The Weekly is a weekly essay where I write about something I’ve been thinking about in the last seven days. They’re under 1000 words, and this week I want to talk about vision and problems when you’re building a new product.
I’m building the lexicon, where I am (un-ironically) trying to build the world’s most useful language learning resources. I like Duo Lingo, but I don’t think it helps you much in the real world ™️. It’s born out of my desire (and previous moderate success) to be come proficient in French. The pandemic slashed my French proficiency, though.
Over the past few weeks I’ve been thinking more about this product idea. I think there’s enough scope in the vision of “best language learning resources in the world” that I can meaningfully do something new. I think I know enough about computers and learning to do something impactful.
The first thing I did? Set up a landing page with a sign-up form for a mailing list, and released two (free) sets of flash-cards with the most common words. If you want them, you know where to find them.
None of these things (which took real time and energy) solve any of the core problems that I’m trying to solve, though. I’ve got a vision for the company, and these things move me generally in the direction of it. I can’t sell anything if people can’t find where to buy it.
But I don’t think I solved the high value problems.
In his essay how to get startup ideas Paul Graham says:
You should only work on problems that exist. And yet by far the most common mistake startups make is to solve problems no one has… The verb you want to be using with respect to startup ideas is not "think up" but "notice”… The most successful startups almost all begin this way.
I created this vision because I noticed a problem: I want to learn a language in an efficient way, that gets me to real-world use quickly. Having a landing page does nothing to solve that particular problem.
What Graham says is (probably) true, but I also think they miss out on another dimension of problems: the specificity / uniqueness of them.
Getting Specific: Flashcards
Let’s get specific.
I produced some flashcards sets. These keeps within the vision, but what problems do flashcards solve?
- Target-language translations of words (être ➡️ “to be”)
- They decide what word’s you’re learning
- The can be used in Active Recall
This list, broadly, runs from most generic to most opinionated solutions.
1. Bilingual translation Could you learn French with resources which just showed you the words in the target language. Verser, manifestation, et vaccin.
Have you just learned to pour/contribute, protest (noun), and vaccine ? No. You maybe could have guessed the latter, though.
To remove the bilingual elements of language-learning flashcards would be an unexpected feature. Maybe replacing them with image/sound can help (an auditory component for French ↔️ English translations would be beneficial) but largely, you need the bilingual text. This isn’t a high-value problem.
2. Decide what wordsWhat words should I learn? This is a more interesting problem. It’s actually one of the core problems that lead me to think about the lexicon. I don’t think we teach useful words to new language-learners.
The problem of “what words do we learn, and in what order” is a problem worth solving. So I used data-backed findings to decide on the most common verbs. This is a higher-value problem.
3. Encourage Active Recall Active Recall is, almost irrefutably, a core component of any language-learning practice. When you use flashcards, you should see the prompt and guess the answer (actually guess it, like, say it aloud) before viewing the answer. Anything else is passive recall or lying to yourself.
Active recall is essential to any effective language learning practice, and I was leaving it as optional. “Here”, I said, “are some flashcards. You know how to use them effectively, right?”.
A foolish assumption that was probably false (no one teaches these skills de facto in schools) which essentially let the user determine exactly how effective or useful the language-learning resource was going to be
I had a chance to solve an important and unique problem (how to pair the “right” words with the “right” learning method) and I delivered it in static flashcards. I moved towards the vision without solving the particular problems: actually increasing your French vocabulary.
I left the hard work of solving the problem to the customer. I shifted responsibility.
Visions are slopes, problems are places
I was guided by a vision, without thinking about the unique and important problems.
A vision is like a slope. It’s easy to know when you’re heading up or down hill, and you can conform to a vision by degree: head directly down hill, or go diagonally.
If a vision is a slope, a problem is a place. Or a least a region. The further away from the problem you are, the easier it is to head towards it. When you haven’t started, doing almost anything will get you closer. Setting up a landing page and a mailing list, for example.
But you get a bit closer and you’ve got to change bearing. You were heading east, but now you’ve got to head north-east. And then north-north-east. Anything other than that and you start veering away your destination. Maybe you’ll find a different problem along the way (the infamous pivot) or maybe you won’t.
A mailing list, a landing page, and some static flashcards are not the world’s most useful language learning resources. They’re part of the vision but not the product,
Set vision, solve problems.