Starting A Japanese Candy Subscription Service

Published: December 9th, 2018
Bemmu Sepponen
Founder, Candy Japan
Candy Japan
from Tokushima, Japan
started July 2011
Discover what tools Bemmu recommends to grow your business!
customer service
Discover what books Bemmu recommends to grow your business!
Want more updates on Candy Japan? Check out these stories:

Hello! Who are you and what business did you start?

My name is Bemmu Sepponen (@bemmu) and I run a service called Candy Japan that sends surprise boxes of Japanese candy to people all around the world, on a twice-monthly basis.

After moving to Japan I found the candies here fun to try, so I decided to start a service that would send them to people abroad to try as well. For example we have candy that makes a sound as you blow air through it, ones that start to foam when eaten and a large variety of DIY kits where you mix parts together to create your own candy. Another fun one has you dig through chocolate to find hidden gummies inside.

There are many flavors that you might not see outside Japan, such as rose, wasabi, green tea and Japanese plum (umeboshi). Some candies even have their flavor change when you mix them together, or have mystery tastes where you don’t know whether something will be sweet or sour before you try it. Part of the reason for this variety is the large size of the Japanese candy market, coming up with new concepts for Japanese consumers to try.

Another reason is that every area in Japan has their own regional specialty, and the local ingredient will eventually find its way into candy as well. For instance near us there is a place famous for their onions, so of course they also make onion candy drops for all those onion-loving tourists that flock to that area.

Same goes for every other area with their own regional specialties. I’ve seen beer caramels, octopus jelly, potato caramels, seaweed candy, chinese soup drops… the list is endless. One time I tried to offer a daredevil subscription that would only send the most shocking things to the most daring customers, but that one didn’t have many takers, so now I limit what I send to those I would be personally willing to eat :-)

The site has paid my living expenses in Japan for 7 years now, and over that time I’ve shipped over $1M worth of candy in total.


What's your backstory and how did you come up with the idea?

I’m originally from Finland, but now live in Japan.

I started working full-time as a developer at a tiny software company when I was 17 years old, then changed jobs and worked at a university developing tools related to freetext search. The department encouraged me to try to also apply as a student, I got in and now have a computer science degree as well. Finland can be very flexible like that.

The biggest impact from becoming a student for me was the ability to have an exchange study period. The choice of where to go was obvious, as I had been interested in Japan for a long time and Japanese was my minor subject. I spent a year in Tokyo, then didn’t really feel like going back, and asked the administration if I could maybe stay another year? And to my surprise they said yes, so I ended up studying Japanese full-time there for two years total.

Later on after coming back to Finland I was on vacation with a friend, and we started brainstorming ideas about projects we could do online. I had heard about this new site called Birchbox, or possibly another one of the early subscription services, where instead of ordering specific products you would just subscribe and random stuff would start appearing in your mailbox.


I thought maybe we could send something from Southeast Asia like that, enabling us to travel to different countries and ship items from a new place each time. I was busy with my work, studies and other side-projects, so we did not end up doing that. However the idea of starting a similar “recurring surprises” service stayed in the back of my mind. He was busy too, and I’m glad I couldn’t talk him into anything, as he eventually ended up becoming a professor at a prestigious university.

When a few years later my wife landed a great job in Japan, we moved back, this time to an island in Japan called Shikoku.


After reading about what it’s like to work as a software developer in a Japanese company, I decided to find another project I could start by myself instead. One of my side projects during the exchange study had been an online store, so I knew that I could sell items from Japan to abroad. The annoying part was the randomness of orders coming in, having to stock items and guessing what might sell, often getting stuck with unsold inventory.

Then I recalled that holiday brainstorming session, where that “recurring surprises” idea had come up. With a recurring service I wouldn’t have the issue of buying something that I’d be unable to sell, as I could know in advance exactly what I need to stock. I could also batch shipments, so I wouldn’t necessarily have to visit the post office every single weekday.

To choose what product the surprise subscription should be about, I figured I could start with anything, and then expand to other things later. I thought candy would be a good first choice, as they are small to ship and inexpensive to get started with, unlike starting a luxury watch subscription for example. They have enough variety that people could enjoy subscribing for a long time, and as most people periodically like to treat themselves to some sweets, you could potentially subscribe even for years.


Take us through the process of designing, prototyping, and manufacturing the product.

When I started in 2011, the term “subscription box” was not in popular use, and it wasn’t even clear that people would want to subscribe to anything besides BirchBox makeup samples.

So before spending too much time on the project, I wanted to make sure I could actually find paying subscribers. I did this by emailing past customers of the online store I had been running before during my exchange study, asking them if they would be interested in random Japanese candy. Three people agreed to subscribe, and I started shipping candy to them even before putting up a website.


I didn’t even have a box, just envelopes stuffed with some candy. The upside of these envelopes was that the service was very affordable as they are cheap to ship.

After sending candy to my three customers for a while, I set up a basic website where visitors could subscribe, at first only by PayPal. I also needed a system that could track who I should be sending candy to, print shipping labels, and so on. No readymade platform existed for starting a subscription box, so I had to do a fair bit of Python programming.

If one were to start a new subscription box now, I would recommend the CrateJoy subscription platform. However I started my site in 2011 and CrateJoy wasn’t released until 2013, so using it was not an option for me.

Describe the process of launching the business.

After the site was up and running, I posted about it on Hacker News. The original blog post no longer exists, because the blogging platform Posterous shut down, but you can still see all the comments here. After posting it a few blogs also featured it (thank you Tofugu, I Heart Chaos, The Daily What, Laughing Squid, and Japanator). I didn’t do any outreach, they simply read my post, liked the idea, and decided to post about it on their own sites.


With the blog backlinks and no competition in this space, I became the #1 result for the search term “japanese candy”. Around this time videos about Japanese candy had also started trending on YouTube, thanks to hugely popular channels such as RRCherrypie, so there were a lot of people searching for the term and ending up on my site.

Thanks to all these factors, within four months of launching I had 300 paying subscribers.

There were no costs involved in starting the site, as I receive payment for items in advance, and only buy the necessary amount to ship them. The main challenge in the beginning was to figure out how to conveniently package and ship hundreds of items. Initially I did this from my living room, recruiting some local students and my wife to help me on occasion.


Since launch, what has worked to attract and retain customers?

Here the story takes a bit of a turn from your typical “success story”. I am yet to find a long-term source for attracting new customers. I have occasional campaigns that seem promising, but they never seem to work out in the end.

I have tried many approaches. Initially showing ads on Google next to relevant searches was slightly profitable, but the bids on good AdWords keywords have doubled. Not that there were so many good ones to start with, as a Japanese candy subscription isn’t a problem in your life you’d often use Google to solve.

Besides search keywords, I’ve tried Google content network ads, Facebook ads, Reddit ads on relevant subreddits, and even more obscure things such as StumbleUpon ads and banner ads through a niche ad network. On the niche network I managed to have a second profitable campaign, but perhaps they really were selling ads too cheaply, as now the entire network has shut down. I suppose everything has its lifespan, and you always need to keep trying new approaches.

One major push was buying preroll ads on YouTube, the ones that show up before your selected video starts. My theory was that since making a video ad is more difficult than a text ad, the bids for showing them might be low right now, so perhaps it would be an opportunity I should try.

I spent over $10k which included getting an ad made and show it to a large number of people, but in the end it didn’t make sense to continue running it. While some customers did sign up after seeing the ad, it again cost too much for it to be profitable.

You can see the ad below, it was aimed at anime fans.

YouTube is so promising that I keep flirting with ways to promote through it, the second attempt was sending samples to YouTubers. I spent a fair bit of time and money contacting YouTubers and then hauling boxes of samples to the post office, but this time it did even worse, resulting in no sales at all.


As you can see, I’ve had many different marketing experiments. Since I do this openly and usually blog about each experiment, some have even suggested that I’m failing on purpose to get juicy blog posts out of them. While the posts do sometimes bring in new subscribers, I would rather find a viable marketing channel rather than publicly blog about my failures.

How are you doing today and what does the future look like?

Future looks uncertain. The site is down to about the same number of subscribers as it had 7 years ago.

Why the decline? The reasons seem to boil down to changes introduced to stop fraud that have reduced conversion rates, increases in costs because of exchange rates and improved packaging, changes in AdWords costs and from no longer ranking as well in organic search results.

Unlike selling software, sending physical packages to foreign countries has costs such as shipping, packaging and the candy itself. Sales are now about $8200 / month, of which about $1500 is profit.

That’s still great as a side-project, and mostly pays for my living cost in Japan, but I’ve started looking for new projects to do as well. So far I’m finding it hard to clear up the mental space to focus on something else. All the daily and weekly tasks involved in running Candy Japan always take me back to thinking about it.

Here’s a chart showing the number of subscribers over time, so you can see the rise and fall in all its glory. If you are wondering about the huge peak in the middle, a lot of it is explained by credit card fraud.


Through starting the business, have you learned anything particularly helpful or advantageous?

I’ve learned a lot about credit card fraud, which isn’t something that I really wanted to learn about. One day fraudsters started placing hundreds of orders with stolen credit card numbers, which took me a while to notice, as it hadn’t ever happened before.


By the time I noticed it, a lot of damage had already been done. I had to cancel and refund hundreds of subscriptions, leading to the sharp decline that was visible on the chart. The fraud was also very expensive, as there are fees involved when you let fraudulent orders slip by. On top of that I had already shipped hundreds of boxes of candy, which I now wasn’t going to get paid for.

To be able to better discern which orders were true and which fraud, I started requiring customers to enter their phone number, confirm their email address and tell me how they found out about the site. This helped with the fraud issue, but made it more annoying to subscribe, leading to less sales afterwards. Possibly very significantly so, but I had to act quickly to stop bleeding money.

Since the fraud issue happened, I have switched my payment processor to Stripe, which comes with anti-fraud features built-in. I have also gotten better at spotting fraud, and am experimenting with undoing some of the changes to make it easier to subscribe again. Hopefully this will help turn things around. If you want even more detail about the fraud experience, I wrote a long blog post about it.

Credit card fraud can come in waves. Instead of having the odd fraudulent transaction mixed in, it can happen as a sudden attack. I had experienced no fraud at all for years, so I was unprepared for it when it did come. I’ve gone from barely knowing credit card fraud exists, to being able to look at an order and quickly spot whether it looks suspicious.

What platform/tools do you use for your business?


The subscription system itself is homegrown, but now uses Stripe for payment processing, which has been a joy to integrate and use. For servers I use Google App Engine, which is overkill for my needs, but I was familiar with it from past projects. I spent a lot of time writing Python in Sublime Text and PyCharm.

I’ve found ZenDesk great for helping customers. Support requests end up in the same place regardless of whether they were sent from email or through social networks. You can track the status of each request, and can hand off requests to others, which is useful when you have people helping out.

When I need a small task done, I’ve often found UpWork to be a good place to look for freelancers. It’s incredibly hit-and-miss, with some borderline scammers in there, but also talented people if you spend a bit of time looking.

For shipping labels I use a Brother QL-710W when I just need to print a few, and a normal desktop printer with larger label sheets when I need to print hundreds.

What have been the most influential books, podcasts, or other resources?

Back when I was a computer science student, I read a lot of Paul Graham’s essays and listened to many talks, such as this great one from David Heinemeier Hansson who wrote Ruby on Rails and this one from Paul Buchheit who created AdSense.

The most inspiring blogger I know is patio11, who also lives in Japan and started a business from nothing, while daily blogging about the idea, launch, marketing and the growth in sales. He is a great writer, and inspired me to also spend more time trying to improve my own writing.

Podcasts are great for consuming entertaining startup stories, as there’s often time in transit where you can’t do anything else, but can still listen. I’ve found that currently How I Built This with Guy Raz has the best-quality interviews. Indie Hackers podcast also has some stellar ones, such as this one with Product Hunt #1 creator Mubashar Iqbal and this one with Chris Chen who launched Instapainting.

Advice for other entrepreneurs who want to get started or are just starting out?

Try a lot of stuff. Sometimes something works. If you can quickly test if people want it, you can increase your chances of finding something workable, as you’ll be able to try more things.

Where can we go to learn more?

I publish a few posts a year on the “Behind the Scenes” blog, where I write about the latest experiments I’ve done and challenges I’ve had, and share yearly revenue reports.

Finally if you would like to try some Japanese candy for yourself, you can subscribe to Candy Japan here.