Costs
Part of Is money already moving before market validation?
Which market validation tools are actually data sources you can trust?
Market validation tools are data sources: what each category honestly supports, six questions before a figure enters the memo, and the estimator trap.
Most market validation tools that matter are data sources, not software. The list includes the US Bureau of Labor Statistics, the US Census Bureau, Companies House, SEC EDGAR, SAM.gov, Statista, Google Trends, and the audience estimators in Meta Ads Manager and LinkedIn Campaign Manager.
The skill is reading them, not choosing them. A source built for a different purpose will still give you a confident answer to your question.
What follows names the sources, says what each can honestly support, and gives the questions that decide whether a figure belongs in your memo.
What to take away
- Judge a source by who produced it, who paid for it, what was counted, and when. A number without those four is decoration.
- The definition is where most errors enter. Your category and the source's category are rarely the same shape.
- The cheapest instrument on this page is a conversation with somebody who already sells to your buyers.
The categories and what they can carry
| Source | Honestly supports | Cannot support |
|---|---|---|
| National statistics agencies: BLS, Census Bureau, Eurostat, UK Office for National Statistics (free, method disclosed) | Population counts, industry employment, business counts, over time | Your specific segment, or anything about willingness to pay |
| Regulators and public registers: Companies House, SEC EDGAR, USPTO, EUIPO | Who exists, what they are licensed for, formal filings | What they buy, or what it costs them |
| Industry associations: National Association of Manufacturers, American Bankers Association, Specialty Food Association | Category definitions, member counts, how the trade describes itself | Neutral sizing; the membership is the sample |
| Published market reports: Statista, IBISWorld, Gartner, Forrester | That a category has a name and someone sized it | Your reachable slice, unless the method is disclosed |
| Advertising audience estimators: Meta Ads Manager, LinkedIn Campaign Manager, Google Ads | Roughly how many accounts match a targeting description | A count of people, or that they are buyers |
| Search demand tools: Google Trends, Google Keyword Planner, Ahrefs, Semrush | That people type certain words, and roughly how often | What they meant, or that they will pay |
| Job listings: LinkedIn Jobs, Indeed, Glassdoor | That somebody funded work on this problem | How many, or for how long |
| Directories and membership lists: Crunchbase, LinkedIn Sales Navigator, Apollo.io | A findable list of the right people | That the list is complete or current |
| Public procurement portals: SAM.gov, USAspending.gov, TED | How organizational buyers describe what they buy | Anything about smaller or private buyers |
| Survey panels: Prolific, Amazon Mechanical Turk, Qualtrics panels | Answers from people who take surveys | Behavior, or a sample resembling your market |
| Patent and trademark databases: USPTO Patent Full-Text, EUIPO eSearch | Who is filing, in what technology class, and when | Whether anyone bought the product |
| Community and review sites: Reddit, G2, Capterra, Trustpilot | The words buyers use about problems and alternatives | A representative sample, or sales figures |
| Crowdfunding platforms: Kickstarter, Indiegogo, Republic | Whether a pitch and a price attract backers | A market total, or repeat purchase behavior |
| Company registries: OpenCorporates, Companies House | Legal existence, officers, filing history | Revenue, headcount, or buying intent |
The first row is the one people never actually open. The federal series on business births, deaths and survival is free, dated, and states its method, which puts it ahead of most figures that end up in a plan.
Access terms matter. BLS, Census, Companies House, SEC EDGAR, SAM.gov, USPTO and EUIPO searches are free. Google Trends is free, and Google Keyword Planner is free with a Google Ads account.
Ahrefs and Semrush plans typically start near $130 a month as of 2026. Statista runs from a few hundred dollars a month to five figures a year. LinkedIn Sales Navigator is around $100 a month, and Crunchbase and Apollo.io each have free tiers.
What Each Source Can Carry
Source
- National statistics
- Population, employment, business counts
- Regulators
- Who exists, licenses, filings
- Industry associations
- Category definitions, member counts
- Published market reports
- Category has a name
- Audience estimators
- Roughly how many accounts match
- Search demand tools
- People type certain words
Honestly supports
- National statistics
- Your segment, willingness to pay
- Regulators
- What they buy or pay
- Industry associations
- Neutral sizing; membership is sample
- Published market reports
- Your reachable slice
- Audience estimators
- A count of people or buyers
- Search demand tools
- What they meant or will pay
Cannot support
- National statistics
- Regulators
- Industry associations
- Published market reports
- Audience estimators
- Search demand tools
Read the right-hand column first. Almost every bad market memo is built by taking a source past it.
Six questions before a number goes in the memo
Who produced it. A body that counts things for its own purposes is a different instrument from one that publishes to attract clients.
Six Questions Before the Memo
- Who produced it
- Who paid
- What was counted
- Over what period, and when
- How the sample was drawn
- Whether the definition matches yours
Who paid. Sponsored research is not worthless, and the sponsor's interest tells you which direction to discount.
What was counted. Establishments or companies. People or accounts. Transactions or customers. Different sources use the same word for these, and the difference is often larger than the number.
Over what period, and when. A figure with no date is unusable. Anything that moves with the economy needs both the period covered and the publication date.
How the sample was drawn. An undisclosed method means the figure cannot be checked, which means it belongs in your memo as a claim rather than a fact.
Whether the definition matches yours. This is the one that catches people. If the source's category is broader than your segment, you cannot narrow it by multiplying by a fraction you chose.
Write the answers beside the figure. A number with a source and a date is evidence. The same number without them is something you will end up defending without remembering why.
Free instruments people skip
- Somebody who already sells to your buyers.A distributor, an agency, a consultant. They know the buying cycle, the objections and the approval path, and they will usually tell you. LinkedIn Sales Navigator, free for a month, can find the right few.
- The refusal.Take an offer to a small number of people and record what they say when they decline. This produces information no dataset contains.
- Trade press archives.How the category described itself a few years ago against today tells you what changed and what buyers now expect.
- Public filings of larger players.Where a listed company operates in your category, its own reporting describes the market in its own terms, with dates attached, and full text search across those filings will find the passage rather than the company.
- A count you make yourself.Building a list of a hundred qualifying organizations by hand is tedious and produces a number you can defend, which no purchased figure will. OpenCorporates and Companies House make this cheaper.
All five are the same move: taking evidence from behavior rather than from a published total, which is where problem discovery starts as well.
The estimator trap
Audience estimators inside advertising systems are the most seductive instrument on the list, because they return a large, precise, instantly available number.
That number counts accounts matching a targeting description. It is not a count of people, and certainly not of people with the problem or the authority to buy. It also reflects the platform's own inference about those accounts, which you cannot inspect.
Use it for one thing: comparing the relative size of two targeting descriptions on the same platform. That comparison is meaningful. The absolute number is not, and it should never be the top line of a bottom-up estimate. What that estimate should be built from is set out in market validation.
Building a source list you can hand over
Keep one sheet with six columns: the claim, the source, the address, the date you accessed it, what was counted, and one sentence saying what this does not establish.
The last column makes it useful later. When the plan changes you will want to know which figures still apply, and the only way to answer quickly is to have written down what each one was actually about. It is the same discipline that makes a competitive read honest rather than reassuring: see competitor validation.
Common questions
Do I need to buy a market report?
Rarely, and read the method before deciding. A report with an undisclosed method gives you a number you cannot defend, which is worse than no number, because you will use it anyway.
A Statista subscription typically runs from a few hundred dollars a month for single-user access to five figures a year for enterprise seats. A single IBISWorld report often costs four figures.
What if no data exists for my segment?
Common for anything narrow, and not a blocker. Build the count yourself from directories and lists, and be explicit that it is your own count with a stated method. A defensible small number beats an indefensible large one. Apollo.io and Crunchbase both have free tiers that export a starting list.
Which of these tells me whether people will pay?
None. Willingness to pay comes from an offer with a price attached and a person who can refuse it. Everything here draws the boundary, and the conversations that fill it in remain the only route to what happens inside.
How current does a source need to be?
It depends what moves. Population and business counts move slowly; prices, channel costs and platform behavior move fast. Put the publication date next to every figure so a reader can see which kind they are looking at, and refresh it before the gate that depends on it.
