The order matters because each step removes a decision the next one would otherwise force. Choosing small at the start is what makes a single polish afternoon realistic, and a finished page is what makes a one-line story possible. Jumping straight to the share step is the most common way a portfolio dies unpublished.
Step 1: Choose three finishable projects first
Starting here protects the whole plan from collapsing in week three. The portfolio's job is to exist in public, not to be impressive, and three finished small things beat one unfinished ambitious thing. Choose projects you can complete in a weekend: a single landing page, a repaired chair, a 40-line script that renames files, a four-recipe zine. If nothing is started, pick three tasks you have already done for someone for free — those count. Write each one on a separate slip of paper so step 2 can attack them one at a time. The failure mode here is choosing work that would look best if finished, which is exactly the work that never gets finished. Ambition is cheap at the selection stage and expensive at the polish stage; you are buying the ability to reach step 5.
Step 2: One timed afternoon per piece
One afternoon per piece, timed, no more. Polish means fix what breaks, cut what confuses, and stop. Open the project, list the five ugliest things about it, fix those five, close it. A script gets real error messages and a two-line README; a wooden box gets sanded edges and one coat of finish; a zine gets its typos killed and its margins evened. What people get wrong is treating polish as improvement — adding features, redesigning, chasing a better version. That is rewriting, and it moves the finish line past step 5 permanently. Since you already chose small in step 1, treat "done enough to show" as a real threshold you can name out loud before the timer starts. Set the timer. When it rings, the piece is done whether or not you agree.
Step 3: Put them on a simple page
By this point you have three finished pieces, so the page's only job is to hold them, and it should take under ninety minutes. Use a free host with a template — Carrd, GitHub Pages, or a Notion page published to the web — and skip the custom domain for now. One column, three entries, each with an image, a title, and a link. Resist an About page, a blog, or a contact form; those are separate projects and they will eat this week. The usual derailment is turning the site itself into project number four, which then needs its own polish afternoon you don't have. Check how it looks on a phone before calling it done, because that is where nearly everyone who opens the link will see it. Then copy the URL into a note you can paste from in steps 4 and 5.
Step 4: Give each project a one-line story
A single sentence per project does more work than any redesign of the page you just built. Each line should answer two things: what was the problem, and what did you decide? Not the tools. "Wrote a script to rename 400 client files after doing it by hand took an afternoon" beats "Python, automation, CLI." Show the decision, because that is what a reader is actually judging — whether you can think, and whether working with you would be easy. Keep it to one line under the title, in the tone you would use telling a colleague at lunch. The trap is padding it to three sentences because the project feels too small to stand alone; that feeling usually means the project is well chosen, not badly chosen. Read the line aloud once. If you stumble, cut the first clause.
Step 5: Show one person and ask one question
Send the link to exactly one person whose judgment you trust, and make the ask narrow enough to answer in two minutes. "Does the second project read as a decision or just a task?" works. "What do you think?" does not — it invites a polite "looks great" and teaches you nothing. Choose someone who does work adjacent to yours rather than your closest friend, and send it as a direct message, not a group thread or a public post. Whatever they answer, you now have three finished pieces, a live URL, and one piece of outside evidence, which was the point of the whole checklist. The place people stall is waiting until the portfolio feels ready. Readiness is not a feeling you arrive at; it is a state you can only confirm from outside. Ask the question, then fix the one thing they flagged.
Three finished pieces and one live URL is the whole goal, and the strip is a countdown to it rather than a reminder of it. A phone reminder disappears the moment you dismiss it, leaving no trace you ever saw it; a paper strip taped where you will actually look stays put, gets visibly shorter with each cut, and is gone only when the work is.