How do I know if my startup idea is good?
By Bonaventure Ogeto, Founder and CEO, Mctaba Labs
Published
A startup idea is good when a specific group of people has a problem often, already spends money or hours working around it, can be reached by you cheaply, and can pay a price that covers your costs. You cannot judge this by how the idea feels. Score it against evidence, then test the weakest point first.
You have carried the idea around for months. You have explained it at a wedding in Nyeri, in the office kitchen and to your cousin who "knows tech". Everyone nodded. One person said you should hurry before somebody else does it. And still you do not know whether it is any good.
That doubt is healthy. An idea is not good because it is clever, new, or praised. It is good when it matches how a real group of people already behaves. The useful news is that you can check most of that from your phone, in a few weeks, before you pay anyone to build anything.
Why can't I trust how I feel about the idea?
Because the feeling measures the wrong thing. Excitement tells you the idea is interesting to you. It tells you nothing about whether a shop owner in Thika will open her M-Pesa and pay for it every month.
Three things distort your judgment in the same direction:
- You know the solution too well. You have imagined the finished product so many times that it feels real. The customer has never seen it and is not waiting for it.
- People are polite. When you describe your idea, the easy response is praise. It costs nothing and ends the conversation pleasantly.
- You count agreement as demand. "That would be useful" and "I will pay KES 800 a month for that" are different sentences. Only the second one matters.
So set the feeling aside and look for signs you can observe.
What are the signs of a good startup idea?
There are seven. A strong idea does not need a perfect mark on all of them, but it needs most.
- You can name the customer. Not "SMEs" or "the youth". Something like "owners of private primary schools in Nakuru with 100 to 400 pupils". If you can name the group, you can find it.
- The problem comes up often. A problem that appears every day or every week keeps a product in use. A problem that appears once a year is forgotten between occurrences.
- The problem costs something you can count. Lost sales, hours of a paid employee, penalties, spoiled stock. If the cost cannot be put in shillings or hours, the price conversation will be hard.
- People already work around it. They use an exercise book, a spreadsheet, a clerk, a relative, or a tool they complain about. A workaround proves the pain is real enough to act on.
- You can reach the buyers cheaply. You are in their WhatsApp groups, you sell to them already, or they gather somewhere you can go. An idea whose customers you cannot reach is a good idea for someone else.
- The buyer can pay enough. The price they can afford, times the number you can realistically win, covers your costs with room left over.
- The first version is small. You can describe one job the product does on day one. If the idea only works once it has ten features and thousands of users, it is a hard idea to begin with.
How do I score my own idea?
Use the table below. Give each row 0, 1 or 2. The rule that makes it honest: you may only score from things you have seen or been told by customers. An assumption scores 0, however confident you are.
| Sign | Score 0 | Score 1 | Score 2 |
|---|---|---|---|
| Named customer | "Everyone" or a broad sector | A group, but you cannot list any | You can list ten by name |
| Frequency | Yearly or rarer | Monthly | Weekly or daily |
| Countable cost | Annoying only | Costs time nobody pays for | Costs money or paid hours |
| Workaround | None seen | A free, rough workaround | They pay for a workaround today |
| Reach | No route to buyers | You can reach them with paid adverts | You already have a direct line to them |
| Ability to pay | Price would not cover your costs | Covers costs with a large volume | Covers costs with a small number of customers |
| Small first version | Needs many features or a large network | Needs three or four features | One job, deliverable by hand at first |
Add it up out of 14. These cut-offs are a rule of thumb, not a science:
- 11 to 14. Strong. Move to asking for money.
- 7 to 10. Promising with a hole in it. Fix the lowest row before anything else.
- Below 7. Keep the customers, change the problem. Ask the same people what costs them the most each month.
A worked example. Say your idea is a fee reminder and tracking tool for private primary schools in Nakuru. You know four school owners, so the named customer scores 1. Fees are chased every term and every week within it, so frequency scores 2. A bursar spends days each term on it, so the countable cost scores 2. Schools use an exercise book and bank slips, so the workaround scores 1. Your aunt runs one such school and sits in an owners' WhatsApp group, so reach scores 2. You have not discussed price with anyone, so ability to pay scores 0. A first version that sends reminders and records M-Pesa payments is one job, so that scores 2.
Total: 10 out of 14. The idea is promising, and the scorecard has told you exactly what to do next. Talk to owners about price, and grow the list of four schools to ten.
What does a weak idea look like from the inside?
It rarely feels weak. It feels like an idea that just needs more time. Watch for these patterns.
- "Everyone needs this." Products for everyone are sold to no one, because you cannot find, message or design for everyone.
- You have to explain why it is a problem. If people do not raise the problem themselves, they do not feel it.
- The user and the payer are different, and you have only spoken to one. Students may love it while the school refuses to pay. Parents may pay while children refuse to use it.
- It depends on someone else saying yes first. A bank, a county office or a telecom has to agree before you can serve a single customer. That is a partnership project, not yet a product.
- It requires people to change a habit they like. Asking traders to stop using cash, or customers to stop ordering on WhatsApp, is a far bigger job than the software.
- You keep adding features to make it sound better. Good ideas get simpler as you learn more. Weak ones get longer.
One of these is a caution. Three together mean the idea needs to change shape.
Does a good idea have to be new?
No, and chasing newness leads founders astray. If nobody has ever tried to solve the problem, the likeliest reason is that it does not hurt enough, not that you are the first to notice.
Existing products are evidence that customers pay. What you need is a gap: a group the current options do not serve well. In Kenya that gap is often practical. The existing tool bills by card in dollars when customers pay by M-Pesa. It assumes a laptop when the owner runs the business from a phone. It was designed for a company with an accounts department when your customer is one person with a notebook.
"The same thing, built properly for this customer" is a perfectly good idea.
Is this idea good for me in particular?
An idea can be sound and still be wrong for you. Three questions decide it.
Do you have a way in? A nurse building for clinics, or a former sacco clerk building for saccos, begins with trust and phone numbers. An outsider begins with cold calls.
Do you understand the work? If you have done the job your customer does, you will spot which part of the problem is the expensive part. If you have not, plan to spend longer listening.
Can you afford the wait? Selling to schools or companies can take a term or a quarter per decision. If your savings cover three months, choose an idea where customers decide in days.
What should I do with the score?
Go after the lowest row, because that is where the idea is most likely to fail.
If reach is the weak row, spend two weeks getting into the groups and places where your customers are before you think about software. If ability to pay is weak, have price conversations. If the first version is too large, cut the idea down until one person could deliver the result by hand with WhatsApp and a spreadsheet.
This matters for the build as well. A high score on a small first version means a short, cheap first release that does one job. A clear named customer means the screens, the language and the payment method can be chosen for that customer instead of guessed. Founders who skip the scoring tend to ask for everything, because they have no basis for leaving anything out. Founders who do it can say what the product is, who it is for, and what it must do first. That is what a good idea looks like when it is ready to be built.
Key takeaways
- Excitement, yours or your friends', is not evidence. Behaviour is.
- A good idea has a named customer, a frequent problem, an existing workaround and a way for you to reach buyers.
- Score the idea out of 14 using only things you have seen or heard, not things you assume.
- Competitors are a good sign. An empty market usually means an absent problem.
- The same idea can be good for one founder and bad for another, depending on access to the customers.
- Work on the lowest scoring row before you spend anything on a build.
Frequently asked questions
- Is my idea bad if someone else is already doing it?
- No. Someone else doing it shows that customers exist and that money changes hands. Your question is whether a group of customers is badly served by what exists, for example because it is priced in dollars, needs a laptop, or ignores how people here pay. If you can name that group, the idea is still open.
- Should I ask friends and family what they think of my idea?
- Ask them for introductions, not opinions. People who care about you will encourage you, and people who do not have the problem cannot judge it. The only opinions worth collecting come from people who live with the problem, and even then you should ask what they do today, not what they think of your plan.
- Can an idea be too small to be worth building?
- Yes, if the people who have the problem are few and each can only pay a little. Multiply a realistic number of customers by a realistic monthly price. If the result would not pay one salary after costs, it may still be a fine side business, but it will not support a company.
- How long should I spend deciding whether an idea is good?
- Give the scorecard an evening and the conversations two to three weeks. If you are still unsure after speaking to a dozen people who have the problem, the uncertainty is itself an answer. Strong ideas tend to produce clear signals quickly.
Sources
- The Mom Test: How to Talk to Customers and Learn If Your Business Is a Good Idea When Everyone Is Lying to You, Rob Fitzpatrick (book, 2013). Checked on 19 November 2026.
- The Lean Startup: How Today's Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses, Eric Ries (book, 2011). Checked on 19 November 2026.
