Before You Build, Get Honest With Yourself
Are you building something people admire, or something they urgently need?
If your product disappeared tomorrow, who would actually care?
When I started building Preplings, I assumed the hard part would be the technology. AI orchestration, CEFR mapping, adaptive logic, structured writing correction. That was exciting. That felt like progress. I could see the architecture in my head — the feedback loops, the scoring engine, the way a learner would move through levels with precision.
It wasn't the hard part.
The real difficulty had nothing to do with code. It had everything to do with market clarity — the kind that's slow and uncomfortable and forces you to question whether you're building something people actually need, or just something you find interesting. These are different things. They feel the same from the inside, and that's precisely what makes them dangerous.
I spent 6 months building before I stopped and forced myself to answer a set of questions I'd been quietly avoiding. Not because they were hard to find — but because I was afraid of what the answers might mean for everything I'd already put in.
Here's what I had to confront before I could build seriously.
1. A Real Problem Is Not the Same as an Urgent Problem
Language learning is a real problem. Millions struggle with it. The global language learning market was valued at roughly $69 billion in 2025, and it's projected to nearly double within the next several years. Over 1.5 billion people worldwide are actively learning a second language right now, with two-thirds of them using digital platforms to do it.
But real doesn't automatically mean urgent.
Most people wake up thinking about deadlines, work pressure, kids, money. "Improving my German" is important — just rarely top-three urgent. Products that don't attach themselves to pressure drift quietly into the nice-to-have category. They get downloaded, opened twice, and forgotten. The app stays on the phone as a kind of aspirational decoration.
That forced me to reframe the whole product. Not AI-powered language learning. But measurable CEFR progress for people under real time pressure — exam deadlines, visa requirements, career transitions. Urgency lives in consequences. When the stakes are real, the motivation is real. When the stakes are abstract, people procrastinate forever.
The reframe sounds subtle. It changes everything about how you build, who you talk to, and what you put on your landing page.
2. What Happens If They Ignore It for 12 Months?
This question hurt. If someone ignores a language app for a year, what actually happens?
They progress slower. They stay stuck. Maybe they miss an opportunity. But that's vague — and if the downside is blurry, the willingness to pay will be blurry too.
Vague consequences produce vague motivation. Vague motivation doesn't convert.
So I pushed myself to get specific. What does "ignoring it" actually cost a real person?
A student who delays preparing for their B1 certification misses the application window for their university course. A professional who puts off reaching C1 watches a promotion go to someone who can present in German. An immigrant who doesn't hit B1 in time faces delays on their settlement permit — a process that is already exhausting, bureaucratic, and emotionally draining.
The Goethe-Zertifikat B1, for example, is formally required for a German Citizenship. It isn't optional. It isn't a nice credential to have. Failing or missing it has hard, bureaucratic consequences that delay someone's life in a very tangible way. That's the kind of downside that makes a product necessary, not optional.
When the cost of inaction becomes concrete, the product becomes sharper. I stopped describing Preplings as a learning tool. I started describing it as exam preparation infrastructure for people who cannot afford to fail.
3. Are People Already Paying to Solve This?
Originality is seductive. But markets don't reward originality!
they reward existing spending behavior.
This was one of the most clarifying things I learned. Before building something ambitious, the question isn't "is this a good idea?" It's "is money already moving here?"
In the German language certification space, money is very clearly moving. People already pay €150–€250 for a single Goethe exam attempt — and if they fail, they pay again. They pay for TELC preparation courses that can run into the hundreds of euros. They pay private tutors €40–€80 an hour to drill them on writing and grammar. They buy prep books, practice tests, and structured online courses. The Goethe-Institut alone conducts exams at hundreds of locations worldwide, and the B1 level alone saw substantial volume in 2026.
This matters deeply. If nobody is already spending money in your category, you're not just building a product — you're first educating an entire market about why they should care. That's an enormously expensive problem. It can take years and millions in marketing just to create demand that doesn't yet exist.
In the German language certification space, I didn't have to create demand. I had to intercept it.
That shift — from creating demand to intercepting existing demand — changes your entire go-to-market strategy. It changes how you write copy, who you target, what promises you make, and how you price.
4. Is This a Company or Just a Feature?
This was the most uncomfortable question to sit with, and I avoided it the longest.
Could Preplings become a feature inside Duolingo? Or LinkedIn Learning? Or some AI mega-platform that decides German certification prep is worth a product sprint?
Duolingo generated $748 million in revenue in 2024. They have 500 million registered users and an engineering team that can ship new features in days. If a company like that decides to build what I'm building, and they build it broadly, what happens to Preplings?
If my answer was "we'd probably lose," then I didn't have a company. I had a capability.
The only real answer to this is depth.
Not features — depth.
A large platform can copy a feature in a sprint. They cannot easily copy a rigorous, purpose-built methodology that is specifically optimized for CEFR exam success, that has structured writing correction calibrated to the exact rubric that Goethe examiners use, and that tracks measurable skill progression in a way that's meaningful to learners under real deadline pressure.
Duolingo is built for engagement. Preplings is built for certification. Those are genuinely different goals. Engagement maximizes time spent on platform. Certification maximizes demonstrable outcomes before a specific date. The product decisions that follow from each goal diverge completely.
That distinction is a moat — not a perfect one, not an impenetrable one, but a real one. It requires a competitor to care enough about this specific outcome to rebuild from different first principles, not just ship a new module.
Depth is harder to copy than features.
5. "Everyone" Is Not a Market
Early on, I told myself Preplings was for "anyone learning German." That felt expansive. Ambitious. It was also completely useless from a product, messaging, and growth standpoint.
Here's the problem with building for "everyone": you end up building for no one in particular. Your features try to serve too many different needs. Your messaging tries to speak to too many different motivations. Nothing resonates sharply with anyone because everything is tuned slightly wrong for every specific person.
Broad audiences feel safe. Narrow audiences convert.
The language learning app market had 231 million downloads in 2023. The users behind those downloads have radically different goals. Some are casual learners building a holiday vocabulary. Some are heritage speakers reconnecting with a family language. Some are students trying to pass a standardized test. Some are professionals facing a career requirement. These groups don't just need different features — they need entirely different products, different tones, different success metrics, and different pricing.
When I narrowed to specific segments — Gymnasium students preparing for ZP10, professionals targeting B1 for a visa or settlement permit, adults aiming for C1 before a career move in a German-speaking environment — everything sharpened almost immediately. Messaging improved because I knew exactly what pressure I was speaking to. Feature priorities became obvious because I knew what outcome I was optimizing for. The product finally had edges, and edges are what make something memorable.
There's a counterintuitive principle here that took me a while to believe: the more specifically you define who you're for, the more confidently people self-select in. When someone who is three months from their B1 exam lands on a page that speaks directly to their exact situation, they don't think "this is too niche." They think "this was made for me."
6. Can You Sell the Outcome Before You Build It?
This was the most humbling section of my pre-build checklist, and it's probably the one most builders skip.
The question is simple: if you stripped away all the technology, all the AI, all the polished interface — and you offered the core outcome manually, would someone pay for it?
If I couldn't convince someone to pay for a structured CEFR audit, or focused writing correction done by hand, or a personalized study plan built from scratch, then adding AI layers on top of that wouldn't change anything fundamental. It would just be a faster, cheaper version of something nobody wanted in the first place.
Automation scales value. It doesn't create it.
When people hesitate once money enters the conversation, it's almost never about the price. It's about the framing. It's about whether they believe the outcome you're promising will actually happen. A weak "yes, maybe, I'll think about it" at €10 is far more informative than enthusiastic nodding at €0. The money forces honesty.
Code won't fix weak positioning. If the problem framing doesn't create enough pressure to open a wallet, no amount of engineering elegance will compensate. Only sharper clarity about the outcome — and the consequences of not getting it — fixes that.
7. Are You Solving Market Pain or Personal Curiosity?
I need to be direct about something, because I suspect many builders reading this are in the same position I was.
I love systems. I love designing engines that measure progress with precision. I find CEFR-mapped learning architectures genuinely fascinating — the way you can decompose language competency into discrete, testable skills and build feedback loops around each of them. I could spend months building the most elegant scoring system in the world and call it work.
But markets don't pay for intellectual elegance.
They pay for pain relief.
It's easy to build something that proves you're smart. Every technical decision can be a small act of self-expression — choosing the more sophisticated architecture, the more interesting AI approach, the more nuanced evaluation framework. It feels like craft. In some ways it is craft. But craft without a paying customer is a hobby.
The distinction between solving market pain and satisfying personal curiosity now sits behind every feature decision I make. When I'm considering adding something new, I ask: is there a specific user, with a specific deadline, who is experiencing a specific frustration that this feature resolves? If I can't answer that with a real person in mind — not a persona, a real person — I don't build it.
8. The Checklist I Now Use Before Building Anything
Before adding features or starting something new, I force myself to answer each of these in writing. Not in my head — in writing, because writing exposes the places where my thinking is vague.
Is this tied to real pressure? Not a vague aspiration. A specific consequence that arrives on a specific date if the user fails to act.
What happens if someone ignores it for a year? If the honest answer is "not much, really," the problem framing needs work before anything else happens.
Are people already paying to solve this? If not, what evidence do I have that they would? And how long would it take to educate the market to the point where they would?
Is this a company-level move or just a feature? Does it require depth that a larger platform would have to rebuild from first principles, or could it be absorbed into an existing product in a sprint?
Who, specifically, is this for? Not a demographic segment. A type of person, with a specific deadline, a specific consequence, and a specific thing they're afraid of.
Can I validate it in 30 days without heavy engineering? If the answer is no, the question becomes: why not? What assumption am I making that I haven't yet tested?
If I can't answer those cleanly, I don't build. Not because I've lost faith in the idea — but because I've learned that ambiguity at the validation stage becomes expensive confusion at the execution stage.
The Hard Truth Nobody Tells You
Product development feels like momentum. Every commit, every design iteration, every new feature gives you the sensation of moving forward. It's viscerally satisfying in a way that thinking hard about market positioning simply isn't.
Clarity feels slow. It feels uncomfortable. It sometimes feels like an excuse not to build. It isn't.
Clarity saves years.
The language learning app market is growing rapidly — the global online language learning market was estimated at over $22 billion in 2024 and is projected to reach nearly $55 billion by 2030. There is no shortage of opportunity. But there is also no shortage of well-funded, well-engineered products that failed because they were built for a real problem that wasn't urgent enough, for an audience that was too broad, with outcomes that weren't concrete enough to justify the cost.
Preplings only started to feel real when I stopped asking "What impressive system can I build?" and started asking "What is painful enough that someone will pay to fix it right now?"
Those are different questions. The first one leads to elegant technology that not enough people need. The second one leads to something that might actually matter.
Build last. Validate first.
Earn the right to the interesting engineering by proving the outcome is worth building toward.
That shift changed everything.
Deepak Upreti