What we would’ve done differently in the first year of bootstrapping our company
We've learned a lot in the last year -- from specific growth tactics to the importance of planning for the long-term.
tl;dr
We wished we would have…
Started content marketing earlier
Targeted specific keywords with our SEO content
More deeply understood the experiences of individual users
Started doing one-on-one tutoring earlier
Taken compliance stuff more seriously
Taken finance and accounting more seriously
Approached Arno as a multi-year marathon, not a few-month sprint
Things we wouldn’t have done differently
Talked to a lot of users before investing a lot of time into building the product
Shipped quickly & learned by doing
Did the work ourselves
1. Started content marketing earlier
⭐ Key takeaway: Building out an acquisition channel takes a lot longer than shipping products.
The main reason to start content marketing earlier is it takes take time to build a following. The algorithms powering platforms like Google, YouTube, Instagram, TikTok, and Twitter are designed to take months and years to build an audience. We learned that we can ship products much faster than we can get users. As we discussed in our other post about how we got 60,000 users with SEO, it took 3 months before we saw any meaningful traction from our SEO articles. We’ve been on YouTube for over 6 months and have published 26 videos. We have a little over 3,000 subscribers and we average about 1,500 views per day. These numbers aren’t terrible, but the channel that dominates the “Duolingo English Test” topic on YouTube has over 500,000 subscribers and probably gets tens or hundreds of thousands of views per day. He has also been creating content for over 3 years. It takes time and consistency.
Another reason to do content marketing earlier is that we learned that marketing can inform what we build. When we started out, our view of marketing was that it is simply how you acquire customers, a view that I think many other technical/product founders share. Getting customers is indeed the main role of marketing. However, it can be a lot more than that. Marketing can help you figure out who your customer is and, in particular, what problems they have, which is the core work of a startup.
With content marketing especially, you are trying to answer some specific question that your target users have. If you make a bunch of great content marketing, you can easily test out a bunch of theories about what users are actually struggling with and what they actually care about. It’s like you’re shipping a bunch of MVPs to validate your theories and learn about users. Moreover, in order to create the content, you have to become an expert in the current solutions available. This enables you to spot gaps in the market and get ideas for what you could build.
For example, some of our best-performing content teaches people vocabulary. The success of this kind of content has compelled us to focus on shipping features that help people learn vocab. If you put marketing before product, you can use marketing to learn about your users and help you to build better products.
2. Targeted specific keywords with our SEO content
⭐ Key takeaway: Each article you write you should target a specific keyword that people search for.
Our first set of articles were in-depth guides to each of the question types on the Duolingo English Test. This seemed like useful content to me, so I figured it would do well in Google.
It did not. I learned that users don’t actually search for information about specific question types. This means that our articles with titles like, “Fill in the Blanks - Guide to Duolingo English Test Question” does not appear in a lot of search results because people don’t search that much for things like “fill in the blanks duolingo english test”.
Not targeting keywords is a classic rookie mistake with SEO. You have to be strategic about the content you write. We have a full walkthrough of the SEO strategy we eventually used for our content that will help you avoid this mistake — and others that we made 😄
3. More deeply understood the experiences of individual users
⭐ Key takeaway: It’s important to have rich qualitative data, not just quantitative data.
I think we have done a great job of using data. In everything from our marketing content to our product’s features to our finances, we keep a close eye on the numbers. But we learned that quantitative data doesn’t give you the whole picture. You also need the qualitative data, which is just a fancy way of saying that you need to understand the actual experiences people have using your product, ideally in their own words.
Here’s the biggest example of how we got this wrong. We had a user whom I call Martin. Martin used the product a ton. He answered thousands of practice questions. He bought hundreds of credits for feedback and even did a number of tutoring sessions with us. From our perspective, Martin was a power user who was getting a ton of value out of Arno. Otherwise, why would he keep using it?
A few weeks ago, we got an email from Martin informing us that despite all his work in Arno over many months, he did not get the score that he needed. He took the test 4 times but always came up just 10 points short. Martin trusted Arno to help him get the score he needed, and we let him down.
Had we been talking to Martin, had we understood his lived experience using the product, we likely would’ve understood that all his practice was not a good sign. Rather, it was a sign that he was stuck and couldn’t improve his English. He needed more than what our features could offer him. This experience allowed us to set a clear focus for the next phase of our roadmap: No More Martins. This is what we’re trying to solve now — preventing users from getting stuck at a certain level of English.
It’s critical to put yourself in the shoes of your users. When was the last time you went through your sign up flow? When was the last time you used your app for more than 10 minutes? When was the last time you watched full user sessions? For us, the answer to all these questions, we realized, was: “A long time ago, maybe never.”
The importance of having relationships with users is a lesson we seem to learn and forget over and over. As I write this blog post, I’m realizing that we are not practicing what I’m preaching here. This tells me that for whatever reason, this is a difficult thing to do, even though we know how valuable it is. So, I’m re-prioritizing getting users on the phone now 🙂
4. Started doing one-on-one tutoring earlier
⭐ Key takeaway: Don’t be afraid to get creative in order to talk to users and build relationships with them.
One of our early struggles was simply getting users on calls. We knew it was important to talk to users, but we just couldn’t make it happen. First, we didn’t have that many users to ask. Second, our users are not native English speakers, so doing a full interview in English is intimidating. Third, they are high school and university students, so they are unaccustomed to meetings and video calls. The no-show rate was maddeningly high.
Eventually, Makoto proposed that we start offering tutoring sessions. In these sessions, we could ask people about how they were using the product and get their feedback. We got students right away. Since then, I spend several hours each week working with users one-on-one on their English. I get to see where they are struggling. I get to test out different methods of helping them improve. I get to learn about how they are preparing. These tutoring sessions are a great way to deeply understand our users, their problems, and what they really want.
As Makoto and I talk about new features, I constantly reflect on the tutoring sessions for insights. We’re able to think about how it would work for specific users that we know well. At the beginning of Arno, we were talking about some vague concept of “our users” that really just consisted of our guesses about what they wanted. In our discussions now, we refer to users by their names all the time. We say things like, “I think that Maria would find that helpful, but I don’t think Franc would use that.” These opinions are grounded in our understanding of real users, which means they’re much more likely to be accurate than whatever we think up in our imaginations, like we were doing before.
Of course, doing one-on-one tutoring doesn’t scale very well. But that’s why you should do it. Things that don’t scale well are things that require you to get hands-on. You have to deeply understand the user and their problems so that you can try to help them. Once we started doing that, we started shipping features that users engaged with a lot more.
I would encourage you to get creative about how to get hands-on with users. You could start by doing consulting. You could offer to build a custom solution for very cheap. Whatever the method, a good test is whether you can think about individual users by name instead of some faceless, vague entity called “our users”.
5. Taken compliance stuff more seriously
⭐ Key takeaway: If you’re building a business, there is a lot of unsexy, tedious paperwork that you have to be on top of.
Starting a business entails a lot of paperwork. No one told us this. The incorporation process alone entails 12 different pieces of documentation. After you incorporate in one state, you have to register in all the other states you do business in. Then, you have to register with all the tax agencies at the federal, state, and local levels. You have to set up payroll, which is more paperwork. Oh and don’t forget about bank accounts and setting up Stripe and such.
In short, there is a lot of compliance you have to take care of when you start a company.
We learned the hard way that if you do not take care of it promptly, you will pay. For example, we didn’t file an annual report with Delaware (where we are incorporated) on time. We had to pay a $300 fine. Ouch. We didn’t set up all the accounts properly in Pennsylvania when we started paying me, and we had to pay a few hundreds dollars in fines. Ouch ouch.
It’s not that we completely ignored these things or planned not to pay what we were supposed to. We just put it on the back burner. Now we know that you can’t do this. Compliance stuff goes on the front burner, always.
You can pay to have these problems solved for you. But these services are expensive. We didn’t want to pay for them before we saw for ourselves how difficult it actually is. (We found that it isn’t that difficult, so we concluded that these services are way overpriced.)
Unfortunately, a lot of this compliance stuff is really confusing. To make matters worse, there is shockingly little good information out there to help you through it. If any accountant, tax specialist, or business backend process expert of any kind is reading this: There is an SEO goldmine waiting for you if you write clear, step-by-step instructions for compliance stuff for bootstrapped C-corps.
(I plan to write about what we figured out as far as compliance in future posts, so at least that will provide people with something better than what we managed to find.)
6. Taken finance and accounting more seriously
⭐ Key takeaway: Money is the lifeblood of your company, so you should always thoroughly understand its financial state.
A business can be boiled down to one simple question: How much money is coming in (revenue and funding) and how much is going out (expenses)? If revenue and funding is greater than expenses, then you can survive. If expenses are greater than revenue and funding, you are on track to go out of business.
Everything you do — marketing, product, strategy, customer support — ultimately comes down to your finances. That’s why you have to keep a close eye on them. For the first 6 months or so of Arno, we didn’t do this. We spent money and brought in some money. We didn’t know if we were profitable. We didn’t really know if expenses or revenue were growing. We could get away with this because we had full-time jobs that paid us well. But as soon as we got more serious about going full-time on Arno, we needed to put into place some processes to keep track of our money.
The good news is that it’s relatively simple. At this point, it takes me like 3 hours a month to “do our books”, meaning tally up our revenue, list out our expenses, add some numbers together, and figure out how much in profits we made. But, it took a few months to make the process this streamlined.
I will write a whole blog post about how we keep our books but for now, here are the key aspects of our process: At the beginning of every month (after all the bills have come in and the bank and credit statements are processed), I look at the previous month. I create a new tab in a spreadsheet and:
List out all our revenue (Stripe, Paddle, App Store, etc.)
List out all our expenses
Add up all the revenue
Add up all the expenses
Subtract total expenses from total revenue
Get our final profit number
Compare the different numbers to previous months to see changes over time
It’s important to list out all the numbers like this so that you can see changes in different line items over time. By doing this, we discovered that we were paying an arm and a leg to Mailgun for our marketing emails, so we built our own system and it’s going to save us hundreds if not thousands of dollars this year alone 🥳
7. Approached Arno as a multi-year marathon, not a few-month sprint
⭐ Key takeaway: Adopting a long-term mindset allows you to enjoy the journey more and stay on it longer, which increases your chances of success.
We started Arno when we both had full-time jobs. As we started getting users and money started coming in, Arno became a lot more exciting than our day jobs. We wanted to go full-time, but Arno wasn’t bringing in nearly enough revenue. We wanted to be prudent, so we planned to first get to ramen profitably and then quit.
We looked at our growth rates and made some projections and anticipated that we could get to ramen profitability in 6 months. That became a mantra for me: 6 more months in this job that I hate then I would be free and living off Arno.
This was a bad mindset. First, our projections about how long it would take to get to ramen profitability were basically bullshit. “If we can grow 25% month over month, then in 6 months we will get there” is what concluded. 25% growth each month didn’t seem like much. Plus, we told ourselves, you just have to set a goal and hit it by any means necessary because that’s what startups do!
We learned that startups are not that straightforward 🙂 Sure, it’s good to have aggressive goals, but you can’t just make growth happen. “Growth is not linear” is how Makoto phrases it. Many companies experience minuscule, but steady, growth for a long time, before they hit some inflection point and it takes off. This is why many startups are “overnight successes, years in the making.”
Thankfully, I only made it about 2 months longer in the job before I knew I had to quit because I was so unhappy. I was planning to get another job, but then Makoto floated the idea of me just going full-time on Arno. That’s what I ended up doing. I’ve been full-time for 8 months now and Makoto for about 5. We’re still not ramen profitable 😅
Looking back, the 6 month forecast until ramen profitability was unrealistic. We should have said 18 months, if not more. I don’t think that that would’ve made us any less aggressive. It certainly would’ve allowed me to think more seriously about the sustainability of my lifestyle. I would’ve thought about finding a day job that wasn’t so deflating or building up some part-time work writing content that I could do for 18 months or more.
At this point, I’m pleased to say that both of us wholeheartedly appreciate that building Arno into a business that makes us decent money is going to take years. We don’t even fret about the next few months, to be honest. We just show up each day, do the work, and trust that over many months, it will all add up to something successful.
Makoto recently got into jiu-jitsu and I’ve gotten into CrossFit. (I promise — we’re not actually bro-y in person.) It occurred to both of us that building a startup is like becoming an elite athlete. Getting a purple belt in jiu-jitsu, which is the intermediate stage, takes 5 or more years. Maybe in 10 years I can be in the top 10% of CrossFitters for my age group. Knowing that it’s going take a while allows us to sit back into the journey and just enjoy where we are now, instead of being unhappy that we are not where we ultimately want to get to.
Although part of me thinks that we needed to suffer under the burden of unrealistic growth expectations before we were able to change our mindset, part of me can’t help but feel that I wasted some good months of my life — stretched far too thin between my day job and Arno, constantly filled to the brim with stress, feeling like I had to half-ass everything. I very much hope that you can avoid doing this to yourself by adjusting your expectations now about how long it will take to get your business humming and paying the bills. If you can commit to the long-term nature of this pursuit, I think you will make the adventure more enjoyable and sustainable, which will also mean that your odds of success will be higher because you will be less likely to quit — and not quitting is basically what success in entrepreneurship comes down to.
8. Saved more money
⭐ Key takeaway: Your savings won’t last as long as you think they will.
While we had full-time jobs, we saved money, but we could’ve been more frugal. We ate out a lot instead of cooking at home. We took Ubers when we could’ve walked and taken the subway. We went out for drinks instead of having beers at home. We didn’t go to Ibiza or buy expensive things. But all the little expenses, mainly related to convenience, add up quick.
We also learned that our expenses didn’t decrease that much when we left our day jobs. It turns out that food, even in other countries, isn’t so much cheaper than what we were paying at Trader Joe’s. Once you work for yourself, you also have to pay for health insurance. If you’re digital nomads like us then you have to pay for flights and visas as well. And there’s always unexpected costs. For example, last week I learned that a tube of Neosporin costs $40 here in Tulum! 😲
Your savings will go a lot faster than you think they will. $50,000 in the bank may seem like a lot now, but you will be amazed at how easily it can be spent.
Since moving to Mexico, I’ve started tracking all of my expenses. I wish that I had done this for just one month while still living in Philadelphia because I know it would’ve made clear why I felt like I was earning so much money but never able to save as much I wanted to. It would’ve allowed me to get my spending under control and start saving more. I don’t think it would’ve changed my lifestyle much at all. So if you’re thinking of going full-time on your own thing, I would encourage you to get a firm handle on your expenses so that you can, first, save as much as you can, and, second, reliably predict your runway once you go full-time.
9. Things we wouldn’t have done differently
There were some things that we’re proud to say we feel that we did correctly.
a. Talked to a lot of users before investing a lot of time into building the product
⭐ Key takeaway: Aim to gain a general understanding of your users’ experiences around a particular topic or problem before going deep building a solution.
In our last startup, we spent 6 months building a product and only talking to our non-technical co-founders who were telling us what needed to be built. It was a financial modeling tool for real estate, and the co-founders were analysts themselves. Our belief was that if we built something they loved, then other analysts would love it to.
We launched with an amazing product and started trying to get people to use it. Immediately, it was clear that they weren’t going to use it. It would have been funny if it wasn’t so sad. In call after call, people would say, “Wow, this is so cool!” quickly followed by questions like, “But could you change this one thing?” or “Could you add this?” or “I like it but there’s no way my boss will go for it.” In 6 days of calls, it became clear that 6 months of coding was more or less a waste of time 🤦
This time around, we vowed to not repeat the same mistake. First, we talked to a lot of users before we wrote any code. We wanted to learn about their journey with English. What were their current struggles with English but also what were their past experiences and what were their future goals?
After about 20 interviews, we shipped an MVP for a product that helped busy professionals learn English while working, and we kept doing user interviews. Everyone said they liked the product, but no one actually used it. Our conclusion was that non-native English speakers working in the US want to improve their English, but they don’t actually have to. They already have a job. Their English is good enough. A lack of English is not really a pain point for them.
So we asked ourselves: When was a lack of English a pain point? We looked back at our user interviews and there was a clear moment experienced by person after person: When they needed to pass an English proficiency test to apply to college. At that moment in their lives, they had to improve their English, otherwise they would not be able to study abroad. That’s when we pivoted to focusing on the Duolingo English Test.
This pivot was only possible because we talked to so many users. When you are starting out, I would encourage you to adopt the mindset of a journalist or an anthropologist. Your goal is simply to learn about your users and their experiences around a certain topic, process, aspect of their lives, etc. Hold off thinking about solutions and just focus on understanding the problem as deeply as you can, as if you were going write a newspaper article or research report about it. This breadth of understanding will allow you to better define the problem your product solves and also pivot more easily if you need to.
b. Shipped quickly & learned by doing
⭐ Key takeaways: You learn how to build a business by building a business.
Once we did user interviews and had some hypotheses about what people wanted, we did a good job of building a simple MVP that we could get out there quickly. This helped us not only quickly test our hypotheses, but it also forced us to learn by doing.
The importance of putting something out there into the world so that you can learn by doing is something I’ve heard from entrepreneur after entrepreneur but, like a lot of advice, it didn’t really register until I learned it for myself. I’ll join the chorus: The best way to learn how to build a business is by building a business.
I’m a big reader, and I’ve read pretty much all the startup books and essays. I’ve listened to hours and hours of the Indie Hackers podcast and How I Built This. Makoto and I have spent a lot of time talking about startups. A lot of this has been helpful, but it does not come close to how much we’ve learned by shipping something that real people pay for and use. It’s very much like trying to become a good basketball player. You can watch a lot of basketball and study it intensely. But to actually get better, you have to go out and play.
This doesn’t mean that you have to learn how to code and ship an entire product. Just create something that you want people to participate in, read, use, and, ideally, pay for. You just want to get experience creating something new and trying to get people to care about it.
c. Did the work ourselves
⭐ Key takeaway: Don’t outsource a task before you do it yourself.
We’re bootstrapped, so we’re always eager to save money where we can. This means that by default, we never outsource tasks without first doing them ourselves. Even things like taxes we decided to try and do ourselves and not just hire an accountant.
This habit has allowed us to learn a ton. For example, we learned that taxes, surprisingly, are not that difficult. I’ve written over 100 SEO articles and have learned a ton about SEO. I’ve learned how to be a YouTuber. Makoto has greatly improved his design skills. We’ve learned how to operate every single aspect of the company. When we decide to hire someone to handle one of these aspects, we are much better equipped to know what needs to be done and therefore make sure we’re paying a fair price and that the job is being done well.
Personally, this diligence with the business has been immensely valuable in my personal life. I needed to get health insurance recently. In the past, I would’ve dreaded having to wade through all the complexities. Frankly, I probably just would’ve procrastinated until I started getting serious warnings or even fines that I really, really needed to buy health insurance. But, after having to figure out so much stuff for Arno, I took a few hours and figured out my health insurance. Simply put, it feels tremendously empowering to be on top of things and not have blindspots that make you vulnerable.