hypothesis testing
Hypothesis testing is checking whether your business guess is actually true before you spend real money finding out the hard way.
In plain English
Think of the woman who sells beans and rice at the bus park in Ojota. She has a hunch that customers would buy more if she added a small side of fried plantain to every plate. She could just start frying plantain for everyone tomorrow and hope for the best, or she could test the hunch first on twenty customers and watch what actually happens.
That is hypothesis testing in plain clothes. You take a guess about your business, something like 'customers will pay ₦500 more for faster delivery' or 'this new packaging will sell better than the old one', and instead of betting the whole business on that guess, you set up a small, honest test to see if reality agrees with you.
A founder in Lagos building a fintech app might guess that adding a savings feature will keep users longer. Rather than build the full feature over three months, she puts up a simple waiting list page and a mock screen, then checks how many people actually tap to sign up. The numbers either support her guess or calmly correct her.
In the UK, a small bakery in Leeds might hypothesise that Sunday opening will bring in ₦200,000 worth of extra sales a week, in local currency terms. They try it for four Sundays, track the till, and let the receipts settle the argument instead of joy or gut feeling.
For someone in Toronto sending money home to fund a cousin's new skincare line in Abuja, this matters too. A small test batch sold to fifty people tells you more truth than a big confident speech ever will, and it costs far less than finding out after the whole shipment has been paid for.
The point is never to be right immediately. The point is to be willing to be wrong cheaply, early, and on purpose, so the business learns faster than the market can punish it.
A shop owner in Enugu guesses that offering a ₦1,000 discount on orders above ₦10,000 will increase weekly sales. She runs the offer for two weeks only, tracks that sales rose from ₦300,000 to ₦420,000, and confirms the guess before making the discount permanent.
Questions people ask
Do I need fancy statistics software to test a business hypothesis?
No. Most small business tests just need a notebook, a spreadsheet, or even a simple tally on paper. What matters is comparing what actually happened against what you expected, not the sophistication of the tool.
How small can a test be and still mean something?
Smaller than you think. Testing a new product on twenty real customers before a full launch already tells you far more than guessing, as long as those twenty are genuinely representative of who you plan to sell to later.
What if the test proves my idea wrong?
That is the test working exactly as it should. Finding out an idea will not work with ₦50,000 and two weeks of your time is a far better outcome than finding out with ₦5,000,000 and six months.
Last reviewed