Shift：Laravel自动升级工具10年做到5万刀MRR，单人日工作仅数小时

Ten years in

I've been programming since I was 15. I studied Computer Science in college. I've always had side projects. Some have done well — particularly a "March Madness" iOS app back in the early App Store days. I still work on side projects, especially now with AI. But Shift is by far the most successful and remains my primary focus.

Shift automates the process of upgrading Laravel applications between major versions. You sign in, connect your repo, and in less than a minute, get a pull request with all the changes in nice, atomic commits. It costs less than lunch.

I built and run Shift solo. In its ten years, Shift has performed over 175,000 upgrades. While a bit seasonal around the release cycle, it consistently generates over $50k MRR. Growth has slowed a bit with the adoption of AI and fewer Laravel changes. But customer retention is very high. Those who Shift, keep Shifting!

Building at a conference

I gave a talk at a PHP Conference about upgrading Laravel in November of 2015. Taylor Otwell, the creator of Laravel, attended. After the talk, I asked him if any scripts automated the upgrade process. He said, "No. But I'd use it."

I started building Shift that night at the conference hackathon. When I had a working prototype, Taylor mentioned it on Twitter to help me find testers. I continued working on it when I got home. A few weeks later, I launched laravelshift.com on December 23, 2015. Over the Christmas holiday, I made $80 — about 20 runs. Pricing was stupid back then. Having built iOS apps, I charged only a few dollars. But with people buying it over the holidays, I clearly had something.

A scrappy MVP

The initial product was scrappy. A mashup of PHP and shell scripts. It was a true MVP. Although I didn't have a deadline, Taylor mentioned he was working on the next release, so I wanted to be ready. Ideally, I wanted to offer Shifts to cover three major releases (5.0, 5.1, 5.2).

I built a single-page site. No sign-in. Users connected with GitHub, Bitbucket, or GitLab. They chose their Shift. They paid via Stripe Checkout. The initial version immediately redirected users to their pull requests. It took about 60 hours over nights and weekends to make it launchable.

A familiar stack

These days, I run a straightforward, familiar stack: PHP + Laravel, Tailwind CSS, Stripe Checkout, Additional services: Pusher, AWS SES. I host everything with Cloudflare and Digital Ocean for dynamic workers on demand.

The underlying Shift "engine" is still a mashup of PHP and shell scripts, with a much better pipeline architecture. I expanded the website with a user dashboard and redesigned it a few times. But nothing fundamental has changed from the MVP, just a lot more Shifts.

Increasing revenue via pricing, expansion, and adjacent products

Shift started with a single revenue stream: pay-as-you-go runs. I charged $3 to $7, depending on the number of versions upgraded. Embarrassingly low in hindsight. I introduced subscriptions later, once developers trusted the service. Today, revenue splits evenly between the two.

Expanding the catalog has been an obvious driver of revenue growth. Every new Laravel release creates a new Shift. But I also built adjacent products: a test generator, code modernizer, and one-off refactors. These help normalize seasonal revenue.

Absurd margins

The margins are absurd. $100 a month in overhead against $50k+ MRR. Over the years, I've brought in a few contractors. Mostly, when I needed extra capacity or just wanted some company instead of running the business solo. But no employees. No investors. Now with AI, I'm back to doing it all myself. Total overhead is still around $100 a month. I do a few hours of support in the morning. A true lifestyle business. With essentially no overhead, I've had a lot of flexibility to experiment, make mistakes, and figure it out as I go.

Building authority and trust via communities

The biggest growth lever has always been community. From day one, Taylor Otwell backed Shift with a tweet. That one retweet reached exactly the right audience. That definitely gave Shift the initial boost most projects never get. Taylor's tweets were gold. But I didn't always have them. I needed to build my own audience. I got involved. Spoke at Laracon. Got on community podcasts. Wrote blog posts. Built courses where I saw knowledge gaps. None of it explicitly marketed Shift. I'm not salesy. But I was never afraid to mention Shift when it added value to the conversation.

Relying on luck and feedback

The luck part was meeting Taylor at that conference. It was exactly the right time, with exactly the right idea. His early backing gave Shift that initial boost. I'm never afraid to acknowledge that luck is a big factor. The feedback was those early email follow-ups. From day one, I sent an email follow-up every single user 48 hours after they ran a Shift. Two simple questions: What manual changes did you still have to make? How did you find out about Shift? The answers weren't always what I wanted to hear. But I replied to every one. I still do.

Freedom is the goal

Other than that, here's my advice: Ship it! Shift was buggy at launch. I knew it. I shipped anyway. I could have easily spent a year building the full catalog and getting the automation perfect. But I might have missed my window. And it's okay to stay small. Freedom is the goal. Not the exit. Find a real problem in a community you're already part of. Your idea doesn't have to be the next trillion-dollar company. It can be something you and the people around you pay for. That's it.