Here’s the most expensive sentence in business: “Let’s just build it and see.”
We’ve watched it play out many times. An owner has an idea — an app, a booking system, a new digital service — skips straight to building, spends five figures, launches… and hears crickets. The idea wasn’t stupid. It was just never tested. The market got asked for its opinion only after all the money was spent, which is the one moment its opinion can’t save you.
There’s a better order of operations, and it’s the single most valuable habit in design thinking: prototype first. Test the idea cheaply and quickly, then decide whether to build. Done right, you get 80% of the answer for roughly a tenth of the cost. Here’s how it actually works.
What a prototype really is (hint: not a mini product)
Most owners hear “prototype” and picture a rough version of the finished product — smaller budget, same idea. That’s the misunderstanding that makes prototyping sound like slow-motion building.
A prototype isn’t a small product. It’s a question, wearing a costume. The question is “would customers actually use this?” — and the prototype is whatever the cheapest thing is that gets real customers to answer honestly. Sometimes that’s a clickable mockup. Sometimes it’s a landing page. Sometimes it’s literally a person doing manually what software would eventually do. The product comes later, if — and only if — the answer is yes.
Four ways to test an idea without building it
1. The clickable mockup. Screens of your app or service, linked together so a customer can tap through them — no code behind any of it. Put it in ten real customers’ hands and watch: where they tap, where they stall, what they ignore. Days of work, and it exposes most fatal flaws — the confusing flow, the feature nobody touches, the step where everyone gives up.
2. The fake door. Add the button for the thing that doesn’t exist yet — “Book online,” “Try our new service” — to your real website, and count who clicks. (Clickers get a polite “coming soon — leave your email.”) It sounds cheeky; it’s actually the purest demand test there is, because it measures what people do, not what they politely say. If nobody clicks the door, be grateful: you just saved the cost of the room behind it.
3. The landing-page test. One page that sells the idea as if it were real — the promise, the price, a sign-up button — with a small ad budget pointed at it for a week or two. Cost: hundreds. What you learn: whether strangers, not friends, will part with an email address or a deposit for this. That’s demand, measured in behaviour.
4. The concierge test. Deliver the service manually to a handful of customers before automating anything. Thinking of an AI that recommends products? Have a human make the recommendations by hand for ten customers first. You learn exactly what customers value about the service — which is precisely the specification for the software you’d then build. (This one is criminally underused, and it’s the best possible first step for any AI idea: it tells you what the AI actually needs to do before you pay anyone to make it do it.)
The maths that makes this obvious
Put rough numbers on it. A proper build — an app, a custom platform, an AI system — commonly runs S$30,000 to S$100,000+ before you learn anything real about demand. The tests above run from a few hundred dollars to a few thousand, and take days or weeks, not months.
Now the part people miss: the prototype pays off whichever way the answer goes. If it’s no — you just bought the world’s cheapest dodged bullet. If it’s yes — you build with evidence: you know which features matter, what customers will pay, where they got confused. The final build comes in cheaper and better-aimed because of the test. There is no outcome in which testing first was the wrong call — except the one where the delay costs you a genuine first-mover race, and honestly, for most SME ideas, that race exists mainly in the owner’s head.
Why smart owners still skip it
Worth naming, because the trap is emotional, not logical. Building feels like progress — there’s a contract, a timeline, something to show the team. Testing feels like doubt. And by the time an owner is ready to act on an idea, they’ve usually fallen in love with it, and nobody commissions a test they’re afraid might say no.
But that’s exactly backwards. The stronger your conviction, the cheaper it is to confirm — and the more expensive it is to be wrong. The owners who test first aren’t the timid ones. They’re the ones who’ve been burned once.
And here’s the part almost nobody knows: Singapore will help pay for the test
This prototype-first approach isn’t just cheap — it’s the exact behaviour the government now co-funds. DesignSingapore Council’s Good Design Research grant (currently being upgraded into the larger GDRD programme) supports precisely this: researching your customers, prototyping an idea, and validating a proof of concept before the big build. If your idea starts with “we don’t know whether our customers would…”, the finding-out can be co-funded.
Test first, and let the government share the bill for testing. It’s hard to construct a reason not to.
Where Oasis Web Asia comes in
This is genuinely how we prefer to start. Before we build anything significant for a client, we’d rather run the cheap test — the mockup, the fake door, the concierge week — and put evidence on the table, because “understand first, then build” only means something if the understanding comes before the invoice. We design the right test for your idea, run it with your real customers, and hand you an honest answer: build it, change it, or thank the prototype and walk away. And when the answer is build, the research has already told us exactly what to build — often with grant co-funding along the way.
If you’ve got an idea you believe in but haven’t tested, that’s exactly the conversation we like to have.
Start a conversation → — get a free consultation with our Singapore-based team.