For the survey, we were intentionally vague as to what defines an international developer, as we didn’t want to exclude anyone who considers themselves to be one. 7% of respondents were Japanese citizens. However, having Japanese citizenship doesn’t preclude someone from being international, as they could have obtained citizenship by naturalizing or be a citizen who grew up outside the country.
The Greater Tokyo Area, consisting of Tokyo itself, along with the neighboring prefectures of Saitama, Chiba, and Kanagawa, remains the hub for international developers, with 86% of respondents living there.
Outside of the Greater Tokyo Area, the other prefectures that had more than a couple of responses were Aichi (4), Fukuoka (6), Hyogo (4), Ibaraki (5), Kyoto (7), Okinawa (5), and Osaka (5).
Only two respondents indicated they had less than a year of experience. As to be expected, pay scaled with experience.
11% of respondents were women. To put that in perspective, a report based Japan’s 2015 census that indicated 14% of software developers were women, and Stack Overflow’s 2020 developer survey indicates women made up 8% of respondents globally.
While 12% of respondents overall identified as women and non-binary, among those respondents who had less than two years of experience, 32% respondents were members of these gender minorities.
Japanese companies tended to pay worse than other types of companies.
While 20% of all respondents worked for a company with 10,000 or more employees, among those who work at subsidiary of an international company, that number rose to 49%. Larger organizations tended to pay better.
68% of respondents reported using both English and Japanese at the office. Frequent English usage correlated with higher job satisfaction.
56% more respondents selected Python as a language they use than those who selected its made in Japan “competitor” Ruby.
Developers using Swift tended to have the highest job satisfaction, whereas PHP developers tended to be the least satisfied. Developers using Scala were paid the best on average, whereas those using PHP the worst.
Languages that were developed as a successor to another languages tended to have users that were earlier in their career. For example, if we look at Kotlin versus Java, TypeScript versus JavaScript, and Swift versus Objective-C, the newer language always has users with a lower average years of experience.
Respondents were asked to select a single title that best describes their current role. 73% selected a developer role, with full-stack positions being the most popular. Engineering managers had a much higher salary than other positions.
When asked “What do you like most about your current job?”, the most common response was Coworkers. This answer also stood out as the only one that was strongly correlated with gender, with 41% of women mentioning it, whereas only 16% of men did.
Respondents with an annual salary between ¥3 and ¥5 million reported an average satisfaction of 6.0, compared to 7.1 for those making between ¥5 and ¥10 million, and 7.4 for those making ¥10 million and above.
With the exception of a few outliers, programming languages didn’t strongly correlate with job satisfaction. As respondents who used PHP also had the lowest salaries, this could in part explain their low satisfaction. On the other hand, neither Swift nor Objective-C correlated with exceptionally high salaries, indicating that there is more to this correlation than it just being a proxy for salary.
Despite the Developer, game or graphics role correlating with the lowest average salary, they had slightly above average satisfaction. The high satisfaction mobile developers had could explain why developers using Swift, Objective-C, and Kotlin were the most satisfied.
Respondents who frequently used English at their job had a higher average job satisfaction. It’s possible that this correlation was the result of the higher salaries that are associated with English usage.
We didn’t notice a strong correlation between a respondent’s language ability and job satisfaction, beyond those who didn’t speak any Japanese having a slightly lower job satisfaction.
While there was no strong correlation between frequency of overtime and salary, there was a clear correlation between the frequency and job satisfaction.
Engineering managers stood out as being considerably better compensated than other roles. Part of this was because they tended to be more experienced, with an average of 11.7 years as opposed to 8.3 years for non-managers.
However, as Developer, game or graphics, the role with the lowest compensation, also had above average experience at 9.4 years, it’s not only experience influencing salary.
Average salaries scaled with the count of worldwide employees.
Developers who work for a Japanese company had an average salary of ¥7.5 million, versus ¥10.8 for developers working for a different type of entity.
Respondents who said they used Scala were best paid, though this could just be due to our relatively small sample size of 13 respondents who indicated they used it.
Get notified about jobs for English-speaking software developers in Japan. Most positions require no Japanese skills, and many are open to overseas applicants.
TokyoDev is a site dedicated to helping international developers start and grow their career in Japan. We do this through a job board for English speaking software developers in Japan, articles about finding a job here, a forum to discuss life as a developer in Japan, and the stories of international developers working here.
Paul McMahon runs TokyoDev, which started off as his personal blog about being a developer in Tokyo. You can reach him on Twitter at @pwim or via email at [email protected].
Tom Coombs does Product, UX and Data. He founded and runs Man Woman & Child, and co-founded Lendstreet Financial. He focuses on complex, data-rich systems.
Mathieu Mayer is a designer and developer based in Kyoto. He helps companies of all size with Design Systems and CSS. He also contributed to redesigning doorkeeper.jp and TokyoDev.com. At the moment, he works with the folks at fundbook.jp to produce, organize and maintain their products and components library.
A special thanks to everyone who helped giving us feedback on early versions of the survey and results, and helped to promote it, including Ben Watanabe, Paulo D'Alberti, Clement Chidiac, Tutti Quintella, Mustafa Dualeh (Dev in Japan), and Women Who Code Tokyo.
Tom's written an article about building the charts with d3.
We'd love to hear from you! Please send an email to [email protected].