On Valentine’s Day 2025, I officially announced the release of my first book Autumn: A Local’s Guide to Travel in Japan. It’s been a long journey from initial idea to the final product, more than 14 months from start to finish. It still feels slightly unreal to hold a physical copy in my hands. And now that the final product has been realized, I want to take a moment to celebrate and reflect on the entire book-writing process.

Writing and self-publishing a book is not for the faint of heart. In preparation, I read many other writer’s thoughts on producing their first book, and everyone agreed that it’s truly a terrible time-to-money investment. Nobody gets into the business of writing books to get rich, and for every one runaway success there’s thousands toiling in obscurity. People write books in order to distribute ideas, to gain visibility, to archive and reflect on their life, to nourish their undeniable creative spirit–and in the same vein, this has very much been a passion project for me as well. My goal was to write the book that I wished I had access to, back when I first started visiting Japan.

View this post on Instagram

The book is now available in digital and print formats! It’s available here on my website and also through your favorite retailers like Amazon, Barnes & Noble, and Bookshop.org.

Here’s how it happened from beginning to end, plus thoughts on what’s next.

• Beginnings
• Schedule
• Design & illustrations
• Tools
• Publishing and bureacratic purgatory
• Next steps

A peek into the interior of the Autumn guide book.

Beginnings

It all started, like with most personal travel recommendations, as a Google Doc.

The last updated version of the Google Doc from November 2023 was 2.7k words.

Dru and I have traveled to Japan many times, for business and vacation, both separately and together. We recently did a count of our combined travel history and it added up to fifteen trips to Japan in last ten years. But our knowledge and immersion in Japanese culture really reached a new level starting in 2018. We moved our whole life, including a border collie dog, to live in Tokyo for a year. And while we lived abroad, we were traveling almost every other weekend to see new places within Japan.

Friends and family started to regularly reach out for tips and recommendations. We put together a Google Doc to consolidate our travel knowledge and make sharing easier. This doc eventually grew and grew in length, especially as we started visiting lesser-known cities and countryside towns.

I noticed that people really appreciated the information, especially since we could tailor the advice to our style of travel. We heard about our suggestions being used to plan trips, supplementing the mass-produced advice given by the standard travel books like Lonely Planet. As Kate Lindsay writes in The Most Sought-After Travel Guide Is a Google Doc,

It’s about getting the right recommendations, and not just from people who share your taste, but who have been to the city enough to have reliable intel.

But we had created more than a travel doc. During our year living abroad, I also produced a podcast about the highs and lows of being an expat in Japan. It was an unconventional way to keep in touch with our loved ones, and I had a lot of fun learning how to make a podcast. We had so many new observations about Japan that we were able to fill twelve episodes, with plenty of extra material that was left on the cutting room floor.

When we eventually returned back to the US in late 2019, the podcast naturally came to a close. I started thinking about my next creative project. I had accumulated a lot of photos and (relatively) esoteric knowledge about Japan, but the Google Doc didn’t feel the right place to put this content.

I started flirting with the idea of writing a travel guide. It was appealing to me because guidebooks are a form factor better suited for long-form writing. I could incorporate many kinds of visuals, like illustrations and infographics. It wasn’t strange to mix practical travel tips with personal anecdotes in a book—in fact, it was a more enjoyable to read about travel stories. I started getting excited, looking forward to doing deeper research to expand on the topics that we already had written in the Google Doc. But then COVID happened and no one could travel anymore, so that put my book-writing aspirations briefly on hold.

It would take more than three years before we could visit Japan again, this time for our honeymoon. At the same time, once travel restrictions lifted, combined with the yen falling to a 24-year low in 2022, the number of tourists to Japan exploded.

Now in 2025, global tourism has exceeded pre-COVID highs. News articles are being written about the dangers of overtourism in Japan. And the fears of overtourism are exacerbated by travel guides and videos recommending the same few spots in Tokyo, Kyoto, and Osaka.

New record high of visitors to Japan in 2024 (+15.6% compared to 2019), from the Japan National Tourism Organization (JPNTO).

One thing we have gradually internalized about Japan over the last decade is a better way to travel—a slow, ultra-seasonal, and hyper-regional style of travel preferred by locals. This means planning your visits around festivals and smaller cities with special events during the time of your vacation. So at the start of this project, we tested this travel method on our honeymoon-slash-book-research-trip in November 2023. And, despite tourists causing trouble and congestion in all the “golden route” cities, my husband and I were frequently the only foreigners at places on our itinerary.

We were some of the only foreigners at the Obara Shikizakura Festival, a celebration for a special varietal of cherry blossoms that bloom in autumn (Instagram).

As Craig Mod puts it, in his lovely recommendation for visiting Toyama City,

It turns out, too, that those who do make the treks out to these places tend to be the most interesting of tourists. They tend to be folks who have come to Japan three, four, five, fifteen times. Who are not part of giant group tours. Who are deeply engaged with the culture and move respectfully through the landscape. Who know how to speak in a restaurant without screaming. And in this way, are ideal visitors to Japan or anywhere.

So collecting all my accumulated knowledge, I knew I wanted to write four different guidebooks, one focused on each season. Autumn was a natural place to start – it was not only my favorite season but also the most popular season for overseas visitors.

Autumn (September-November) has the highest combined visitors, although December is the single most popoular month, according to the JPNTO.

Now that I had a clear idea of the project, I got to work.

Schedule

This project’s schedule roughly looked like this:

• Started the project in November 2023
• Began hand-drawn illustrations in Feb 2024
• Completed drafts of all the essays in May 2024
• Started the design formatting in June 2024
• Cover design completed in August 2024
• Completed first draft of layout in September 2024
• Revised essays and overall book layout in October 2024
• Completed all illustrations and started editing photos in November 2024
• Completed all formatting edits in December 2024
• LLC formation and opened distribution accounts in Dec 2024
• Ordered author’s proofs and final edits in Jan 2024
• Announced the book’s release on Gumroad/Amazon/etc on Feb 14 2024

I was funemployed for a short period from November 2023 to the end of February 2024, but the rest of the time I was working on the book on nights and weekends while holding a full-time job. I am lucky that I have very few obligations outside of work at this point in my life, which enabled me to spend so much time on book-writing.

Design & illustrations

The most challenging aspect of producing the book was developing a visual language for the book. My graphic design skills are survival-level; good enough in a pinch, but not guaranteed to be especially good quality or refined. I put pressure on myself to finish this project to a high level of polish since this first book would set the style for the entire series.

To help build the aesthetic muscle, I researched many Japanese design books to help create a color palette. I also referenced many of my favorite design-forward magazines like Monocle, Drift, and Japanese-language print publications I had collected on our trips to Japan to construct a design system for the book.

Design inspirations for the guide book. Links to the books here.

I also wanted illustrations to be a core component of the guide. My fondest memories of studying Japanese are centered on the cute illustrations in Japanese textbooks. These iconic clip art drawings are created by the illustrator Takashi Mifune and freely distributed on his website Irasutoya for both personal and commercial use, with some minimal restrictions. And because of the generous usage terms, his art is ubiquitous throughout Japan. It is especially used in educational and informational materials, which was a perfect fit for the goals of this project.

I love the quirky and cheerful vibe of Irasutoya art. I wanted this mood to really suffuse through the entire book, and so I made sure to develop my custom hand-drawn illustrations to be complementary in style. This way, the artworks could sit alongside each other without feeling jarring or out of place. Plus using illustrations in the guidebook also gave me more visual flexibility. My travel photographs often were not perfectly framed or shot, so I could supplement with a drawing instead of needing to source additional photographs.

Examples of Irasutoya artwork and my custom illustrations used in the book.

I considered hiring an artist to do the book cover but ultimately decided to do the cover design myself. It wasn’t necessarily a cost issue since estimates range from $200 to $500 for a high-quality cover design. However, I wanted to do every part of design development myself. Early in the project, I combed through my backlog of Japan photographs to curate a set of autumn photos that I might use for the guide. I realized that I had so many options for a stunning cover photo, so I just needed to roll up my sleeves and iterate on the cover design until I was satisfied.

Sample of the photos in the Google Photos album for the Autumn guide.

In terms of the overall project, the visual and design aspects were definitely where I spent the most time. The writing only took seven months and I was multi-tasking with design-related tasks. Designing & illustrating the book took nine months end-to-end, and that work was the sole focus for five of those months.

Now that the patterns and visual language has been established for this first book, I think the process will go much faster for the subsequent books.

Tools

I wrote the first drafts of all the essays in Obsidian, where I had been compiling research and links for many months beforehand. While everything was written and revised by me, I did use generative AI tools to assist in summarization of online research materials and brainstorming. Google’s NotebookLLM was really helpful in extracting high-level takeaways from the dozens of online websites I was sourcing for my essay topics. I also used a free plan on Anthropic’s Claude to generate consistent title/subtitle pairings for each chapter in the book. Marketing and branding continue to be my greatest weakness in any creative project, so I was grateful for the LLM-based assistance in coming up with catchy headings for the table of contents.

The formatting of the book itself was done completely through Adobe Indesign. I luckily had taken a class in high school where I was taught how to use Indesign, so I was able to resurrect those long-buried memories. My hand-drawn illustrations were created in Procreate on an iPad. I heavily used Google Photos search and photo albums to select and organize the hundreds of travel photographs I had taken over the years. The photos were then edited using Adobe Lightroom. I used Oliver Keys’s Color Film Emulation Profile & Presets as a baseline for color grading, and spent many hours watching Youtube videos to understand how to fiddle around with settings in Lightroom.

Irasutoya website provided the remaining illustrations used throughout the book. I used fourteen of his clip art illustrations, in accordance with Mifune-san’s Terms of Use. The chapter artwork for the “Step-by-step” essay was generated with a free account on AI Irasutoya, a website which allows you to generate artwork in the style of Irasutoya with a generative AI LLM trained his work.

“Winding road through the countryside in autumn” generated output (1,2) on AI Irasutoya.

Publishing and bureacratic purgatory

I completed the development of the book in early December 2024, almost exactly one year since starting the project. I was so happy to have a finalized draft of the book. Riding high on that feeling of accomplishment, I wanted to get the book into the hands of friends and readers as quickly as possible. I naively thought that I was 99% done with the project, and assumed it would be as simple as uploading my files and pressing “Submit” to get the book distributed.

Alas, self-publishing has a steep learning curve. It quickly became clear to me why a lot of writers end up working with agents and publishers, rather than go through self-publishing themselves. It often felt like navigating through a bureacratic maze, with random roadblocks and unnecessary platform complexity.

It took me an additional 2.5 months to work through all the logistics, adding an additional 18% of time to the project timeline. A lot of that time was sitting around and waiting, working in fits and starts after I finally got a response from the publishing platforms. Filing a support ticket for help was like shouting into a dark void.

Here were the key hurdles I had to clear in order to publish my book:

• Forming a LLC
• Purchasing ISBNs
• Creating a gumroad website
• Selecting print distributors
• Setting book pricing
• Resolving an automated content integrity block

There are many possible choices that ultimately produce a physical book so it’s very easy to get stuck in analysis paralysis. I am grateful to Nicole Tietz-Sokolskaya and Will Larson for their self-publishing write-ups. They both have clear suggestions which helped resolve a lot of the ambiguity of self-publishing. It’s great to learn directly from folks who have already gone through the process, and now it’s my turn to give back.

Forming an LLC

I wanted to form my own publishing company, Odd Imprint Co LLC, in order to have separation of responsibilities and liabilities. It’s unlikely (but legally possible) that someone could sue the company for getting injured while following my travel guide recommendations, and having a separate incorporated company with it’s own funding means that my personal assets are protected. Since I’m based in California, I followed the CA Secretary of State guidelines for the formation and filing steps. I use Northwest Registered Agents to provide privacy so I don’t have to list my home address on company records available to the public. I chose Mercury as free business banking for the company since my husband had a good experience with them when he founded his own startup.

The initial costs to form the LLC totaled up to $1,026:

• $70 for articles of organization
• $125 for registered agent service
• $20 for statement of information
• $800 for annual CA business tax
• $11 for the oddimprint.co domain

View this profile on Instagram

Odd Imprint Co (@oddimprintco) • Instagram photos and videos

Purchasing ISBNs

When self-publishing, you need to acquire an ISBN in order to uniquely identify your book. Most publishing platforms like Ingram Spark and Amazon Kindle Direct Publishing can provide a free ISBN, but it limits your usage of the ISBN to that specific platform. For flexibility and tracking sales/reviews across the publishing ecosystem, purchasing your own ISBNs is a better option.

Bowker runs a profiteering scheme by selling 1 ISBN for $125 and 10 ISBNs for $295. Since I plan to produce at least four travel guides, one for each season, the 10-pack of ISBNs was the most sensible option.

These ISBNs are owned by Odd Imprint Co LLC, which will be my imprint for any additional books I publish too.

Creating a gumroad website

I planned to distribute a digital version of the guide as a PDF because I didn’t want to compromise the book design by converting it into ePub or Kindle file formats. I chose Gumroad for my online storefront since it was much simpler to set up. It does take a bigger cut (10% of sales) and offers less customization options compared to Stripe, but it’s much quicker to set up. I can also switch the digital distribution platform if my sales ever get high enough to warrant building out a more custom Stripe checkout process.

I have also seen Gumroad being used effectively by a lot of other self-published book authors, such as:

• Koya Bound by Craig Mod
• Cool Tools by Kevin Kelly
• The Tech Resume Inside-Out by Gergely Orosz
• The Art Director Toolbox by Jason Murray

Check out the book’s Gumroad page here.

Selecting print distributors

I use Amazon KDP and Ingram Spark as my print distributors. Most self-published authors choose both because you get higher royalties by submitting to the platforms directly. Every platform has the option for expanded distribution which helps automate the process of connecting to the other platforms, so you could use KDP or IS to distribute everywhere. However, they take an additional fee to act as a middleman and by listing directly you don’t need to pay that unnecessary cost.

KDP allows you to list directly on Amazon and IS has the largest distribution with book retailers and websites, such as Bookshop.org, Barnes & Noble, and even libraries. I briefly considered self-distribution, which would entail using a print-on-demand service like Blurb and MagCloud and selling the physical copies through Gumroad. However, I really wanted the ability for my book to appear on book retailer websites and (maybe even one day!) be sold in physical stores. Just personally for me, it didn’t feel like “real” self-publishing unless I could see my book listed in all the channels where I regularly purchase my own reading books.

Very early on into the project, I knew that I wanted my book to be printed at the standard A5 trim size (5.83 x 8.27 in). I wanted the book to feel compact but not overly dense, and I tested by going through my personal library and hefting various book sizes in my hands. A5 books felt the most balanced and familiar to me, since it’s the industry standard trim size for memoirs and novels.

I also needed to decide what grade of paper and print color to use. The guidebook is filled with painstakingly edited photographs and hand-drawn illustrations, so I wanted to use the most premium paper and print quality. I chose these corresponding print settings:

• KDP: premium color interior with white paper
• IS: ultra premium color (perfect bound) on white paper

Print options available for book configuration on KDP.

Reformatting the PDF files for print distribution was extremely tedious and a major part of the self-publishing learning curve. Each platform has their own guidelines for cover layouts, interior file formats, and bleed settings (KDP, IS). Every platform also has a quality review step before printing is even allowed. This review steps is mostly automated, but times can vary from 10-15min on KDP up to a few days on IngramSpark.

Then there was even more revision rounds once physical printing commenced. I would order a print proof, waiting several business days (up to a week) for delivery, check for print errors and formatting mistakes, and then correct the files. Each round I would notice something incorrect or off, so would have to reinitiate the process all over again. If I thought it was a printing error, I would have to open a customer support ticket and wait another week for a response.

Ultimately, I ordered 4 copies from Amazon and 3 from IngramSpark, including rushed delivery fees. This cost me $130 on just proof copies alone.

All the author proofs I purchased to spot printing and typesetting errors.

While the magnitude of my print review experience was perhaps exaggerated because of the complexity of the design layout for my book, which made it more prone to printing errors, I now understand on a deep emotional level why most authors would rather go take their chances with traditional publishing. With self-publishing, you not only act as author and editor, but also need to have skills to act as a production editor and quality control expert.

Setting book pricing

I knew beforehand that book publishing was not a lucrative business, but I was still surprised by how low the profit margins are on print books. For the digital PDF, I priced the guide at $20. Gumroad charges 10% plus 50¢ and sales tax for every sale, which means after all the fees, my take-home is a nice $16.46.

There are extra printing costs and wholesale discounts for print sales. Amazon pays a royalty rate of 60% but subtracts the printing costs from the royalty.

On top of printing costs, Ingram Spark suggests incentivizing retail stores to stock your book by providing a wholesale discount and making it returnable. The typical recommended wholesale discount is 55%.

Given the high printing costs of the ultra premium color, I had to lower the discount to 40%. I would have been printing at a net loss if I set the discount percentage even higher. Compared to the digital copy, I make only $1-6 on every print book sale.

Distributor	Retail Price	Costs	Profit
Gumroad	$20	$3.59	$16.46
KDP	$29.99	$11.98	$6.01
IngramSpark	$29.99	$16.10 (plus 40% wholesale discount)	$1.59

Resolving an automated content integrity block

Perhaps the most infuriating part of my self-publishing experience was being unfairly blocked on Ingram Spark. I had uploaded my formatted files and submitted them for quality review. I wasn’t expecting any issues since I had already submitted my files to KDP and requested my first round of print proofs. At this point, the administrative drudgery was annoying but manageable, and I was still coasting on the feeling of being so close to finished with the project. Then, unexpectedly, I received an ominous email titled “Catalog Integrity Notice” which brought all my hopes and dreams crashing down.

Email I received on the morning of Christmas Eve informing me of the catalog integrity violation. Merry Christmas to me, I guess.

At the end of 2023, Ingram Spark announced new guidelines to guarantee the quality of their distribution catalog. In particular, they wanted to ensure that “original content” was being produced. This seemed to mean removing content that was not only directly copying but also similar, whatever that means, to other copyrighted content. Content that was mass-produced, such as writing “generated using artificial intelligence”, would also be blocked. Finally there was a confusing stance about public domain works, stating that the platform “do[es] accept public domain content but may choose not to sell a public domain book if its content is undifferentiated or barely differentiated”.

Finally, with almost a hand-wavy air, Ingram Spark concluded their announcement with the barest of reassurances:

Most IngramSpark authors and publishers won’t be impeached by our catalog integrity guidelines.

Yet here I was, almost at the finish line of producing my most complex original work ever made with all original copyright or free-to-use content, getting slapped with an integrity violation notice. The consequence of being in violation was that I could still print my book through Ingram Spark, but it would not be available for distribution on their network, meaning that none of the retailers would have access to it. The entire of goal of self-publishing was in danger of failing! And in the notice itself, there was no explanation of which rule in the guidelines was in violation.

I had almost no recourse. You could fill in an integrity appeal to explain why you were not violating the rules. However, since IngramSpark did not explain what guidelines you had violated, I was only making my best guesses. I wrote at length about how this was all custom work, and tried to give proof with screenshots of my Google photos and writing timestamps. I filed multiple customer support tickets but the kind human support staff who wrote back said they had no access or visibility into the content integrity decisions.

One of the five identical rejection notices I received.

Thus started a month of blind back-and-forth with “Gale” from the Ingram Content Group, who I am pretty sure is just an automated bot. Each time I would file an appeal, within a few hours I would receive the same unhelpful rejection notice from Gale. I spent hours looking reading forums, watched Youtube videos, and found other authors in the same dire straits. Thad McIlroy ironically had his book about AI flagged by the content integrity AI, and he was only able to get his violation resolved by going viral on LinkedIn and catching the attention of real humans on the content team.

Desperate times called for desperate measures. After two unhelpful responses from Gale, I finally decided to compromise the integrity of my content by systematically testing potential violations. I wasn’t sure exactly what was tripping the automatic integrity detector, but I had a few theories in mind:

• Public domain woodblock prints
• Screenshot of a NYT cover image
• Snippets of lyrics from music on Spotify

Possible suspects for what was flagging the automated content integrity system.

So I resorted to basic guessing-and-checking. I would remove an image or piece of content from the interior files, upload it to Ingram Spark, wait for it to pass quality review, and submit a new content integrity appeal. I did this six separate times, each time receiving an automated response saying it could take up to “10-15 business days for my appeal to be reviewed and addressed.”

Finally, one of my attempts removed the problematic image that was secretly getting flagged. It was in fact the New York Times cover image, which I was using to actually encourage readers to go read the article more in-depth. I switched to a personal photograph instead and finally “Cassandra” from the Ingram Content Group team approved me for distribution.

The required change for interior files in order to pass the content integrity checks. Fortunately, I did not have to make any additional changes beyond this spread.

At last! I was finally approved for distribution.

I’m not sure if there’s anything to learn from this episode besides being prepared for all kinds of unexpected roadblocks. If platforms continue along this path, adding more automation that operate on a set of hidden and esoteric rules, it’s the smallest creators that will always bear the brunt of the pain and inconvenience.

Next steps

Despite all the stops and starts, I would recommend writing a book. There was a lot of frustration and roadblocks along the way, but I don’t regret it in the slightest. I’m excited to start writing the second book in the travel series since I feel like I’ve finally figured out many things which were only scary because I was doing it for the first time.

I should also probably be spending significantly more of my time marketing and promoting this first book, a set of behaviors that honestly is very unnatural for my personality. I’ll probably need to write another set of reflections on my attempts to muddle through self-promotion.

Thanks for reading, and once again¹, here’s how to get the book:

• Digital PDF is available here on my website
• Print book is available on Amazon, Barnes & Noble, and Bookshop.org

1. See how awkward I am at this transition? ↵

---