After about 7 months in the making and 7 months in operation, I am pulling the plug on my latest side project, a remote jobs board for software engineers in Japan.
So now that it is over, I thought to write a retrospective about the making-of and the lessons learned on this journey to nowhere.
I tried to fill a niche with a format I saw in the English language with sites such as weworkremotely.com or remoteok.io, because I though that could work for the Japanese market as well. Remote Working is a topic of my interest, as it is my current work style in the past two years, and of course because I am an engineer myself. So why not Yet Another Remote Jobs Board™️?
The making
So I made this little site from scratch using the well-proven Rails monolithic stack, with Turbolinks driving the front-end and very minimum javascript. Plain vanilla JS and also some Stimulus towards the last days for a spam protection mechanism that I will leave for another post.
Quite fast and responsive, thanks to the minimal mini.css framework. I actually started working with the Bulma framework, as I was tired of the also-starting-with-B framework behemoth and was looking for some less beaten path.
Bulma looked great out of the box, unfortunately midway in development I had problems with the input fields, as the Japanese input was somehow slow to a crawl on Safari, and this being a site catering the Japanese market the issue was an absolute no-go.
So there I went and rewrote the front-end with the mini one. I can say it worked for me. I found some quirks, nothing important though, and would have been a perfect match if I had some better css skills to make it more visually appealing. Anyways I am quite satisfied with it, it was a right choice at the time.
I was running it all on a single VPS at Sakura Internet (the major national cloud provider). Nothing else. Running ActiveStorage from the SSD and ActiveJobs on threads like an animal. No redis server or any caching mechanism at all. KISS all the way down. Except maybe, I should have gone with the no code approach instead, but then where is the fun?
When building the production server, I took my time to write Ansible and Capistrano scripts so I could spin up more servers and deploy easily. Capistrano came in handy, deploys are super-easy, but I had actual little use for Ansible. Never reached the point to need a second server ever but I slept better knowing that I could rebuild the server with one command in case of catastrophic failure. I suppose it was worth it just for that.
The operation
When everything was up and running, I started to fill it with job postings. How? Just search remote works and pick up some that piqued my interest, copy and paste, rinse and repeat, all by hand.
I had an admin panel I made for myself. I could have opened it to companies looking to post their jobs, but I had not enough visits so there was no point in opening it yet. First had to solve the eyeballs problem.
I had a contact form though, hoping that interested companies may reach through it. There was one message from someone asking me if I would like to post a remote job opening in their site (wtf), and daily spam. I don’t know who thinks it is a good idea to spam randomly through random web sites contact forms, but anyways. I found a way to stop the spam so I implemented it in a couple of hours, and it was all fun and relax again.
For marketing, I just used twitter, with my account for the launch and also created an “official” account just for it. I am very bad at marketing in general and Twitter in particular. Like terribly bad. It is something I have to work out for good. No lessons to give, sorry.
As a redeeming point, I was originally going for the long run. My original plan was to run it for two years so that it catches wind and The Google ranking along the way, but unfortunately I had to shut down midway for the reason I explain next. It’s a shame as I was seeing improvements with the ranking, went from an average position of 50 at the beginning to about 40 , which to be honest is pretty bad as I was only seeing like 1-2 accesses from google every day.
Here take a look at the home page in all its glory!
The end
So on my original plan I should run this for another whole year, but I am closing it right now. Why? I just renewed the domain a couple months ago! There is more interest than ever on remote work because of COVID19!
The reason is lack of time. I have too much on my plate already. I have a full-time job, and I have two children who wants to play with their daddy all the time, and as you can guess, working from home is a blessing and a curse in such situation. And to worsen things, there is the occasional consulting gig on the side, pushing this side project even more to the side.
I have very little spare time for this stuff. So little time for doing marketing, development, maintenance, looking for jobs to post. On a usual week, I just spend my scarce free time picking up the job posting and crafting the tweet to announce it, and that’s it, time over.
Sometimes I managed to develop some unsubstantial tweak to the site, or reaching out to some random established remote work related blogger to beg for him/her to post about my site and not receiving any answer at all.
And then there is the issue of the Japanese language. At first I assumed I could rely on my Japanese wife, as she gladly accepted to help me translating my texts to Japanese. But there was also very little time or motivation for her, she is not really into remote work for engineers. And understandably she’d prefer to do her own things instead of mine.
No problem, I am quite proficient at Japanese. I work at a Japanese start-up and use Japanese daily. But, yes problem. writing Japanese takes me quite some time, and the mistakes. I don’t spot my own mistakes, but they are there. I know you are.
Ok, If wife cannot translate, maybe she could just read-proof my texts? Well, something like a tweet, maybe. But a full-blown article? Still waiting. Published it anyway without proof-reading. It doesn’t matter, it seems no-one ever read that about page.
So besides to developing the site and looking for jobs to post I also wanted to write articles in Japanese to improve the SEO, market it to peer engineers and to hiring managers as well, and all has become overwhelming.
There was a time I could wake up at 5:30 in the morning when the distractions are sleeping and hustle on everything. But no more. Ironically it was the wake-up call that made me realize it was over. So I’m done with it for good. Sayonara baby. Hasta la vista. Finito. Kaput.
The next
So taking the previously said into account, I am ready to start a new chapter, but I will take some precautions. I am scaling down on the scope of my next pet projects. No more full-blown service time-intensive to maintain, I am aiming for something more basic.
Whatever it is, I will be writing about the process of creation from the beginning, not when the horse is dead. Yeah I think it is one of the big missed opportunities of this last project (or any of the others for the matter). It could have worked as marketing material, and I hope it will work for the next one.
I say that because I came to the realization that I am not very active on the Internets and it played against me. Without an audience it is so difficult to launch anything. So I am going to work up on marketing and put it to use with something minimal.
Maybe I can develop some cool simple web application, or write some book on programming / engineering stuff. At first probably I will just be writing technical articles for this blog as warm-up. Something manageable I can do on spare time without pressures, on my own schedule.
And from now on I will be writing English only. No more Japanese, why torture myself. I am not native English but it is pancakes in comparison.
So looking forward you can expect me to be writing more often, as I am ready to share my work in public. Follow me on twitter for updates!
