Most portfolios fail for the same reason: they show a person who can do many things, and a client who needs one thing done. This guide is the version that fixes that: a short, clear, live portfolio that says who it is for and proves it can do the one job.
This is a craft question, not a market one; the specifics below are judgment, not a fact.
The one decision that decides everything
Before a single screenshot: what is the one thing you are selling, and to whom? A portfolio that answers that in one sentence and then proves it with two or three live pieces outconverts a portfolio that shows everything. The offer is the filter; every project in the portfolio has to be evidence for that one sentence, and everything else goes.
What to put in, and what to cut
| Put in | Cut |
|---|---|
| 2-4 real, openable pieces | Ten screenshots of the same kind of thing |
| One line of what each did | Wall-of-text case studies |
| The result, not the tool list | "Used React, Node, Postgres, Docker" with no outcome |
| A link that opens in ten seconds | A password-protected or slow demo |
The cut list is the hard part, and it is the part that makes the portfolio honest. If a piece does not prove the one offer, it is noise, and noise reads as "I do not know what I do."
No paid work yet: the clean move
You do not need a client to build the proof. You need a real, small, shipped thing. A site for a friend or a local business, a class project cleaned up until it is presentable, or a small tool that solves a problem you had. The test is whether it is live and openable and has one line about what it did, and that is a portfolio with a spine, not a folder of screenshots.
The shape that converts
One page. One sentence: who you help, and with what. Two to four live pieces, each with that one line. A way to reach you. That is the whole portfolio, and it is enough. The offer-first structure is the same instinct as the proposal, which is the next step once a client is talking, in writing freelance proposals that get you hired, and it sits on top of the broader how to start freelancing in Nepal path.
A sensible default
Pick the one offer, cut the portfolio to two or four live pieces that prove it, give each one line, and make the page open fast and look openable. Do that and the portfolio stops being a show and starts being a yes/no machine for the right client.
Bottom line: a portfolio wins when it says one thing and proves it with a few live pieces, not when it shows everything. Pick the offer, cut to two or four, give each a result line, and make it open in ten seconds, and it converts the right clients instead of impressing the wrong ones.
FAQ
How many projects should a portfolio have?
Two to four, real and linked, beat ten screenshots. A client does not need depth of variety, they need proof you can do the one thing you are selling. Cut everything that does not prove that.
I have no paid work yet. What goes in the portfolio?
Real, small, shipped work. A site you built for someone, a class project cleaned up and made presentable, a small personal build that solves a real problem. The test is whether it is live and openable, not whether it was paid.
What is the one thing that decides whether it converts?
A single clear offer. A portfolio that says exactly who you help and with what outconverts a fuller one that shows everything you can do and nothing you are for.
Sources and how to verify
Platform policies, payout routes and NRB rules for moving money home change; confirm the current receiving method and fee with the platform and your bank. General guidance, not financial advice.