How do I do market research for a business idea or product?
Settle three things with evidence: who exactly has the problem, how they solve it today and what they pay, and how many of them there are. Start with desk research. Count your customers with government data on data.gov.in, read industry reports such as Blume's Indus Valley report for how Indian consumers split by income, check Google Trends for whether interest is rising, and study competitors' sites, pricing and reviews on G2. An AI research agent such as Gemini Deep Research drafts a cited report in minutes; check every number against its source. Then do primary research: talk to people who have the problem about what they did the last time it happened, not whether they like your idea, as The Mom Test teaches. Finish with a bottom-up count of customers times what each would pay a year. Rule: decide from what people have done and paid, never from what they say they would buy.
Migicovsky built Pebble on hundreds of these conversations, and this lecture is the specific playbook: how to find your first ten users, what to ask them, and how to tell a polite no from real interest. Pair it with the Alströmer talk above and you have most of what YC teaches on this in under an hour and a half.
The US government's plain guide to market research: demand, market size, economic indicators, location, saturation and pricing, plus how to analyse competitors.
Blume's annual report is the single best data-grounded read on what is actually happening in Indian internet and consumption, not what is trending among San Francisco VCs on X. It shows you adoption curves (UPI, quick commerce, ONDC, how few households actually shop online) so you can pressure-test whether a hyped trend has the demand base to land here. Treat it as a starting map of Indian reality, then judge your specific trend against it.
From
Blume Venturesby Sajith Pai, Anurag Pagaria and team (Blume Ventures)~180 charts
Grounds trend-spotting in real Indian adoption data (income distribution, online-shopping penetration, UPI-native monetization) instead of imported hype.
Indian startups are building a distinct playbook (micro-subscriptions on UPI Autopay, DPI rails) that has no clean US analogue, so copying a US trend directly often misses the real opportunity.
A trend can be huge in raw numbers yet thin in monetizable demand: the report repeatedly separates users from paying users.
The single best thing ever written on customer conversations. It teaches you to ask about the customer's life and past behaviour, not your idea, so you can't be lied to. If a founder reads one thing before talking to a single customer, it's this.
Most market sizing advice online is generic TAM SAM SOM filler. This one is written by an investor, backed by a survey of 30 VCs, and it is honest about the thing that matters: a big number pulled from an industry report proves nothing. It walks you through building the number bottom up (customers times what they pay you per year), which forces you to confront whether real people will actually pay, and that is the honest test of whether an idea can grow past a niche.
Size the market bottom up (count of real customers times annual revenue per customer), not by claiming a percent of some giant top down figure.
TAM, SAM, and SOM are used loosely across the industry, so state your assumptions plainly instead of hiding behind the acronyms.
Project the market out five or more years and include how you would actually reach and acquire customers, since a market you cannot serve is not your market.
India's official open data portal: datasets from ministries and states on households, businesses, agriculture, health, education and more, useful for counting a market.
When you cannot read the code or the spec sheet, real buyers can. G2 lets you pull up any technical competitor, sort to the one and two star reviews, and read exactly where the product frustrates people in plain language. You do not need to understand the technology to understand "onboarding took six weeks" or "the API kept breaking", and that is the signal you are actually hunting for.
Go straight to the low-star reviews of each competitor: the complaints are your market's unmet needs, written by people who paid for the product.
G2 breaks ratings into fields like ease of setup and quality of support, so you can spot a repeatable weakness without decoding a single technical feature.
Cross-check the same product on Capterra and Trustpilot, since each site skews to a different buyer, and treat any one glowing or scathing review as noise until a pattern shows up.
You do not need to be in anyone's network to see interest rising, you need the search data everyone already generates. Google Trends lets you compare terms over time and break interest down by Indian state and city, so you can catch a curve bending up before it becomes obvious. A free, honest first read on whether a trend is real or just loud in your own circle.
