How to Know If There Is Demand for Your Business Idea
Short answer: demand is proven by triangulation, never by a single signal. Take five independent measurements — search and query demand, transaction evidence, competitor health, direct outreach response, and a small paid test — and see whether they agree.
Any one of them can mislead you. Four of them pointing the same way rarely does. And the only measurement that fully settles the question is the one where somebody hands you money.
Finding a problem worth solving is one thing. This guide asks a colder question: are there enough people with that problem, reachable by you, at a price that works?
Those are three separate conditions. Most failed businesses satisfied one or two of them.
The five signals, and why you need all of them
Each signal answers a different question and fails in a different way. Used alone, each produces its own characteristic false positive.
| Signal | What it answers | Typical false positive |
| Search and query demand | Are people actively looking? | High volume, zero purchase intent |
| Transaction evidence | Is money already moving? | Money moves, but only to incumbents at scale |
| Competitor health | Can this support a business? | Competitors look busy, are quietly failing |
| Direct outreach | Will these specific people engage? | Politeness inflates the response |
| Paid test | Will people actually pay? | Sample too small or badly targeted |
1. Search and query demand
Look at what people type into search engines — and increasingly, what they ask AI assistants, which tends to be longer, more conversational and far more revealing about the real decision being made.
What to measure:
- Volume for the core commercial term
- Volume for “near me” and location-modified variants, if you’re a local business
- Volume for problem-phrased queries, not only solution-phrased ones
- Trend direction across 24 to 36 months, not three
Read it properly. Low volume isn’t automatically bad news. A B2B service with 90 monthly searches at £8,000 a contract is a much better business than a consumer product with 40,000 searches at £14. Multiply volume by realistic value before you react to either number.
The trap: informational search volume is not commercial demand. “How to fix a dripping tap” is enormous and worth almost nothing to you. “Emergency plumber” plus a town name is small and worth a great deal. Sort your queries by intent before you add them up.
2. Transaction evidence
Find the places where transactions are visible from the outside:
- Marketplace listings with review counts — reviews are a rough proxy for volume, and a product with 400 reviews has sold vastly more than 400 units
- Sold listings on auction and resale sites
- Booking systems showing availability; a competitor with nothing free for three weeks is a demand signal you can read for free
- Published price lists and rate cards
- Tender and procurement portals for B2B and public sector work
- Job adverts, again — hiring is spending
This is the cheapest high-quality signal available, and the one most people skip entirely.
3. Competitor health
The existence of competitors proves demand exists. Competitor health proves the demand is profitable, which is a separate and more important claim.
Read for: how long they’ve traded, whether they’re hiring, whether they’re opening additional locations, whether review volumes are still growing, whether they discount constantly, and whether their listings look maintained or abandoned. Where published accounts are available, ten minutes with them will tell you more than a week of speculation.
| What you see | What it usually means |
| Several healthy competitors, clearly differentiated | Good market. Demand is real and there’s room to position |
| One dominant player and nobody else | Either a hard market to enter or a genuine opening — find out which |
| Many competitors, all discounting heavily | Commoditised. Demand exists, margin doesn’t |
| No competitors at all | Almost always a warning rather than an opportunity |
“No competition” is the most misread finding in business research. Genuinely unserved markets are rare. Unservable ones are common. Before celebrating, find out who tried this and stopped.
4. Direct outreach
Contact fifty people who match your target customer. Not friends, and not through a survey.
Ask for something that costs them a small amount of effort: a fifteen-minute call, a waitlist signup with a real email address, a reply describing how they handle this today. Then measure the response rate, not the warmth of the responses.
Benchmarks worth holding yourself to:
- Cold B2B outreach: 5 to 15% reply rate is normal. Below 3% means the wrong list or the wrong message.
- Warm network outreach: below 25% is a serious signal that the problem isn’t really felt.
- Waitlist conversion from a landing page: 5 to 20%, depending heavily on traffic quality.
The discipline that matters: count only replies containing specific detail about the person’s own situation. “Sounds interesting, keep me posted” is a polite no wearing a disguise, and counting it as interest is how founders talk themselves into eighteen months of work.
5. The small paid test
The only signal that settles the argument. Ask for money, or for something adjacent to money: a deposit, a pre-order, a paid pilot, a booked and confirmed slot.
Five payments beat five hundred survey responses. The mechanics — thresholds, sample sizes, what to build and what not to — are covered in How to Test a Business Idea Before Spending Money.
Sizing the market from the bottom up
Top-down sizing — “the UK pet care market is worth several billion, and if we capture just 0.1%” — is worthless. It produces a number you can’t act on and can’t check.
Bottom-up sizing is arithmetic you can defend. Here’s a worked example for a dog daycare in a mid-sized town.
| Step | Calculation | Result |
| Households within a 20-minute drive | Census and local data | 34,000 |
| Households with a dog (~28%) | 34,000 × 0.28 | 9,520 |
Working households needing weekday care (~18%) | 9,520 × 0.18 | 1,714 |
| Willing to use paid daycare regularly (~25%) | 1,714 × 0.25 | 428 |
| Existing capacity across four competitors | 4 × 30 places | 120 |
| Unserved or switchable demand | 428 − 120 | ~308 |
| Realistic year-one capture (10%) | 308 × 0.10 | ~31 regulars |
| Revenue at £28/day, 2.5 days a week, 48 weeks | 31 × 28 × 2.5 × 48 | ~£104,000 |
Now you have something you can argue with. Every assumption is visible and testable. If someone thinks 25% is too generous, you change one cell and watch what happens to the answer.
Always run the pessimistic version too.
Halve your capture rate and drop your price by 15%. If the business still works, the idea is robust. If it collapses, you’ve just identified the single assumption you’re betting the business on — and that’s the one to test first.
Growing, stable, or declining?
Three questions decide this, and only one of them looks backwards.
Is the underlying driver expanding? Demand rarely moves for its own reasons. It follows demographics, regulation, technology, housing, employment patterns or cost pressures.
Identify the driver behind your demand and check its direction. Pet daycare follows pet ownership and office attendance. Compliance services follow legislation. Home improvement follows housing transactions and interest rates.
Is the category being substituted or absorbed? Declining markets often look stable right up until they don’t, because incumbents mask the decline by consolidating. Ask what technology or behaviour change could remove the need altogether, and how far along that change already is.
Is spend per customer rising or falling? A market with a flat customer count but rising spend is growing. A market with more customers each spending less is commoditising, which is usually the worse of the two.
Worth saying plainly: a stable market is perfectly investable. Founders systematically over-value growth and under-value stability. Stable markets come with known economics, patient competitors and buyers who aren’t constantly re-evaluating their options.
Seasonality: measure the trough, not the peak
Almost every business has some seasonality. It only becomes dangerous when the trough can’t cover fixed costs.
Do this calculation before you commit to anything:
- Estimate monthly revenue across a full year, honestly.
- Identify your worst three consecutive months.
- Check whether those three months of revenue cover three months of fixed costs.
- If they don’t, work out the cash reserve needed to survive them — and treat that reserve as a mandatory part of your startup capital rather than an optional buffer.
Then go looking for counter-seasonal work. Gardening firms take on Christmas lighting. Wedding suppliers take on corporate events. Tutors run exam-season intensives. The best counter-seasonal revenue uses the same equipment and the same people you already have.
Watch the cash pattern as well as the revenue pattern. Retailers spend heavily on stock in October and get paid in December. That two-month gap has closed businesses that were, on paper, profitable.
The demand verdict
Score each signal from 0 to 4, based only on evidence you have actually gathered.
| Signal | Weight |
| Search and query demand at commercial intent | 15% |
| Transaction evidence | 20% |
| Competitor health and headroom | 20% |
| Direct outreach response rate | 20% |
| Paid test result | 25% |
Above 70%. Demand is established. Move on to customer segmentation and economics.
50 to 69%. Demand is plausible but unproven. Find your weakest signal and go and strengthen it before you spend anything.
Below 50%. Don’t launch on this. Either the demand isn’t there, or you haven’t yet found the segment that has it — which is what How to Identify Your First Profitable Customer Segment is for.
One override rule sits above the arithmetic: if the paid test scores zero, the total doesn’t matter. Four strong signals and nobody willing to pay means you’ve found an interesting phenomenon, not a business.
Frequently asked questions
How much search volume is enough?
There’s no universal threshold, because the answer depends entirely on what a customer is worth. Multiply monthly commercial-intent searches by a realistic conversion rate and by your average customer value, and see whether the result covers your costs. Ninety searches a month at £8,000 a contract is a strong business. Forty thousand searches at £14 with a 1% conversion rate is not. Volume without value is just traffic.
What if my idea is genuinely new and nobody is searching for it?
Then search volume is the wrong instrument, and you lean on the other four signals — particularly direct outreach and the paid test. Look for search demand around the problem rather than your solution, since people search for what they know to call it. But be careful here: “nobody is searching because it’s new” is also what founders say when the truthful answer is that nobody wants it. The paid test is what tells those two situations apart.
Can I trust surveys at all?
For measuring demand, barely. Surveys are reliable for descriptive facts about the present — what people currently use, what they currently spend, how often something happens. They’re close to useless for predicting future behaviour, because saying yes to a hypothetical costs nothing. If you run one, ask only about the past and the present, and never about intention.
How long should demand validation take?
Two to four weeks for the first four signals, and another two to six for a meaningful paid test. Longer than that usually means you’re gathering comfort rather than evidence. If you find yourself on week nine still researching, ask what finding would actually change your decision — and if nothing would, you’ve already decided and you’re just delaying the risk.
What if the demand exists but it’s all going to one big competitor?
That’s a positioning question rather than a demand question, and it’s often a good position to be in — the market is proven and someone else paid to educate it. Look for the segments the incumbent serves badly: customers too small to interest them, a geography they cover thinly, a use case their product handles awkwardly. Concentrated markets are hard to attack head-on and frequently soft at the edges.
My outreach response rate was low. Is the idea dead?
Not necessarily, but you need to work out which of three things went wrong: the wrong list, the wrong message, or genuinely absent demand. Test the message first, since it’s cheapest to change — rewrite it around the problem in the customer’s own language and try fifty more. If a second, better-targeted attempt also comes back below 3%, treat that as real evidence rather than bad luck.
Should I validate demand before or after building something?
Before, in almost every case, and the paid test doesn’t require a finished product. Deposits, pre-orders, booked slots and paid pilots all measure real willingness to pay against nothing more than a clear description of what someone will receive. Building first inverts the risk: you spend the money, then discover the answer, and by then the money is gone whichever way it goes.
Does high demand mean the business will be profitable?
No, and conflating the two is a common and expensive mistake. Demand tells you people want the thing. Profitability depends on what it costs to deliver and what it costs to acquire each customer, both of which are independent of how much people want it.
Plenty of high-demand categories are unprofitable for new entrants precisely because demand attracted so many competitors.
Run the numbers in Can This Business Actually Make Money? before you treat proven demand as a green light.
Comments (0)