In 2014, I began my search for a software engineering job in Tokyo.
But I didn't want just any old job. I wanted one that was — for lack of a better term — actually good.
Because I'd heard some scary stuff about Japan's tech industry. Tales of overwork. Low wages. The dreaded Japanese "black company".
But despite all those stories, I believed there were good tech companies out there too. So I set out to find them.
8 years later, my wife and I run Japan Dev. It's a job board focused on that same mission: Helping people find tech jobs at great companies in Japan.
Last month, our ultra-niche job board earned $62,197 in revenue.
Here's the story.
It all started with a Trello board
I set up a board on Trello board called "Companies 2014" and started adding companies I wanted to work for.
My requirements were pretty simple. I wanted to find a tech company with the following:
- Good work/life balance
- Good salaries
- Modern tech stack
- International vibe
So I went online and searched. I mostly used Japanese resources like Wantedly, Forkwell, and Green Japan.
But as a foreigner, most of the companies weren't a good fit. Some were too domestic. A lot didn't have any other westerners working there. Some were using old-school programming languages like COBOL instead of more modern languages.
But eventually I did find some gems.
I discovered companies like Mercari and Cookpad. These were product-focused startups. And they were building web services and apps using the same tools as companies in Silicon Valley. They cared about expanding overseas, and they had other foreigners working for them.
So I added them to my Trello board.
I also attended meetups in Tokyo where I discovered some bigger international companies like Rakuten.
And of course, I researched foreign tech companies with offices in Japan (so-called "gaishikei" companies). I realized that places like Google and Amazon had a major presence here, so I added them to the board.
After a few months I had a pretty good list of 30-40 companies. It was a mix of modern Japanese startups and international companies with offices in Japan. In other words, a list of good places to work for foreign software developers like me.
For the next few years I kept updating this list while working as a developer myself. And I started to share what I was learning with others.
I'd post my list of "modern tech companies in Japan" on places like Reddit and Hacker News. But it wasn't until 2017 that I wrote actual code for the project that would become Japan Dev.
In April 2017, I created a mock-up for my idea.
Here's a screenshot:
The idea was simple: Find the "good" tech companies in Japan for foreigners, and share them. Then people like me could find them all in one place.
No more trawling through job boards with thousands of jobs, not knowing which ones are a fit. Or spending all your nights at meetups trying to figure out where other foreigners want to work. Or getting the hard sell from recruiters, just to be ghosted when an interview doesn't go well.
Instead of having to go through that, I decided I’m just going to tell people the good companies. Then people could do their own research and apply directly.
But like most side projects, it didn't go anywhere. I wrote some code, my wife Manami designed a logo and a basic UI, but we barely showed it to anyone and the idea faded away.
I kept updating my Trello board though.
I picked the project up again in 2019.
I started from scratch and built a new prototype. The idea was the same: I had found 50 or so tech companies in Japan I thought were genuinely good places to work, and I wanted to share them.
There were no jobs. Instead, it was more like a "Glassdoor" type site for Japan. We had "pros and cons" for each company, and users could submit their own. Things like "Sponsors visas from abroad" or "Offers 20 days a year of PTO".
The first MVP for this iteration of the project was called simply “Good tech companies in Japan”... and it was pretty simple:
But I continued to work on it, ultimately ending up with the name “Japan Dev”.
I input all the company data I had gathered over the last several years, Manami created a whole new design, and we released the site in April, 2019.
Here's what it looked like:
It got some decent buzz when we first launched.
Our data was pretty good. The site included some good but not-very-well-known companies, so it provided some value to job seekers. It even had unique data like salary ranges that — at least at the time — weren't well known.
...but it was still just a list of companies. And it turned out that was a problem.
Pivoting to a job board
We got an initial spike in traffic thanks to the launch, but it soon died back down to a trickle of users.
And every user we did have was asking the same thing: How do I actually apply to jobs?
Turns out people didn't want a "company discovery platform". They wanted to see jobs. They wanted an apply button.
And I realized I had built the wrong thing.
Also, the whole "Glassdoor for Japanese tech companies" idea had some serious flaws. For one thing, companies are extremely protective of their image. And a site where users could write whatever they wanted about them made them nervous.
Plus there just wasn't a good way to monetize it. Especially in a tiny niche like ours with maybe a few hundred total companies.
It was clear no company would partner with us in our current form, and we knew we would need companies to monetize the service. So we gave up on the idea and set to work implementing a job board instead.
We did keep the Company List, and we still update it to this day. It still showcases data on things like work environment, benefits, language and visa requirements. So this original vision has survived somewhat.
Still, our focus switched to gathering actual job posts, and I got to work turning Japan Dev into a job board.
How Japan Dev makes money
Japan Dev isn't like most indie job boards.
Most job boards charge a flat fee for each job a company posts. Click a button to pay $300 via Stripe, and you can display your job post on the platform for a month.
We don't do this. Instead, we charge a fee when a company actually hires someone.
So we share your jobs on our site. A candidate applies. You interview them and decide to hire them. You notify us you hired them. We send you an invoice. You pay us a fee.
Why do we do this? A few reasons.
The first reason is that it makes it easier to convince companies to post jobs. And seeing as I'm a software engineer, NOT a salesman, I needed all the help I could get on that front.
Why is our model easier to sell? Well, it means there are no up-front costs. In other words, there's no risk for companies. In the worst case if you post jobs but no one actually joins your company through the platform? Oh well. No harm done. You still got some free promotion.
The second reason is that applying through recruiting firms is still the default in Japan.
I don't want to go into the history of recruiting in Japan, but lifetime employment used to be the norm here. That means that changing jobs was a big deal. So the industry was dominated by recruiters who work with candidates one-on-one and provide a high-touch experience.
And while this is changing, even for job boards the "success fee" approach (what we use) remains popular today.
This has one more important effect on Japan's recruiting industry: recruiters charge high fees. 30-35% of a candidate's yearly salary is typical. Some recruiters charge more than that.
Lastly, there's very little churn with this approach. Companies aren't paying you every month, so there's no incentive for them to take their jobs off the service. This is really nice.
However, there are some cons too.
The main one is operational complexity. Instead of slapping a Stripe button on the site, we had to hire a lawyer and create a contract. We still maintain this contract in both English and Japanese and go through a signing process with each company.
We also have to manage invoices and track applicants ourselves.
How we keep companies honest
Quick aside since I get asked this all the time.
Unlike recruiting firms, we don't get involved in the interview process. Candidates find a job on our site, click "Apply", and from there it's between them and the company.
So the question becomes: How do you know when someone actually joins a company? Can't companies just lie, and not tell you when they hire someone?
It's true: We rely on companies to notify us of successful applicants. But we have a lot of measures in place to prevent lying.
For one thing, our contract imposes a late fee for failing to notify us of a successful hire. And the fee increases every month that they don't tell us. This is a pretty good deterrent.
Secondly, we track applicants. Candidates can apply via email (in which case we have their info), or via direct link to the company's job site. In the second case, we show the modal below before allowing them to click through to the application.
Next, we give out 3,000 yen Amazon gift cards to anyone who notifies us that they joined a company through Japan Dev. This is a great way to incentivize people to tell us when they join a company.
And since we have everyone's emails, we can check in with them and let them know about the gift card promotion.
We have some other checks in place too (I'd rather not reveal 100% of my tactics), but these are the main ways that we keep our clients honest.
Speaking of clients...
Finding our first clients
Let's talk about how we actually went about acquiring our first customers.
Japan Dev is a 2-sided marketplace, and 2-sided marketplaces are notoriously difficult because they suffer from the "cold start problem".
To attract job seekers, we needed companies to post jobs...
To get companies to post jobs, we needed job seekers...
It's a catch-22. And if you're going to be successful with a job board or any other 2-sided marketplace, you need a plan to overcome the cold start problem.
There are a few ways to do it.
Job boards like RemoteOK and Startup.jobs did it by crawling jobs on other sites and adding them for free. This attracts applicants, and once you have those you can start making companies pay.
I considered doing this, but instead I decided to try a more manual approach.
Overcoming the cold start problem
When we first started searching for clients, Japan Dev was a side project.
I was still working full-time as a software engineer at Mercari.
And this ended up coming in handy. Because our first customer? It was Mercari. My own company.
This is a great hack to use for B2B services like ours. You have an existing relationship with the company. They trust you (as long as you're a decent employee), and it's easy for you to reach out to decision makers.
I talked to HR, got the contract signed, posted Mercari's open jobs and voila. Japan Dev was a real business!
This was a huge step, but it wasn't enough.
Our second contract was with Indeed.
Some background: Back when Japan Dev was just a company list, it was ranked vaguely by how sure I was the company was a great place to work for foreigners. And in that ranking, Indeed was #1.
A few weeks after posting Mercari's jobs, I came across a post on Reddit by an HR person at Indeed Japan. So I sent them a DM. They'd seen Japan Dev and knew we were promoting Indeed, so they agreed to a meeting. And I convinced them to try posting some jobs.
The next few companies were still a tough sell, but having 2 well-known companies on board helped a lot.
Most of our leads came from cold outreach and networking. This was before COVID-19 so we went to some companies' offices to sell them on the idea in person. We also went to tons of meetups to try and meet potential clients.
Through these efforts, we signed more companies and added more jobs over the next few months. Things were starting to come together.
Now, I'd love to tell you that from this point on, it was smooth sailing. That all these new job posts attracted a flood of qualified applicants. That before long, we were getting placement fees left and right. Our business thrived and we got rich and life was amazing.
I wish I could say that. But dear reader, I cannot.
Because that is not what happened.
Japan Dev had the potential to make money.
We had actual customers. It was an actual business. But it was still a side project — my wife and I were both working full-time, and running Japan Dev at night.
We figured it would be a matter of time before placement fees started rolling in, and technically it was....
It was a matter of 12 hellish months of earning no revenue at all. 12 months of working harder than we ever had before. We'd get off work for our day job, dead tired, and start our work for Japan Dev. We worked every weekend.
And yet we made literally no money for a year. It was terrible. BUT we learned a lot during this time.
We learned how to sell. We learned the importance of distribution and how to attract users. We studied copywriting, email marketing, and a bit of advertising.
We started tracking everything and redesigned the site to improve conversion.
We wrote blog posts. Posted on social media. We built lead lists and designed cold email sequences, and so much more. In other words, we tried everything we could think of to attract more people to the platform.
And we found that some channels worked better than others.
Finding the right marketing channels
Like most software developers who start businesses, I underestimated the importance of marketing.
Turns out it's incredibly important, so we started to really focus on it.
And after trying every channel imaginable, we found that the ones that worked best for us were SEO, email and social media.
SEO is our biggest distribution channel. And we tried a million different tactics to get more traffic from search engines. Things like creating separate pages for long-tail keywords and adding structured JSON to job posts so they show up in Google's "Jobs" searches worked well.
My blog posts also attracted a lot of SEO traffic.
My most popular article was viewed over 50,000 times in the first 3 days of release. And several of my top posts still send thousands of users to Japan Dev every month.
We also worked with external writers to generate content for our blog. This was hit-or-miss, but can work really well if you do proper keyword research (we used Ahrefs) and structure the posts well.
One frustration with SEO is that the feedback loop is long. You change something, and find out 3-6 months later if it worked or not. But it's worth it.
Email is probably our next best channel. We built up an email list with over 10,000 subscribers, and I send job alerts to the list every week.
Once we realized how valuable this list was, I optimized the site for collecting emails. I also wrote a lead magnet (special content that you have to sign up to access — in my case a salary guide). This helps entice people to sign up by giving them concrete value in return.
We also had success with social media.
Specifically, I try to build in public on Twitter and LinkedIn. This is a great hack for tech people like me who don't have a marketing background. I just share what I'm working on, and it gets pretty good engagement. No need to actually learn traditional marketing skills.
We share new jobs pretty much every day on Japan Dev's Facebook, Twitter and LinkedIn accounts too. And we run ads on Facebook and Instagram to increase our social media exposure.
These are far from the only marketing tactics we tried, but they're the ones that worked the best.
And after a year of painfully slow, almost imperceptible progress, we finally earned some actual money.
It had taken one year since signing our first contract. Two years since I started writing the code for the project. And 7 years since I created that Trello board of companies in Japan I wanted to work for.
But our business was finally making money.
Quitting my job
Things continued to pick up in mid-to-late 2020.
Revenue was starting to increase. And as we scaled, it became nearly impossible to run the business while working full-time.
So in April 2021, I decided to quit my job to spend all my time on Japan Dev.
I liked my day job so this was pretty scary. But I tried to quit in the least bridge-burningly way I could to keep the door open in case it ended up being a mistake.
I was afraid our revenue would drop and I'd regret leaving, but thankfully that hasn't happened.
It turned out those months of steady revenue weren't a fluke. The money from Japan Dev continued to come in, and quitting my job proved to be the right decision.
Making a company
At this point, our business was doing well but it was still a sole proprietorship. We had a trademark for the name, but no actual company.
Setting up a sole proprietorship in Japan is super easy. You basically just have to fill out one form. But it doesn't offer the liability protections of a company and the taxes get high after a certain point.
So in mid 2021 we realized we needed to incorporate Japan Dev. Setting up a company was a lot more complex than the sole proprietorship, and I'm not sure I could've done it without my wife (who's a Japanese native speaker).
We opted for a Japanese "KK" (株式会社), mainly because it's the standard here. We thought it would look the most respectable to our clients, and it has some advantages like the ability to issue stock (which makes selling the business easier).
One of the hard parts of incorporating a business in Japan is opening a business bank account. Like many foreigners (and even some Japanese people), I struggled with this.
We applied to 8 different banks before finding one that finally agreed to open an account for us.
There was a bunch of paperwork, some fees and a few other steps like finding an accountant and choosing where our headquarters would be. But in July 2021, we were able to set up our company, Japan Dev株式会社.
Getting a recruiting license
In Japan, certain recruiting activities require a recruiting license.
As a job board, we weren't engaging in any of these activities (we constantly checked with the government to ensure we were in compliance with their rules.
However, pretty much every Japanese company we talked to asked about the license. We ended up spending an enormous amount of time and effort explaining why we didn't need a license to these companies.
Most were satisfied with our explanation, but it still made some of them uneasy.
Plus, there was always the chance the rules could change and we'd be unable to keep doing business. This felt like a risk.
Getting the license would also allow us to do a lot of new things, like manually screen candidates (you can't do this without a license).
In the end, we decided to just get a recruiting license to protect ourselves and make our lives easier.
I'll be honest though: It was pretty annoying.
I had to attend an all-day course about recruiting and pass a test (all in Japanese of course). We had to put 5 million JPY (~$45k at the time) in our business bank account and leave it there. For... some reason.
And since we wanted to use our apartment as our headquarters, we had to choose a place that fulfilled their requirements. So we needed a special room with a lock on it. There had to be a path from that room to a bathroom without going through our living quarters.
We had to buy a plaque with our company name on it and stick it on our apartment door. We needed a special filing cabinet. With a lock on it.
But after jumping through all these hoops, we successfully received our recruiting license in November 2021.
Getting the license provided peace of mind, and it gives us a lot of options for the future.
The current state of Japan Dev
As of August 2022, Japan Dev is still a 100% bootstrapped business. There are no employees — it's still just my wife Manami and me.
But it's going well. In July 2022 (last month), it earned 8.3 million yen ($62,197 at the time of writing), and I'm hoping we'll beat that this month.
We get around 250,000 page views a month — that's more than we got for that whole first year of grinding, when we were making no money. We have contracts with nearly 150 of the best tech companies in Japan.
Here's what it looks like now:
Going forward, we want to scale Japan Dev a lot more.
To do that, we plan to focus on automating and delegating our current tasks. We still do a lot of things manually like customer service, helping companies refine their job descriptions and company profile copy, and adding and removing job posts.
Right now I'm building systems to allow companies to post and update jobs themselves. We're also streamlining our process for signing contracts to decrease the amount of work.
Of course, we're constantly searching for new tech companies in Japan that fit our requirements so we can share their jobs with our users.
We also plan to focus a lot more on our blog. I want Japan Dev to be a one-stop-shop where software engineers and tech folks in Japan can get all the information they need to find a great job in Japan.
That includes information outside of job posts themselves, like visa requirements, salary data and advice on finding apartments in Japan. So we plan to keep adding more content to cover these topics.
But overall, my goal hasn't changed. I want to build the job board I wish had existed when I first arrived in Japan.
And while I think my past self would find the current Japan Dev useful, there's still a ton of work left to do.