Skip to content

Market Analysis & Customer Research

Foundations of Business Development - NIT Northern Institute of Technology / TUHH, Hamburg · part of my Technology Management MBA · study notes for revision.


Once you have an idea and a rough sense of the value it creates, two blunt questions decide whether it is a business or a hobby: is the market big enough, and do you actually understand the people you want to serve? This chapter is about answering both with evidence instead of hope - first by sizing the opportunity, then by segmenting the customers, and finally by going out and researching them directly.

When start-ups die, the single biggest category of reasons is not a weak team, tough competition, or even running out of money on its own - it is market and customer-related: building something that too few people actually wanted, or wanted enough to pay for. The money runs out because the demand was never really there.

There’s a useful phrase for the thing you’re chasing: product-market fit - the moment your product clearly solves an urgent problem for a well-defined group of customers who will pay for it. Founders describe failure again and again in those exact terms: “we failed for the same reason 90% of tech start-ups fail - we didn’t find product-market fit before the cash ran out.”

The failure usually runs as a chain, not a single event:

Not an urgent problemor the solution doesn’t really fix it
→
Too few paying customersweak or stalling revenue
→
Investors lose faithpressure rises, next round refused
→
Cash runs outteam conflict, shut-down
A simplified model of entrepreneurial failure. Notice where it starts: not with money, but with a problem that wasn’t urgent and a solution that wasn’t fitting. Market research attacks the root, not the symptom.

The famous E10 fuel launch in Germany (a petrol with ten percent bio-ethanol) is a textbook case. The product was technically fine - independent scientists found roughly a two percent rise in consumption and no damage to almost any modern car - yet drivers boycotted it. Why? Nobody explained the risks and benefits, gas-station staff couldn’t answer questions, and the perceived risk was never reduced. Politics and industry assumed “green, eco and bio” would sell it on their own. It didn’t. A good product with no understanding of the customer’s fears is still a failed launch.

Before you commit, you want a number for the prize. The standard way to express market potential is three nested figures, each smaller and more realistic than the last.

TAM · Total Addressable Market
total demand for the whole category - everyone who could ever buy this type of thing
SAM · Serviceable Available Market
the slice your specific product actually fits - the product-market-fit portion
SOM · Serviceable Obtainable Market
what you can realistically win in the first years, given your resources and reach
Market potential as three shrinking circles. TAM is the dream, SAM is the fit, SOM is the plan. Each is usually quoted in euros per year (or in number of customers/units).
LevelWhat it measuresHow you narrow to it
TAM - Total Addressable MarketTotal demand for the wider category or type of product/serviceStart here: the whole pie
SAM - Serviceable Available MarketThe demand your particular offering can best satisfy (the product-market-fit portion)Filter TAM by what your product is actually suited to
SOM - Serviceable Obtainable MarketThe share you can realistically capture given your resources and targetsFilter SAM by your sales reach, budget, channels and segment focus

A worked example - private EV charging stations. Sizing the German market for 11 kW home wall-boxes ran roughly like this:

  1. TAM - take the forecast number of battery-electric cars in use by 2025 (about 3.3 million), assume a ratio of chargers to cars, multiply by an average station price (about 2,500 €). Result: a multi-billion-euro category.

  2. SAM - only some of those chargers are the 11 kW wall-box type your product serves. Apply that share (about sixty percent) to the TAM. The number shrinks to the part your product actually fits.

  3. SOM - you can only reach four major cities plus suburbs (about a quarter of the population) and only customers willing to buy online (about thirty percent of buyers). Multiply those filters onto the SAM. You land at a realistic first-years figure - tens of millions, not billions.

There are two directions you can travel when you estimate these numbers, and they tend to disagree in a telling way.

Top-down
  • Start from the big category number (from market studies, statistics) and cut it down with assumptions: TAM → SAM → SOM
  • Fast, uses published macro data
  • Danger: “we only need one percent of a huge market” - easy to say, rarely true
  • Best when the industry is established and well-measured
Bottom-up
  • Start from one thing you can actually count - visits a salesperson can make, one region, one customer group - and build upward
  • Slower, but grounded in real capacity
  • Ties the forecast to what you can do, not to what the market theoretically holds
  • Usually more credible for a start-up
Two routes to the same figure. A synchrotron-detector start-up estimated top-down that it could win about 52 units in year one; building bottom-up from “one salesperson can do 180 visits, close 30 managers ordering 1.5 detectors each” gave about 45. When they disagree, trust the one anchored in real activity.

A third handy trick is analogy: when your market is too new to measure, borrow data from an established, similar market. A jet-surfboard maker sized its market by starting from worldwide JetSki sales and a comparable firm’s market share. And in practice most real estimates are a mix - a bit top-down, a bit bottom-up, a bit analogy - cross-checked against each other. The point is never a single perfect number; it is a defensible range built from stated assumptions you can later test and correct.

A market is not one uniform blob of buyers. It is a mix of people and organisations with different needs, budgets and buying habits. Segmentation means slicing that population into groups so that inside each group people are as alike as possible, and between groups they are as different as possible. Get that right and every later decision - product, price, message, channel - has a clear addressee.

The classic split is between describing who customers are and what they want. The first is easy to measure but crude; the second is more useful but harder to observe.

Identifier variables (a priori - who they are)Response variables (post-hoc - what they want)
B2C (consumers)Demographics (age, gender), socio-economics (income, job, education), psychographics (values, lifestyle, attitudes)Benefits sought (price, reliability, service), usage situation and frequency, price sensitivity, loyalty and switching, openness to new things
B2B (businesses)Company size, industry, location, financial situation, level of internationalisationSame behavioural and needs-based cuts, applied to the buying organisation

You cannot serve everyone at once, and trying to is a classic start-up trap. The move is to pick one narrow but big-enough, underserved, profitable segment that fits your firm and product - a beachhead you enter first, prove yourself in, then expand from.

The project-software firm HQ Labs learned this the hard way. Version 1.0 targeted “SMEs in any industry” with generic features - and drowned: it had to learn the language of every industry, fight low software-affinity, and run long cold-sales cycles. Version 2.0 narrowed to agencies alone, with features built for exactly them (shared projects, freelancer integration, agency benchmarking). Suddenly the value was obvious, acceptance was high, and marketing could run inbound instead of cold-calling. Narrow won.

Once the segment is validated, you make it vivid with a persona: a rich, semi-fictional portrait of a typical customer in that segment, based on real research. Segmentation proves the sub-group exists; the persona keeps a busy founding team focused on it and makes internal and pitch communication concrete.

Persona - Prof. van Dusen, the “busy academic”
  • 47, married, two kids; a lifelong learner, short on time
  • ”I care about the environment - as long as it doesn’t hurt my wallet.”
  • Trusts scientific proof and numbers; not very tech-savvy; dislikes social sharing
  • Wants precise, visual energy-saving reports he can trust, with zero effort

Design implication: sell “saving green” as much as “being green”; show progress visually; respect his time.

A persona for a smart-home energy product. It reads like a person, not a spreadsheet row - that’s the point. But it must be built from interviews and data, never invented to flatter your idea.

A crucial subtlety, especially in B2B and complex markets: the person who uses your product, the person who pays for it, and the person who decides are often not the same. In healthcare, a new drug has patients (users), physicians (deciders/prescribers), hospitals and pharmacies (buyers/channel), and insurers and regulators (who control reimbursement and approval). Sell to the wrong one and nothing moves. So map the value chain and ask, for your opportunity: who is the user, who is the buyer, and who is the decision-maker?

Now the real work: replacing assumptions with evidence. Two foundational distinctions organise all of it.

Secondary vs primary. Secondary (desk) research reuses data that already exists - market reports, official statistics, competitor sites, trade journals, forums. It’s cheap and fast, so you almost always start here. Primary (field) research collects fresh, first-hand data for your specific question - because not everything you need is sitting in a web page or a tweet.

Secondary (desk) research
  • Existing data: market studies, government and industry statistics, competitor websites, journals, forums, Google Trends
  • Cheap, quick, broad - do it first to frame the market
  • Watch-out: don’t behave like a keyword-typing algorithm; be creative, follow different paths
Primary (field) research
  • New, first-hand data collected for your specific question
  • Interviews, observation, surveys, experiments
  • The only way to reach needs, fears and reasons that were never written down anywhere
Start with what already exists, then go into the field for what doesn’t. Good sources for secondary work: internal records, trade associations, government departments, libraries, exhibitions, online databases, experts.

Within primary research, the deeper split is between depth and measurement.

Qualitative - for depth & discovery
  • Interviews, observation/ethnography, focus groups
  • Answers why: reasons, feelings, needs, the buying and usage process
  • Best early, when you don’t yet know the right questions
  • Output: customer profile of pains, gains and jobs-to-be-done, backed by quotes
Quantitative - for measuring & validating
  • Surveys, experiments, analysis of secondary data
  • Answers how many: frequencies, averages, correlations
  • Best later, to size and confirm a hypothesis on a representative sample
  • Output: market size, segment shares, measured behaviour
Qualitative first to discover what matters; quantitative after to measure whether it holds at scale. Do them in that order - you can’t survey people about a need you haven’t discovered yet.

Here’s how the main methods compare in practice:

MethodGood forWatch-outs
Secondary dataMarket size, structure, trends; cheap first passAges fast; not tailored to your question; may not exist for a new market
Interviews (qualitative)Reasons, feelings, complex buying/usage processes; discovering unknown needsNeeds access and interviewer skill; small, non-representative sample; time-consuming
Observation / ethnographyWatching what people actually do, not what they saySlow; you see behaviour but not always the “why” behind it
Surveys (quantitative)Measuring known, not-too-complex issues on a representative sampleOnly as good as the questions; low response quality; self-reported; can’t discover the unknown
ExperimentsTesting cause and effect; validating willingness to paySetup effort; artificial settings can mislead
Match the method to the question. Rough sequence for a time-pressed team: available data → talk to experts → observe customers online → survey → interview real target customers.

Interviews are where beginners waste the most opportunity, because it feels like selling and the instinct is to talk. Flip every instinct:

Don’tDo instead
Talk most of the timeListen most of the time
Pitch and defend your ideaFocus on their pains and gains; be ready to hear what kills your idea
Ask “what would you do if…?” or “what do you plan to change?”Ask about actual past behaviour: what they do, what they’ve already tried
Ask leading questions (“Wouldn’t you agree that…?”)Ask neutral, open-ended questions; ask “why”
Run through 20 fixed questions identically each timeBuild on about 5 core questions and adapt as you go
Hear what you want to hear; interpret on the spotTake notes on what they actually said

Two traps deserve their own warning. The first is confirmation bias - going in wanting the idea to be right, so you hear agreement everywhere and quietly discount objections. The cure is to try to invalidate your idea, not confirm it. The second is the gap between what people say and what they do: stated intentions (“yes, I’d definitely buy that”) are notoriously unreliable, which is exactly why you ask about real past behaviour and, where you can, observe instead of just asking.

All the tools above share one attitude, and it is the real lesson of the chapter. Market and customer research is straining and awkward - which is precisely why founders skip it and stay at their desks polishing the product. Don’t. Market research is the systematic, objective collection and interpretation of data to reduce the risk in your decisions, and most of that data lives with real people, not in your assumptions.

AssumptionTalk to real customersObserve what they doSize it honestlyEvidence
The whole chapter in one line: turn assumptions into evidence before you scale. Get out of the building.

So: start desk-side with secondary data to frame the market and rough out your TAM/SAM/SOM; pick a narrow beachhead segment and sketch a persona as a hypothesis; then go out, interview and observe real target customers to test it - and be genuinely willing to change your idea when the evidence says so. That loop, run early and cheaply, is what separates the ventures that find product-market fit from the ninety percent that don’t.

Next: Entrepreneurial Strategy: The Four Choices → - the hard choices every young venture must make.