How to start a dev agency with $0

You start a dev agency with $0 by picking a narrow niche in week one, running three free audits for case studies in weeks two and three, closing your first paid contract in week four, and reaching $10k MRR by month three with one repeat client and one new logo. Total tool spend stays under $50/month using Stripe, Cal.com, Notion, Linear free tier, and LinkedIn. No venture funding, no office, no payroll until month four.

Most "start an agency" advice assumes a network, a brand, and a runway. This post does not. It assumes a laptop, working knowledge of how to ship software, and 15 to 25 hours a week. Everything below is the founder-does-everything stage. You are sales, delivery, ops, and finance until revenue justifies a hire.

The 90-day plan: $0 to $10k MRR

The calendar version, with execution detail below.

Week	Focus	Revenue target	Time investment
1	Pick niche, write positioning, publish LinkedIn profile	$0	10-15 hrs
2-3	Three free audits delivered as case studies	$0	20-25 hrs
4	First paid pilot ($2k-$5k fixed scope)	$2k-$5k	25-30 hrs
5-8	Repeat client + first net-new logo	$5k-$8k MRR	30-40 hrs
9-12	Second new logo + first contractor or hire	$10k MRR	35-40 hrs

Aggressive but not fantasy. You trade money you don't have (ads, sales reps) for time you do (founder hours), and compound that into proof (case studies) that closes future deals.

Week 1: pick one niche and one offer

The single biggest mistake new agency founders make is being a generalist. "We build web apps" is invisible. "We rebuild the broken Stripe billing logic for post-seed B2B SaaS companies" gets a meeting.

A good niche has three properties:

1. You can name the buyer in one sentence. "Solo founder of a pre-Series-A B2B SaaS in fintech or devtools, technical, ships their own code, hates managing contractors."
2. The problem repeats. Stripe migrations, Postgres performance, Next.js to Remix ports, internal tools cleanup. Pick a wound that bleeds in 100 companies, not 5.
3. You have lived it. If you have shipped three Stripe integrations in your career, your niche is Stripe. If you have done one of everything, pick whatever you did last and most successfully.

Write a 60-word positioning sentence and a 200-word offer page on Notion (free tier). The offer page lives at notion.site/yourname/offer and is what you send to every prospect. It has: who it is for, what you do, what it costs, what the deliverable looks like, and three named outcomes from past work. Do not skip the named outcomes; vague case studies kill new agencies.

Then update your LinkedIn headline to mirror the positioning. Not "Full-stack developer" but "I rebuild broken Stripe billing for B2B SaaS. Fixed-fee, 2-week pilots." LinkedIn is the only acquisition channel that compounds for free in the first 90 days.

Weeks 2-3: three free audits, structured for conversion

You do not have testimonials yet. You need them before you charge. The cleanest way to get three case studies in two weeks is to do three free audits.

A free audit is not "I will look at your code." It is a structured 90-minute session that produces a written deliverable. For a Stripe-billing niche: 30 minutes of read-only access to their codebase, 60 minutes screen-share on the three biggest billing risks, then a 1-page PDF with prioritized fixes and a quoted price for each. The structure matters more than the content; the PDF becomes the case study.

Source the three audit slots from LinkedIn. Post once a day for two weeks: one observation about your niche, one strong-opinion teardown, one "I am offering 3 free audits this week, comment if you want one." 1-3 people will raise their hand. If LinkedIn yields nothing, post in Indie Hackers, the relevant subreddit, or 2 niche Slack groups.

When you deliver, ask three things at the end of the call:

• "Can I quote one of these fixes for you as a paid follow-on?" (this is how the audit converts)
• "Can I publish an anonymized version of the report as a case study?"
• "Who else in your network has this same problem?" (one warm intro per audit is the goal)

Two of the three audits should convert into paid pilots within 30 days. If zero convert, the niche is wrong or the audit is too vague. The exact structuring and follow-on pricing pattern overlaps with the dev agency client discovery call playbook, worth reading before your first audit.

Week 4: first paid pilot

Price the first pilot at $2,000 to $5,000 fixed-fee for 2 weeks. Not hourly. Not retainer. A 2-week pilot with a defined deliverable is the easiest yes you will get from a stranger.

Use Stripe Payment Links (free to create, 2.9% + 30c on charges) for invoicing. No bookkeeping software needed. 50% deposit before you write code, balance on delivery. Send the link via email or LinkedIn DM. The sales-to-paid cycle: discovery call, scoped proposal in Notion, payment link, kickoff. Five days end to end.

A $3k pilot deliverable should be one fix that takes 30 to 50 hours of focused work. Examples: rebuild a broken billing webhook handler with idempotency and test coverage; migrate Stripe Charges to PaymentIntents; clean up a Postgres N+1 slowing onboarding. Small enough to ship in 2 weeks, large enough to matter.

For the first $50k of revenue, run the agency under your name. LLC formation, business banking, separate brand identity: premature until the offer is proven.

Month 2: second client + the "I am the bottleneck" reality

In month 2 you should be running one client from month 1 (likely on a follow-on engagement) and adding one new logo. Revenue target: $5k to $8k MRR.

This is the month the founder-does-everything model starts to crack. You are doing 30-40 hours of delivery and 10-15 hours of sales and admin. The temptation is to hire too early. Don't. Get to $10k MRR with two clients first; the second client teaches you what the actual delivery process needs before you write a contractor onto it.

Sales channel this month stays LinkedIn plus referrals from your three audit clients. If you treated the audits right, each one is worth 1-2 warm intros. Cold outreach is not yet worth your time at this stage; the conversion rate is too low and the time investment too high. We have an entire breakdown of when cold outreach pays off in the dev agency cold outreach playbook 2026, but the short version is: skip it until $20k MRR.

Keep tracking everything in Notion. One database for prospects (name, source, status, expected close, value), one for projects (client, scope, hours billed, hours actual). The discipline of tracking actual hours vs. quoted hours is what makes your next round of fixed-fee quotes profitable.

Month 3: $10k MRR and the first contractor

By the end of month 3, two repeat clients on retainer plus one new pilot equals $10k MRR. At this revenue, the bottleneck is your delivery hours, not your sales pipeline.

Two paths from here. Path A: stay solo, push prices up, cap at $15k to $20k MRR as a high-touch consultant. Path B: book your first contractor or hire and start trading delivery hours for management hours.

If you take Path B, you have three structural choices for how to add capacity. Hiring a full-time employee at this stage is the riskiest. The payroll obligation hits before the revenue is stable. The two lower-risk options are: a US-based 1099 contractor at $75-$150/hr for spiky work, or booking a Cadence engineer at $500-$2,000/week for steady delivery you supervise. Every engineer on Cadence is AI-native by default, vetted on Cursor, Claude Code, and Copilot fluency before they unlock bookings, which matters when you are spec-driven and need someone who can ship from a prompt without 4 days of ramp.

The pricing math is the post here. At $1,000/week mid-tier billing through your agency at $200/hr, a 25-hour delivery week nets you $5,000 revenue against $1,000 cost. That is 80% margin, no payroll tax, no notice period. Cadence's typical engineer is on first commit within 27 hours of booking, which collapses the usual 2-week contractor ramp.

The structural margin on booked-engineer hours is what lets you grow from $10k to $40k MRR without raising prices on existing clients. The agency growth strategy 2026 breakdown covers the channel mix past $20k MRR.

$0-start vs venture-funded agency: the honest comparison

Most "how to start an agency" content quietly assumes you have either a network or capital. Here is what the trade looks like if you have neither.

Dimension	$0-start (this post)	Venture-funded / VC-backed
Time to first revenue	4 weeks	4-8 weeks
Time to $10k MRR	12 weeks	8-12 weeks
Time to $100k MRR	18-30 months	12-18 months
Founder workload first 6 months	35-45 hrs/wk	50-70 hrs/wk
Risk profile	Low, can't go below $0	High, burn-rate driven
Brand strength at month 12	Niche-strong, narrow	Broad, generalist
Margin profile	70-85% from day 1	20-40% during growth phase
Optionality on Path B	Hire when revenue justifies	Hire ahead of revenue
Failure mode	Founder burnout	Cash-out before PMF

The $0-start path wins on margin and survival probability. The funded path wins on speed. For 90% of dev shop founders, $0-start is right because the agency model has no defensible moat that capital buys. Distribution is the moat, and distribution comes from one credible niche position plus 18 months of compounding referrals.

The $0-$50/month tool stack

Every tool below has a free tier sufficient for the first 90 days. The whole stack costs under $50/month even when you outgrow free tiers in month 4-6.

Tool	Cost (months 1-3)	Cost (months 4-6)	Use case
Stripe Payment Links	$0 (2.9% + 30c per charge)	Same	Invoicing, deposits
Cal.com	$0	$0	Discovery call booking
Notion	$0 (personal plan)	$0	Offer page, CRM, project tracking
Linear	$0 (free tier, 250 issues)	$8/user if you outgrow	Project management
LinkedIn	$0	$0 (skip Premium)	Acquisition channel
Loom	$0 (25 video limit)	$15/mo if you outgrow	Async client updates
Google Workspace	$7/user	$7/user	Custom-domain email

Total: $0/month for months 1-3, around $30-$50/month from month 4. No accounting software, no project portal, no help-desk tool until revenue justifies it.

Don't buy: a custom website (your Notion offer page converts better), accounting software (a spreadsheet works to $50k MRR), paid Slack, CRM software, business cards, an office, paid LinkedIn, ads of any kind.

If you also want partner economics (recurring revenue from referring founders to a platform), Cadence pays partners 10% recurring, which can layer on top of your audit pipeline when discovery calls reveal a founder you can't staff.

What to do this week

You can start this Sunday and have a paid pilot by week 4. Concrete steps:

1. Pick a niche and write the 60-word positioning. Today.
2. Update LinkedIn headline and About. Today.
3. Build the Notion offer page. Tomorrow.
4. Post the "3 free audits" LinkedIn post. Day 3.
5. Book the three audits via Cal.com. Day 4 to 10.
6. Deliver audits with the structured 90-minute format. Day 8 to 18.
7. Convert at least one audit into a paid pilot. Day 14 to 21.
8. Send first Stripe Payment Link for 50% deposit. Day 21 to 28.

That sequence has put hundreds of founders into a first paid agency contract inside 30 days. The hard part is not the steps. The hard part is being narrow enough in week 1 that the LinkedIn post in week 2 finds the right reader.

If you need a second pair of hands once your first pilot closes, Cadence's founder onboarding gets a vetted engineer in a 48-hour free trial, the lowest-risk way to test scaling beyond your own hours without a hire.

Running a dev shop and looking for a way to earn recurring revenue off founders you can't take on? Cadence pays partners 10% recurring on every founder they refer. No volume minimums, no exclusivity, paid weekly. See the partner program.

FAQ

How much money do you actually need to start a dev agency?

Zero. The hard costs in the first 90 days are Stripe processing fees (2.9% + 30c per charge), which only hit when you have revenue, and optionally $7/month for Google Workspace email on a custom domain. Everything else (Notion, Cal.com, Linear, LinkedIn, Loom) has a free tier that works fine until month 4. The currency you spend is time, not money.

How long until I can quit my day job?

Nights-and-weekends, expect 4 to 6 months to consistent $8k to $12k MRR before quitting is safe. The math: 3 months of expenses saved plus one quarter at target MRR before cutting the safety net. Quitting earlier raises the failure mode from "slow growth" to "rent panic," which kills more agencies than bad positioning.

Should I hire or use contractors when I need more capacity?

Use contractors first, almost always. A 1099 US contractor or a booked engineer (Cadence engineers are at $500 to $2,000/week with weekly billing and no notice period) gives you variable cost without payroll obligation. Hire your first full-time employee only when one engineer's worth of capacity is fully booked for 3 months straight and you have 6 months of payroll in the bank.

What niche should I pick if I have no clear specialty?

Pick whatever you most recently shipped well. The niche doesn't have to be permanent, it has to be specific enough to make a stranger on LinkedIn say "that's me." Stripe billing, Postgres performance, Next.js migrations, internal admin tooling, and Shopify app development are all niches with proven demand in 2026. You can rebrand in year 2 once you have revenue and case studies; in month 1, narrow beats clever.

How do I price the first paid pilot?

Fixed-fee, $2,000 to $5,000, 2-week scope, 50% deposit. Do not quote hourly until you have shipped 10 projects and know your actual hours-to-deliver ratio. Fixed-fee forces you to scope tightly and pays for the discovery work; hourly punishes you for being efficient. The follow-on engagement after the pilot can move to retainer ($4k-$8k/month) once the relationship is established.