I get this question often from people who have not started yet: I have a product idea, how do I know there is demand before I spend anything. The instinct behind the question is right, because the most expensive mistake in a small business is buying stock or running ads for something nobody was asking for.
The advice they usually get is either run a survey or ask people around you. Both give you a comfortable answer and a misleading one, because you are asking people about future intent and people answer kindly. Someone telling you it is a lovely idea is not lying to you. They are just not buying.
What works looks different. Instead of asking, you look for demand that is already moving without you, and then you run a small test where real money or real time changes hands.
Demand is not the same thing as interest
These two words need separating before anything else makes sense. Interest is someone hearing your idea and saying it sounds good. Demand is someone who was already looking for the thing before they had ever heard of you.
The difference is practical rather than philosophical. Interest does not repeat on its own, and demand repeats every month with no effort from you. That is what you can build on.
So the first question is not whether people will like your product. It is whether people are looking for this right now. Those are two different questions with two different methods.
The cheapest test: is anyone searching for it
This part costs nothing and takes an hour, and there are two things you need to know about the tools before you trust their numbers.
First, Google Trends does not show you search counts. Google own documentation says each data point is divided by the total searches for its geography and time range, then scaled to 0 to 100. The number is relative, not absolute. The same page says something more important: terms with low volume appear as 0, and statistical noise is added for privacy, which is most noticeable on queries with low or no search interest.
The practical consequence of that sentence is this. If your term comes back as zero, that is not evidence there is no demand. It is evidence the term is smaller than the tool will report. Plenty of workable ideas get abandoned over that zero.
The same documentation also states that Trends is not a scientific poll, and that it should always be considered as one data point among others before drawing conclusions. That is Google own wording, and I like repeating it because it means the caution is not just my opinion.
Second, Keyword Planner gives you average monthly searches, and its help page explains that the average covers 12 months by default and that search volume is rounded, which is why totals across several locations may not add up the way you expect. It also carries a line that matters for anyone testing a small market: small geographic targets yield less accurate results.
So if you are testing demand in one city, the weakest case for the tool is precisely your case. Read it as a direction, not as a number.
Three tests that cost a little money or only time
Once you have a direction, you need something stronger: someone paying a price, and the price does not have to be money.
1. The pre-order test
One page that describes the product honestly, and one button. If you can take a small refundable deposit, that is the strongest test available. If you cannot, make the button reserve a unit against a name and a number. The figure that matters is not how many people saw the page, it is what share of them signed up.
2. The manual selling test
Before any stock, sell five units by hand. Calls, WhatsApp, a market stall, anything. If you cannot sell five while speaking to people directly, advertising will not fix that problem. It will scale it and make it more expensive.
3. The competitor test
If someone is selling the same thing and has been trading for two years, that is the strongest demand signal you will find, and most people read it backwards and get discouraged. A competitor means the market exists. No competitor at all means either you are early or there is no demand, and the second is far more common. How to think about the cheaper competitor specifically is in competing with a cheaper competitor.
If you test with ads, know the floor first
Many people jump straight to an ad test, and that can be the right call, as long as you know that an ad below the minimum viable budget is not a test. It is buying numbers that mean nothing. Working out that floor is arithmetic rather than guesswork, and my colleague Mostafa Faried wrote how to calculate the minimum sensible starting ad budget, in Arabic, which is the step that follows directly from this post.
The rule I work by: if the minimum test budget is larger than the amount you could lose and carry on, then ads are not your testing tool today. Go back to selling by hand.
What number tells you to continue
This is where people want a magic number, and I will not give one because there is not one. There is a way to set your own number before you start, and that is what actually makes the difference.
Write this down before the test: if I reach this many sign-ups or this many sales in this many days, I continue, and if it is less, I stop and rethink. The number comes from your own arithmetic: unit cost, margin, and the smallest volume that makes the work worth doing. Deciding it before the test is what stops you from interpreting the result in favour of your idea afterwards, which is what happens every time.
For the wider frame on setting those numbers in the first place, there is how to identify your target market from the numbers and how to know marketing is working.
The limits of all this
First limit: everything above answers whether demand exists. It does not answer whether you are the right person to serve it. A product with heavy demand, fierce competition and a thin margin is real demand and a bad opportunity for you, and no test here separates those two.
Second limit: the figures and settings above were accurate for the pages I read while writing this, and Google changes its documentation. If you are making a large decision, open both pages yourself.
Third limit, and this is opinion rather than fact: these tests reduce risk, they do not remove it. I have watched ideas pass every test and fail in the market, and the reverse. What you are doing is trading a possible large loss for a certain small loss of your own time, which I think is the best deal available to someone who has not started yet.
Most businesses buy a solution before they know what they actually have. That is why my book Growth Engine starts with diagnosis rather than tactics, and ends that first step with one positioning sentence that has to pass a test: could your competitor say the same thing? It is all in Part 1, Analysis: where are you, free, and written in Arabic.