Every so often I get the same question from someone who wants to launch a product or open a new line of business: "how do I know if there's demand for this?". And once that person gets a quote for formal market research from a consulting firm and sees the price, the question changes to "can I do it myself?". The short answer is yes, at least a fairly decent first version. The long answer is this article: what market research actually is, what types exist, and how to do a market study step by step with free sources, including one I work with every day and that rarely shows up in the manuals: search data.
What is market research?
Market research is the process of gathering and analyzing information about a market before making a business decision. In concrete terms, it tries to answer four things: how many people want what you plan to sell, who those people are, how much they currently pay to solve that problem, and who is solving it for them today.
The analogy I use most: it's the commercial equivalent of the soil study you run before constructing a building. You can skip it and build anyway, and sometimes it works out. But when it doesn't, you find out with the building already standing, which is the most expensive possible moment to find out. The same happens with businesses: the information you didn't gather before launching gets paid for later, in inventory that doesn't move or a store located where your customer never walks by.
One vocabulary point that confuses quite a few people: "market study" and "market research" are used almost as synonyms, and for practical purposes you can treat them that way. If you want the fine distinction, market research is the full discipline, the toolbox. A market study is a specific project that uses those tools to answer one concrete question: whether it makes sense to open in a second city, whether there's room for another specialty coffee brand, whether a small business would pay for this service.
What types of market research exist?
There are several classifications floating around, but the two that actually matter for deciding how to work are these.
By the origin of the data:
- Primary sources: information you collect yourself, directly from people. Interviews, surveys, focus groups, product tests. It's the information best fitted to your question, and also the most expensive and slowest to get.
- Secondary sources: information that already exists because someone else collected it. Public statistics, industry association reports, bank studies, search data. It's cheaper (often free) and faster, in exchange for not being tailored to you.
By the type of data:
- Quantitative: answers how much and how many. Market size, purchase frequency, average ticket, search volumes.
- Qualitative: answers why and how. What annoys people about current solutions, what words they use to describe their problem, what would make them switch.
The practical rule I give anyone starting out: secondary sources before primary ones, because they're cheaper and save you from asking questions that are already answered. And quantitative to size things, qualitative to understand them. A useful study almost always mixes both.
What free sources can you use?
Before spending a single dollar on surveys, all of this is already available:
- National statistics institutes: in Peru that's INEI; censuses, household surveys, demographics by region and district. If your question includes "how many people of this profile live in this area", it has probably been measured already.
- Government open data: open data portals publish records from several sectors that help size markets.
- Industry associations and chambers of commerce: they publish reports with figures for their industries.
- Google Trends: free, shows whether interest in a topic is rising or falling over time and in which regions it concentrates.
- Keyword data: the Google Ads planner (free with an account) or paid SEO tools tell you how many times a month something gets searched, with which exact words, and how much advertisers pay for that click.
- The search results themselves: googling as if you were your own customer and noting who shows up, what they offer and at what price is a mini competitive study that costs nothing.
How do you do a market study step by step?
This is the full recipe I use, in six steps.
Step 1: start from the decision you want to make
A market study exists so you can make a decision, so start by writing it down in one sentence: "I will open a second store in this city only if I find evidence of sufficient demand". If you can't write that sentence, the problem isn't information yet: first clarify what you're deciding. This step also protects you from the most common mistake, which is collecting endless data "just in case" that nobody ever uses.
Step 2: size the market with secondary sources
With the decision clear, look for what's already been measured: how many people match your target profile (statistics institutes), how the sector has been growing (industry reports), what the business press says. The goal is to land on an order of magnitude; the exact number matters little at this stage. Knowing whether your potential market is 5,000 people or 500,000 changes the entire decision, and that difference can be detected with free data.
Step 3: measure real interest with search data
This is where my favorite source comes in, and I dedicate a full section to it below. The short version: check how many times a month your category gets searched in your country, with which words, and whether the trend is going up or down. An illustrative example so the logic is clear: if "electric bike rental" gets 90 monthly searches in your whole country and has been flat for two years, that number tells you something about the current size of spontaneous demand that no survey of 30 acquaintances ever will.
Step 4: map your competition
Google the 10 or 15 most important searches in your category and write down who shows up: who ranks organically, who pays for ads, what prices they display, what their customers complain about in reviews. Your competitors' negative reviews are qualitative gold: they are customers of the exact market you want to serve, telling you for free what remains badly solved. To systematize this part, the full process is in my benchmarking guide.
Step 5: talk to real people
Only here do primary sources come in, once you know what to ask. Between 5 and 10 thirty-minute interviews with people who match the profile you defined yield more than a poorly designed mass survey. Ask about the last time they had the problem and what they did at that moment. Avoid asking whether they "would buy" your product: people are kind and say yes to almost anything hypothetical, whereas what they did last time is a verifiable fact. To define who to interview, it helps to have your buyer persona and your market segmentation clear.
Step 6: synthesize and decide
Go back to the sentence from step 1 and answer it: did I find the evidence I asked for or not? Organize the findings on one page (an analysis matrix helps keep it from becoming a list of loose data points), write down what you assumed without being able to verify, and make the decision. If the process doesn't end in a decision, what you produced was a report, and reports without decisions tend to sit in a folder forever.
Why does search data deserve a central place?
This is what I get to see from the SEO side, and I think it's still underused as a research tool. Every Google search is a person spontaneously declaring they have a problem or an intention, without anyone asking them and without the social pressure of looking good in front of a pollster. Aggregated, those searches give you three things that cost a traditional study real money: the size of expressed demand month by month, the exact words your market uses to describe its problem (which are almost always different from the ones your industry uses), and seasonality, meaning in which months that demand rises and falls.
And there's a new layer that didn't exist a couple of years ago: answers from AI-powered search. Asking these systems which brands they recommend in your category shows you the competitive map exactly as your future customers will see it, because a growing share of them asks an AI for that comparison directly instead of opening ten tabs. How to turn all of this into business decisions with process and criteria is something I develop in data-driven marketing.
What mistakes should you avoid?
The ones I see repeated most often:
- Researching to confirm. If you've already decided to launch and the study is a formality to validate yourself, you'll read the data through that lens. Defining in step 1 what evidence would make you NOT launch is the antidote.
- Surveying your own circle. Your friends and family love you, and they're probably not your market. Their answers will give you encouragement, which is appreciated, but very little information about real demand.
- Confusing interest with willingness to pay. The fact that something gets searched a lot or sparks curiosity doesn't mean anyone will pay for it. That's why step 4 looks at competitors' real prices: if nobody is charging to solve that problem, ask yourself why.
- Studying once and filing it away. Markets move: demand changes, competitors enter, and now the way people search and buy is changing too. A market study has an expiration date, and the free sources in this guide let you refresh it every six months with no budget.
- Analysis paralysis. The perfect study doesn't exist, and in practice the winner is whoever moves when the information is sufficient, even if incomplete. If you can already answer the sentence from step 1 with reasonable evidence, decide.
The takeaway
You don't need a budget to stop deciding blind: about three hours are enough for the first pass. This week: open Google Trends and the Google Ads planner and note how much your category gets searched and where the trend is heading, google the 10 main searches in your field and record who shows up and at what price, read the negative reviews of your three most visible competitors, and schedule two conversations with people who have had the problem you want to solve. With that you have the first version of your market study. It won't replace a consulting firm when the decision involves millions, but for most decisions a small or mid-sized business makes, it's more than enough to decide with data on the table.