Fazley RabbiTesting a business idea cheaply means finding the one assumption that would sink the idea if it were wrong, and checking it with the smallest experiment that asks real people for real commitment: a pre-order, an advance payment, a booking, a signed letter of intent. You do this before renting a shop, buying machines or building software. Praise from friends and likes on a post are not evidence. People paying, or clearly trying to, is.
Write the idea as a list of assumptions
Every idea rests on several beliefs. For a hypothetical frozen-snack business in Bogura, they might be:
- Working families in the town want ready-to-fry snacks.
- They will order through Facebook or WhatsApp and pay on delivery or by bKash.
- We can make the snacks at steady quality in a home kitchen at first.
- We can deliver within town before the food thaws.
- A price that covers our costs is one people will pay.
Write yours down, then ask of each: if this is wrong, does the whole idea fail? The assumption that is both critical and least proven is your riskiest one. Test that first. Many founders test the comfortable things, like the logo or packaging, and leave the dangerous question until after the money is spent.
A ladder of cheap tests
Start at the bottom and climb only when a step gives a clear signal.
| Test | What it tells you | Effort |
|---|---|---|
| Conversations with possible buyers | Whether the problem is real and how they solve it now | Low |
| A post or group question describing the offer | Whether people are curious enough to ask | Low |
| A simple page with a WhatsApp button | Whether strangers act, not just friends | Low to medium |
| A small ad test pointing to that page | Whether you can reach buyers at a cost that might work | Medium |
| Pre-orders or a small advance by bKash or Nagad | Whether people will commit money | Medium |
| A manual, behind-the-scenes version | Whether you can deliver and whether customers return | Medium |
| A small batch, pop-up stall or short pilot | Real operations, real feedback, real repeat orders | Higher |
Conversations come first because they cost almost nothing; my guide to customer interviews explains how to ask without leading people. For the page step, a focused landing page tells you far more than sending people to a general profile.
The manual version deserves a word. Before building an app that matches tutors with students, you could match them yourself with a spreadsheet and phone calls. Before designing an IoT device, you might offer a monitoring service using ready-made sensors and check readings yourself. If customers do not value the manual service, automation will not save it.
Decide what "pass" means before you start
Write the success line before you run the test, not after. For example: "If, within two weeks, enough strangers place prepaid orders to cover one production batch, we continue." Keep criteria about behaviour: orders, deposits, bookings, repeat purchases, referrals. If you decide afterwards, it is too easy to call any result a success.
Set a time limit as well. A test that runs forever is not a test; it is a slow business that never decides.
Read the results honestly
Some signals are much stronger than others.
- Weak: likes, compliments, "I would definitely buy this", orders from friends.
- Medium: messages asking about price and delivery, people sharing their contact details, cash-on-delivery orders. These are useful, but cancellations and refused parcels happen, so confirm COD orders before you count them.
- Strong: advance payments, repeat orders from strangers, customers recommending you unprompted, a business signing a purchase order.
When results are mixed, look for the reason. Weak interest may point to the wrong audience, message or price rather than a bad idea. Change one thing and test again. When results are clearly poor after a fair test, stopping is a good outcome: you learned cheaply what others learn expensively.
Treat early customers fairly
If you take money before the product exists, say so plainly: it is a pre-order, this is when you expect to deliver, and you will refund in full if you do not go ahead. Then keep that promise and record every payment. If the idea involves food, medicine, cosmetics or anything else that may need a licence, ask the relevant authority or a lawyer what applies before you sell, even in a trial.
Mistakes that waste money
- Building the full product, website or app before anyone has paid.
- Testing only with friends and family, who are kind rather than honest.
- Changing several things at once, so you cannot tell which one mattered.
- Ignoring operations: the idea sells, but you cannot make or deliver it reliably.
A checklist before you invest heavily
- I have written the idea as assumptions and named the riskiest one.
- I have spoken to real potential buyers, not only people I know.
- At least one test asked people to commit money or time.
- I set the pass line and time limit before the test began.
- I have delivered to some customers and seen whether they come back.
- I know my rough cost per order and whether the price covers it.
If most boxes are ticked, you are investing on evidence. If not, run one more small test. Once the idea passes, the business development service can help you plan the launch.