Shift: From $80 in Week One to $50,000 MRR in 10 Years — a Laravel Auto-Upgrade Tool Run on $100/Month

All dollar amounts are paired with rough yen equivalents at 1 USD = 150 JPY.

Week one revenue: $80

Jason McCreary launched “Shift” on December 23, 2015. Revenue in the first week was $80 (about ¥12,000), across roughly 20 runs. Ten years later, Shift’s monthly revenue has reached $50,000 (about ¥7.5M), with cumulative runs surpassing 175,000. No employees, no investors. Monthly operating cost: about $100 (about ¥15,000).

What Shift sells is remarkably unglamorous: it automates the work of upgrading major versions of the PHP framework Laravel, submitting the diff as a pull request. That’s it. Ten years of persistence on the single task every developer dreads, “upgrading the framework”, produced these numbers.

Shift’s ten years, in numbers

Dividing week-one revenue of $80 by 20 runs gives $4 per run, consistent with the reported $3-$7 initial price range. McCreary himself looks back on that pricing as “embarrassingly cheap.”

What it actually sells

Laravel ships multiple releases a year, and every major version bump requires code changes on the application side. This work is tedious, high in volume, and can break production if handled wrong. Many development teams put off upgrading and end up stuck on ancient versions for years.

Shift processes this work mechanically and hands back the result as a pull request. Humans only need to review and merge. The technical stack is PHP+Laravel, Tailwind CSS, with Stripe Checkout for payments. Infrastructure runs on Cloudflare, DigitalOcean, Pusher, and AWS SES, which essentially accounts for the whole $100/month.

There’s a deliberate mechanism in the pricing design too: the older the version being upgraded from, the more it costs; the closer to the latest version, the cheaper it is. The cost of procrastination is priced directly into the product, embedding the incentive “the sooner you keep up, the better off you are” into the pricing itself.

The moment the pricing changed

The decision that changed Shift’s trajectory wasn’t a feature or a marketing push. It was a price increase.

The trigger was customer feedback. According to McCreary, users explicitly told him “you could charge more.” The $3-$7 initial price was too cheap for work that would take a developer several days to do by hand. From there, he raised prices multiple times, and added a subscription tier to what had been a purely usage-based revenue model. Today, usage-based and subscription revenue are split roughly evenly.

McCreary himself says price was the biggest driver of growth. The figure of 175,000 runs over 10 years is an accumulated result, but at the same run count, the monthly revenue would be an order of magnitude smaller if the per-run price had stayed at $4. Rather than growing user count first, he raised prices to what existing users could bear. That’s the numerical inflection point.

There’s also a moment worth recording as the origin of the whole business. In November 2015, McCreary spoke at a conference about upgrading Laravel. Taylor Otwell was in the audience. When McCreary asked him, “do you have an automation script for this?”, Otwell replied, “No. But I’d use it.” Four weeks after that exchange, Shift launched, and a tweet from Otwell drove the very first wave of users.

How he defended this solo for a decade

The structure that let him keep running this alone for 10 years without getting crushed by competitors breaks into three parts.

First, the 48-hour feedback loop. From day one in 2015, an automated email goes out to every user 48 hours after a Shift run, asking just two questions: “what did you have to fix manually?” and “where did you hear about Shift?” Response rate is roughly 1 in 10. Ten years of these answers have built up into a massive catalog of edge cases that can occur during a Laravel upgrade. Any new entrant trying to match that precision needs the same years of run logs. The moat here is the accumulated catalog of failure patterns, not any single feature.

Second, cost structure decoupled from revenue. $100/month in fixed costs against $50,000/month in revenue. Growth barely moves headcount or server costs. McCreary can say “freedom is the goal. Not the exit” precisely because this structure gives him time back.

Third, marketing was never treated as a separate function. Speaking at Laracon, podcasts, blog posts, courses, forums, Twitter, Reddit, a weekly newsletter, Wednesday livestreams. He’s stayed constantly present inside the Laravel community, mentioning Shift only when the context is relevant. Paid advertising is limited to one monthly placement in Laravel News, which he says has weak ROI and functions more as community support than acquisition. He himself admits, “I have no idea which channel is actually bringing in the most customers.”

What didn’t work, and the headwinds arriving now

He’s also candid about what went wrong. Shift was buggy at launch. “Ship it! Shift was buggy at launch. I knew it, and I shipped it anyway” is how he puts it, reflecting that polishing it to perfection before release would have meant missing the market’s timing entirely.

Seasonality is also a structural weakness. Demand syncs to Laravel’s release cycle, spiking right after a release and dipping in between. McCreary added adjacent products (a test-generation tool, a code modernization tool, and one-off refactoring services) but positions their role not as growth drivers but as “stabilizers to fill the revenue valleys.”

And the current biggest threat is AI. Revenue has recently been flat after currency adjustment, and he estimates the next year could see “as much as a 20% decline.” The breakdown is uneven too, individual developers who used usage-based pricing are drifting to AI tools, while subscription-based team customers are staying. McCreary’s response is to divide labor rather than compete head-on: let Shift handle the deterministic transformations, and hand the final polish to AI.

How far can this be copied

What’s replicable is clear: narrow down to a single task customers repeatedly hate, raise prices to whatever level the market will bear, and keep collecting failure patterns via an automated post-run survey. These three transfer regardless of industry. In particular, the fact that “he ran at too-cheap pricing until customers themselves asked for a price increase” is concrete material for anyone reassessing their own early pricing.

But there are conditions that can’t be replicated too. The de facto endorsement from Taylor Otwell, the framework’s own creator, isn’t something you can buy after the fact. As McCreary himself puts it, “I’ll always admit luck was a big factor.” On top of that, the premise assumes an ecosystem, Laravel, large enough and with a recurring annual version cycle. In technology areas where major version bumps aren’t frequent, the recurring demand simply doesn’t exist to begin with.

And one more thing: the ten years itself has become part of the competitive advantage. Neither the catalog of edge cases nor the trust built inside the community is something you can buy on day one. Rather than a method for rapid growth, what Shift’s numbers offer is an empirical measurement of how far a small, solo-started business can endure, headwinds included.

Related reading

• Zero ad spend, four people, $1M ARR. How Plausible, the Google Analytics alternative, proved a “SaaS sold on principle”
• Zero marketing spend, an anonymous solo developer, $2M ARR. Carrd’s “don’t add features” strategy