I earned $635K on Upwork. Here's the system.

What I left out of the post

The post got more attention than I expected. People asked the same five questions in the comments. This is the version with the answers — what doesn't fit in 1,300 characters.

1. "One niche I could defend" — what does that actually mean?

I tried generalist freelancing for the first six months. I called myself a "data analyst & engineer & consultant." I got hired sometimes. Most weeks, I didn't. The jobs that came through were random and underpaid.

Then I narrowed to one phrase: "I help SaaS founders set up product analytics in 2 weeks." Same skills. Different positioning. My win rate went from 6% to 23% in the next quarter.

A defensible niche means a buyer can hold one sentence in their head and know whether you're right for them. "Data analyst" can't be held. "Sets up Mixpanel for SaaS" can. The narrower your niche, the easier you are to refer.

2. Rewriting profile around outcomes, not credentials

My old hero said: "Senior data analyst with 8 years of experience and a PhD in computer science." Nobody hires that line. It's true and it's invisible.

The version that worked: "I've helped 50+ SaaS companies turn raw events into dashboards their CEOs actually open. Last year I rebuilt analytics for a YC alum and they cut their reporting time by 70%."

The credentials are there — buyers can see "PhD" further down. But the hero answers the only question they have: can this person solve my problem in a way I can verify?

3. The 5-line proposal structure

I never wrote proposals longer than 8 sentences. The structure I landed on, after 600+ proposals:

1. Mirror their problem. One sentence quoting their job post back. Proves I read it.
2. One past project that's parallel. Not all my work — the one closest to their problem.
3. Concrete first step. "I'd start by auditing your event taxonomy." Not "I'd discuss requirements."
4. Risk reversal. "If we don't see a clear plan in week one, you don't pay."
5. One question that requires a reply. Not "let me know if you're interested." Something they have to answer.

Templates feel efficient. Templates have an 0.8% reply rate in the data I've seen. Custom proposals built on this structure ran 8-12% reply rate for me, even on competitive jobs.

4. Why redacted client work beats personal projects

Personal projects show you can code. Client work shows someone paid for the result. Buyers care about the second one.

Most of my portfolio for the first three years was redacted: company name removed, screenshots cropped, numbers anonymized. I'd write: "For a Series B SaaS company, rebuilt their attribution model. Reduced spend on a non-converting channel by $40K/quarter." No client logo. Still moved buyers.

If you have zero client work, the workaround is paid mini-projects: charge $200 for a 4-hour audit, write up the result as a case study, repeat. Three of those and you've got a portfolio.

5. Saying no to bad-fit clients

This is the one that took me longest to learn. The signals I now treat as automatic disqualifiers:

• Budget posted as "open" with a job description that suggests it's actually $500
• "We've worked with several freelancers but none have delivered" — usually means they're the problem
• Asking for free work as a "trial" before any commitment
• Vague success criteria. If they can't tell me what done looks like, I can't deliver done

Saying no early protects your hourly rate. One bad client at 30% below your rate destroys two months of momentum.

How to apply this in the next 7 days

The whole system above took me five years to build. The first week version is much smaller, and it's where most of the lift happens.

Day 1-2: Pick a niche you can defend in one sentence

Write 5 candidate niches. Score each on: (a) can you point to past work in it, (b) are buyers searching for it on Upwork right now, (c) can you describe it in 8 words. Pick the highest score. Don't overthink — you'll iterate in month 3 anyway.

Day 3: Rewrite your profile hero

One paragraph. Format: "I help [audience] [specific outcome] [in what timeframe]. I've worked with [proof point]. Last project: [concrete result with a number]."

Day 4-5: Apply to 6 jobs using the 5-line structure

Track each one in a spreadsheet: job link, how you mirrored their problem, what concrete first step you offered, whether they replied. After 6 proposals you'll see your reply pattern.

Day 6: Add 3 redacted case studies to your portfolio

Even if your last "client" was an internal project, write it up like a client engagement. Problem statement, what you did, what changed.

Day 7: Audit your inbox

If you have any active client conversations, look for fit signals. Drop the bottom 1 — politely. The 6 days of momentum you just built are worth more than that one engagement.