I’ve been asked to write this down a hundred and seven times. Here is the honest version — five phases, six weeks for a typical Shopify build, what gets made at each step, what you see, what you sign off, and where the project quietly dies if we’re not paying attention.
Typical duration: 3–5 days
One thirty-minute call. No deck. I ask three questions: what are you actually selling, who is the buyer, and what will make this project a failure? The third one is the question every agency skips and every founder remembers.
What you get: a written brief, the scope locked in plain English, fixed milestones, fixed price, the launch date written into Upwork escrow before I begin.
Typical duration: Week 1
If there’s an existing site, it gets audited line by line — Core Web Vitals, conversion funnel, every app you’re paying for, every JS file that shouldn’t be there. If we’re building fresh, I write the information architecture, the design system tokens, and the page-by-page wireframes before a single Liquid file is touched.
What you get: the audit document if applicable, an IA diagram you can read, Figma wireframes for every template, and a Loom walking you through it. Sign-off here, or it doesn’t leave the room.
Typical duration: Weeks 2–4
Same person, every line of code. A staging URL goes up on day one of week two and you get a fresh Loom every working day. No silent “in progress” ticket stalemate. No two-week black box. You watch it come together at the pace of an in-house team because, functionally, you have one.
What you get: daily Looms, a staging URL kept in lockstep with my local, a public changelog, and a Slack/email line where I answer within four hours during your hours.
Typical duration: Week 5
Performance pass — the staging site has to hit ninety-plus on Lighthouse mobile before we cut over, not after. Analytics wired. Pixels firing. 301s in place from every old URL to its new home. We launch on a Tuesday morning your time, when I have the whole day in front of a screen, not a Friday night when the only person who can fix anything is asleep.
What you get: a launch checklist signed off line-by-line, real-traffic monitoring for the first 72 hours, and a written rollback plan I’ve never had to use but always have ready.
Typical duration: Week 6 + 30 days
Real traffic surfaces what staging never could. For thirty days post-launch I’m on call for everything — bug fixes, performance tuning, the device combination nobody tested, the Klaviyo flow that didn’t fire. Included. No new SOW.
About a third of clients move into a lightweight monthly retainer after this — a few hours a week of CRO, new sections, seasonal landing pages, A/B tests. The other two-thirds disappear with a great site and a thank-you email. Both endings are fine with me.
If your last build felt like a black box, this one will feel like a window.
Hourly billing. It rewards the wrong behavior. Everything here is fixed-price, milestone-paid, escrow-protected.
Six-month engagements with no defined v1. Scope without a line in the sand isn’t a project, it’s a subscription to your own indecision. Every engagement here has a v1 we can both point at on a calendar.
“Can you just match this Figma exactly?” Yes, technically. But Figma doesn’t know your customer’s thumb size, your conversion rate, your Page Experience score. I’ll match the design, then earn the right to flag the three things in it that will hurt the number you actually care about.
Weekend launches. I will be asleep. So will Shopify support. So will whoever has to fix the Klaviyo integration. Tuesdays and Wednesdays only.
Send the URL and what you’d change first. I read every note personally and reply within four hours.
Start a conversation