The Math of a Five-Person Studio in 2026

The story I've seen repeated this year goes roughly: "I built X in a weekend, hit $1K MRR in week three, $10K MRR by month nine, quit my day job." It's a great story. It's also a sample of one, told retrospectively, with all the false starts edited out and the time-to-first-dollar measured from the moment a customer happened to find the post.

I want to write the version that isn't a story.

We're a five-person software studio operating out of Islamabad. We ship four products. We don't have outside funding. We're not chasing an exit. This is what the math actually looks like — costs, revenue logic, where the leverage is and isn't, and what it takes for the spreadsheet to balance.

I'm not going to share our exact numbers, partly because they'd be misleading without our exact context, mostly because the shape of the math is what's useful, not the line items.

The cost stack

Five engineers is the dominant cost, by an order of magnitude. Everything else is rounding.

People is the line that does not compress. You can negotiate cheaper hosting; you can switch from a paid analytics service to a free one; you cannot meaningfully pay an engineer "a little less" without it showing up in retention, in interview-readiness, in how excited people are on Monday morning. Pakistan salaries are competitive with the global remote market in a way they weren't five years ago, because anyone good can take a remote job for a US firm and earn US wages. So the studio's people cost is the going rate for the labor we want, not a deep discount.

After people, the cost structure has four real categories:

Infrastructure. Hosting, databases, object storage. Vercel for the marketing site and for two of the products. Convex for two products' backends. Cloudinary for OurTrail's media. Mid three-digit dollars per month, total. This is the line everyone obsesses over and the line that matters least until you're at a scale where it matters a lot.

Per-user services. Clerk for auth, Stripe for payments, Web3Forms for the contact form. These have free tiers we live inside for two of the products and a per-user fee on the others. The Clerk bill is the largest of these and grows with active users, which means it grows when revenue does, which is the correct shape for a cost.

Tooling. Linear, GitHub, the various AI-coding subscriptions, our internal services. Couple hundred dollars a month. Smaller than infra. Worth paying for; this is the part where being cheap costs more than it saves.

Compliance and legal. Almost zero today, because we're early enough that "privacy policy, terms, basic data handling" is the whole list. This number gets larger if and when we take payments at scale or operate in a regulated vertical, which we're aware of but not yet exposed to.

Add it all up and the recurring monthly burn is people-plus-a-rounding-error. The implication is that every product needs to either pay for a fraction of a salary or not exist on the salary side of the spreadsheet — i.e., one of the team builds it as their primary thing and the product justifies that allocation, or the product is small enough that nobody is solely on it.

The product math

For a product to be worth its slot on the roadmap, here's the question we actually ask: at what user count and price does this clear the bar?

Take envpilot.dev. It's a developer tool. Developer tools have a known price band; let's call it $5–$25/month per seat for a small-team SaaS. For envpilot to pay for one full-time engineer at the prevailing rate, we need roughly 400–1,500 paying seats, depending on the average price point. That's a real number; it's not a vanity metric. It tells you what success looks like before you start building, which means you can decide before the engineer-quarter is gone whether the addressable market plausibly contains 400 paying seats for what you're building.

For something like OurTrail, the math is different because the unit isn't a seat. It's a household, maybe twenty dollars a year. The number of households needed to clear the same engineer-quarter is much larger — multi-thousand. The TAM is also much larger because every family is a potential customer instead of every developer team. The question becomes "can we find those households cheaply?" because if customer acquisition cost runs to $30, the unit economics break.

For LUMO, an Android-only consumer finance app, the math is different again — closer to a freemium ladder where the conversion rate matters more than the price point, and the price-bearing tier is a small slice of total users.

Four products, four different math models. That's a feature of the studio, not a bug. It means we're not all-in on one curve; it also means we have to be willing to kill the products whose math stops working. We've not had to yet. We will, eventually.

The 12-to-18-month $10K MRR claim

A common indie figure is "$0 to $10K MRR in 12–18 months is reachable for a focused founder." Plausible. Achievable. Not the same as "everyone who tries this hits it."

I've seen the inside of enough indie products this year to estimate: of the founders who seriously commit to a niche, build for it, and ship continuously, somewhere between a third and half hit some real revenue number in their first year. The rest don't, mostly for one of two reasons. Either the niche didn't actually contain the volume the founder estimated, or the founder couldn't sustain twelve months of low-feedback grinding before the niche started responding.

The thing the highlight reels don't tell you is that the slope between month one and month twelve is almost always flat for the first six. If you compare yourself to a chart that says "linear growth from launch," and your reality is "flat for six months, then bend," you'll quit at month four. The studios that don't quit at month four are the ones who knew about the bend going in.

Where the leverage actually is

A few observations, after a year.

Distribution is the bottleneck, not product. This is the cliché. It's the cliché because it's true. Five engineers can build a lot of product. Five engineers cannot, by themselves, manufacture attention — that costs money, or time, or both, and the ROI is much less linear than building features. The studios I see succeed at the indie scale are the ones who treat distribution as a first-class function from day one, not the ones who build the better mousetrap and wait.

Boring products beat exciting products at this scale. The exciting product needs ten thousand active users to be interesting to anyone. The boring product (env vars, expense trackers, file-based writing tools) needs a few hundred and they pay you money. We chose four products that all skew toward the boring end of the spectrum, mostly on purpose.

The studio model gives you optionality. A solo founder bets on one product. A five-person studio can run four with overlap. If one breaks out, the others get the attention rebalanced down. If one stalls, it doesn't end the company. This is the single biggest reason I'd recommend the small-team studio over the solo-founder path for anyone who isn't trying to optimize their personal narrative around being a solo hero.

Pakistan is a cost advantage, with limits. Operating from Islamabad means our salary base is lower than equivalent US-domiciled studios, even though our individual salaries are competitive globally. The advantage is real and it's not infinite — the talent market for senior people is global now, so we pay close to global rates for senior roles regardless of geography. The cost advantage is most pronounced at the team level (rent, services, shared overhead) and less so at the individual senior-engineer level.

What I'd tell a friend starting one

Run the math before you build. Not the optimistic math. The realistic math, with churn assumptions, with a six-month flat period baked in, with customer acquisition cost in the spreadsheet. If the math works at conservative numbers, build. If the math only works at optimistic numbers, find a different niche.

Don't quit the day job until the math is load-bearing, not aspirational. A month of revenue is a data point. Six months of stable, growing revenue with retained customers is a business.

Build the studio if you have collaborators. The solo route is romantic and it leaves enormous productivity on the table. Two engineers do not produce twice as much; they produce somewhere between three and four times as much, because the parts that block one person are different from the parts that block another, and the conversations between them are themselves productive.

The dream isn't a wrong dream. The math just has to be in the picture from the beginning, or the dream becomes a long, expensive education.

---

Numbers in this post are described in shape, not in precision. Real exact figures vary too much by team and product to be useful as a benchmark.