He copied the market leader's method for four months. Then nearly shut everything down.

Your next paycheck could already exist.

Millions of Americans have unclaimed funds sitting with state agencies—and most don't even know it. A simple search could reveal money that's legally yours.

Look up forgotten bank accounts, tax refunds, insurance payments, pensions, and other unclaimed assets with one quick search.

Reaching $10k MRR in sixty days sounds like luck. In Ivan Nedelkovski's case, it was the payoff of doing the unglamorous work first — the validation, the rubric, the years of operating experience before the weekend build.

Ivan dropped out of university twice — Mechanical Engineering, then Computer Science — and started a series of businesses tackling local problems in Skopje, Macedonia. In 2020 he founded a software development consulting business, MVP Masters, and scaled it to 15 full-time employees and seven figures in revenue before exiting in 2025 to start a product studio.

The studio's first product was Lancer, an AI agent that helps freelancers and agencies get leads and scale on Upwork, fully automated. Within two months of launch it reached $10k MRR. It now sits at $20k MRR, run by a two-person team.

Here is how he did it, including a candid account of the four-month detour that nearly ended the whole thing.

Lancer began as a classic case of building the thing you personally need. While running MVP Masters, Ivan still did not have a consistent channel for acquiring clients. A fellow agency owner recommended Upwork in 2024, and after a few months of learning the platform, the leads started coming.

The results were real and substantial: a $1k consultation, then a $14k iOS app build, then a $10k MVP build that turned into a long-term relationship with a $15k-per-month retainer — a single client that is still with the agency and has generated close to $500,000 in revenue.

But the daily process was miserable. Opening the platform, reading listings, filtering the noise, and writing personalized proposals was endlessly repetitive, and Ivan hated doing it. It was also exactly the kind of task an AI agent could automate. So he built an MVP over a weekend, used it himself, and it worked. He shared it with a couple of friends already on Upwork, and they closed clients with it in the first week.

Then came the validation that mattered most. He connected with one of the highest-rated Upwork coaches on the platform — who happened to also be Macedonian — and the coach was blown away. It turned out he was already an affiliate for an established competitor, whose product he disliked, and he judged Ivan's MVP to be far better. That coach said he would bring a flood of users, and he delivered: two months after launch, Lancer had 30 paying users at around $300 average revenue per user, reaching $10k MRR, most of it from that single coach.

Version 0.1 took a weekend. Version 1.0 — the first commercial release — took much longer, and the gap between the two is instructive. Building a scrappy automation for yourself is a completely different problem from building a product that offers more than 99% job-marketplace coverage and guarantees account safety for every user.

That distinction became a point of pride. Ivan notes that more established players have gotten some of their users banned; Lancer's commercial version was built specifically to avoid that. It took two people building full-time for six months before the first commercial version launched.

The stack:

The most valuable part of Ivan's story is one he tells against himself. Safely scaling scraping and automation has been Lancer's biggest challenge by far, and his first approach to it was a copy of the market leader's — a decision that nearly sank the company.

The mechanism is worth understanding. Upwork lets you set up an agency account and invite other freelancer accounts under a role called "Agency Manager," who can apply to listings on behalf of any agency member. The established competitor used this hierarchy to run its automation: when a user signed up, they were instructed to create an agency account and invite one of the competitor's sourced accounts as a manager.

Ivan is clear-eyed about why that approach is flawed. You never get access to the user's own account, so you lose all insight into inbound leads — which are roughly half of the opportunities established freelancers receive on Upwork. You also cannot touch the inbox, ruling out inbox automation entirely. And the safety argument the competitor makes is, in his assessment, simply false: because verified accounts are hard to source, they get shared across many users, so one sourced account ends up acting as manager in dozens of agencies at once — a pattern practically designed to trigger Upwork's bot detection and cause a chain ban across every connected agency.

He adopted the method anyway, reasoning that it was how the three-year incumbent operated. Four months and thousands of dollars in crypto-sourced accounts later, the team drew a hard line: make direct account connection work, or shut everything down.

They made it work. Lancer now runs automation through users' own connected accounts, and Ivan is direct that if he could start over, he would have skipped the third-party-account approach entirely. The detour is the clearest lesson in the whole story: copying the incumbent's method because they are the incumbent is not validation, and sometimes the established way is established despite being wrong, not because it is right.

Lancer uses a standard subscription model, and Ivan notes the team can set sharper pricing precisely because they are small, lean, and spend nothing on marketing. There are two plans:

The done-for-you setup became a feature by accident, and the story is a useful one. Many users struggled with setup — even those with strong Upwork profiles saw poor campaign performance — but once the team set them up manually, their campaigns took off. Ivan began offering setup to everyone, and when that grew overwhelming, he moved it into the higher tier as a dedicated feature, which raised average revenue per user in the process.

There is also a second, cleverer revenue stream: reverse referrals. Lancer refers its own clients to affiliate Upwork coaches when it believes the client will benefit — the reverse of the usual direction, since those coaches typically refer users to Lancer. Roughly 10% of monthly revenue comes from this.

Lancer's growth started with friends, moved through a single high-leverage affiliate, and then systematized. The first customers were friends on Upwork. Then the one coach became an affiliate and carried the product to $10k MRR within about 60 days of launch. From there, the team built out the channel deliberately:

The UpworkMRR move is the standout. A free, useful tool did double duty: it drove traffic on its own, and it generated the exact dataset that made Lancer's cold outreach precise rather than scattershot. The marketing asset and the marketing intelligence were the same build.

Ivan's central piece of advice is to answer three questions before launch: how big is the market, is there an existing solution and how successful is it, and how will you market and sell it. His framing of why most people skip this is sharp — most indie hackers are more builders than entrepreneurs. They can whip something up over a weekend, especially in 2026 with agentic development, but building is not the same as monetizing and marketing.

He worked Lancer through the exact math, and the process is worth copying. Because Lancer automates Upwork, the platform itself is a hard ceiling on its potential size — so he set out to measure that ceiling precisely rather than guess at it:

The final question in his rubric was whether he could actually improve on what existed. He paid $500 a month to subscribe to the biggest competitor — there was no self-checkout, so he sat through a sales call to do it — and found the product genuinely lackluster for the price. His conclusion is the honest inverse of most founder confidence: he would not have built Lancer at all if that competitor had a better product. Only after all of this did he begin building and marketing.

Ivan's goal is to build the best product in the Upwork niche — a tool thousands of freelancers use to automate the platform. With only around 100 users today, he sees substantial room to grow. The path there, in his view, runs through going deep on agentic-enabling features and adopting a fair, usage-based pricing model.

You can follow along on LinkedIn, YouTube, and Instagram, or check out UpworkMRR and Lancer.

Does your State Treasury Department have money owed to you?

Search your name and state to check if you may appear in unclaimed property records, including funds from accounts, refunds, or deposits.

Search for Money