copy of X. https://api.fxtwitter.com/AlexIsBuilding/status/1887851848146509959. 2025-02-07. Read at source on 2026-10-10 (X itself was not opened).

Corrections

If you are quoted or named on this page and think something is wrong, want your reply shown beside it, or want to be removed, write to [email protected]. We aim to reply within 14 days, and always within one month. Factual errors are corrected with a dated note here. A sentence that is seriously disputed comes down while we check it.

Change log: (empty)