Thesis¶
This page contains my business hypotheses for my language learning app, Banyan Flashcards.
The sections come from those described in The Four Steps to the Epiphany by Steve Blank.
Customer and their problem¶
Customer¶
The original customer was me - I made the first prototype for this app in 2019.
My theories about the ideal customer of this app...
- Learning a language
- Perhaps generally interested in language learning as a hobby
- Possibly at the intermediate level, at which point they want to expand their vocab in order to widely apply the many sentence structures they've already learned
- Exposed to the language they're learning regularly
- Perhaps a heritage learner
- Learning to communicate better with their partner's family, or to get to know their partner's culture more deeply
- Immigrated to a country in which they don't speak the language particularly well
- Has a busy schedule
- In full-time work or study
- Perhaps has a family
- Tech-savvy in some sense
- Sympathetic to novel applications of everyday technology
- Has tried other language learning apps, such as Duolingo
- Has possibly even paid for these other apps
Their problem¶
They want to improve in the language they're learning, but they feel like they don't have much free time in their day, so they struggle to fit traditional studying into their schedule. Even popular, gamified apps like Duolingo don't fit so easily into their life. It isn't just about how much time you have in the day, it's also about how much energy you have within your day for extracurricular learning.
At the same time, methods that rely on fixed windows of practice are less effective, because you only think about them for the 10 or 15 minutes that you're practicing. For the rest of the day, you're not thinking about the language you're trying to acquire at all. The fact that you're not immersed makes it much harder to retain language skills as you acquire them.
Product¶
Core¶
By combining phone notifications with a spaced-repetition algorithm, we can make it possible for a person to learn and gain long-term recall of language vocabulary without ever studying or ever putting effort into memorising vocabulary. This means it requires less time and less mental energy, yet at the same time, it's also more effective.
We let users enter words they want to learn into the app, and from there the app takes care of making sure that the person memorises it by sending interactive flashcard notifications for those words.
Extensions¶
Learners need to practice using these words in context. There is a sentence practise section, which lets users enter sentences and have a large language model (LLM) give them feedback on their grammar and how idiomatic their use of the vocabulary is.
Learners acquire vocabulary best when they learn it in context, for example, from reading. There is a reading list feature which lets users paste entire texts into the app, from which it then extracts words that the user doesn't know yet and proposes them for addition into their vocabulary corpus.
Competition¶
Summary¶
The competition includes Duolingo, Anki, Daily Hanzi, and the multitudes of other smaller language learning apps, traditional self-study using a textbook, attempted immersion via movies and TV shows, and other sources of comprehensible input.
Other apps are more comprehensive in what they teach. They also attempt to teach sentence structures, grammar concepts, listening and speaking. We don't attempt to teach any of that directly. This app is mostly for self-directed learning.
Other apps are more specific in what they teach. There are apps dedicated to specific languages, particularly the Asian languages like Chinese and Korean. This app is generic.
There exist many new apps which use AI to generate comprehensible input, sometimes based on a vocabulary provided by the learner. Plenty of apps use AI to generate flashcards, which users can review and study in the app. This app provides the possibility to use AI to generate individual pieces of vocabulary, but it doesn't (and won't) generate entire texts, because I believe learners should acquire language from native human input.
One thing that all other apps have in common is that they primarily rely on lessons in the app to teach learners. This means that they are time intensive, which makes them difficult for learners with busy and variable schedules to stay consistent with.
Typical pricing for other apps is between $5 and $25 per month, for monthly billing, and slightly cheaper for annual billing. Personally, I think this is too expensive, especially now that such software is relatively easy to produce. This app is cheaper as well as better.
Competitive advantage¶
Large incumbents won't use my idea because they already have effective engines of growth.
Smaller incumbents which use the traditional interaction mechanism (in-app lessons) won't use this idea because it saturates and muddles their feature set and the way they interact with the learner. I think it would be quite difficult for an app that combines both interaction mechanisms to still provide a coherent overall learning experience. Nevertheless, it would be possible. The mechanism alone isn't a durable advantage. A competitive go-to-market approach is necessary for sustained growth.
There are other smaller incumbents like Daily Hanzi who are already invested in passive interactions, for example via smartphone lockscreen widgets. It would be relatively easy for them to incorporate the core product idea of Banyan. Luckily, they are language-specific. If they did decide to reuse my idea, it probably wouldn't be difficult for them to capture the market for Chinese and Korean learners, even with an inferior product, because they have an effective, existing distribution strategy.
Channel and pricing¶
We will price ourselves at the lower end of the range because our solution is simpler and our solution is also best used as a complement to other learning methods.
Customers will find out about our app via the same channels that they find out about our competitors - through social media and hopefully plenty of referrals - we will make referrals a first-class priority, partly because the team (just me) is not skilled at distribution, but also because our app is quite novel, so we don't really yet know how to explain it to users - that is best done by our much-treasured early believers in the product.
A small and unique advantage of this app is that the way it's designed means that it doesn't contribute significantly to your screen time. Many consumers in my age group are actively interested in reducing their screen time - although I don't think it's especially motivating to them, it's more of a nice to have.
Market type¶
We are entering an existing market, looking to re-segment it based on a niche. This means we need to prove that the niche advantage we provide to learners is worth paying for, even at the expense of other features provided better by existing competitors.
Demand creation¶
If we're able to convince our target audience that it's possible to expand your vocabulary without putting in any effort, then I think learners will flock to that solution, because that's such a compelling story. It almost feels too good to be true - that you can improve without even trying. We need to show our target market that this is a reality, not just an oversold story for marketing. If we succeed at doing that, then people will buy our solution.
Created on 2026-07-21
Updated on 2026-09-08