Last updated: 1 September 2026
Entry research answers two questions that get merged into one budget line and should not be. Sizing asks how many people could plausibly buy, which is an arithmetic problem built on public data. Validation asks whether anyone in that market actually wants what you are selling, at your price, against what they currently use.
Only people can answer that. A business case with a confident number and no respondents has answered the first and skipped the second.
The failure this produces is specific and common. A team assembles a market size from published reports, each of which cites an earlier report, none of which discloses its own method, and arrives at a total addressable market that has never touched a customer. The number is precise, defensible-looking, and has no error bar because nobody knows what the error is.
Sizing is the cheaper half and the half that misleads. Validation is the expensive half and the one that changes decisions, and what that validation costs is set out in what consumer research costs in 2026.
Where Desk Sizing Numbers Come From, and Where They Break
Most published market sizes are derived rather than measured, and the derivation is usually invisible. A figure quoted in a press release traces to a syndicated report, which was modeled on a national statistic plus assumptions the report does not publish. Compounding three such figures produces a number whose uncertainty nobody can state.
Build from primary public data instead, where the method is documented and the assumptions are yours:
- National statistics offices and international bodies. World Bank open data for income and population structure, UN statistics for demographic and economic series, Eurostat for European detail.
- Connectivity and device data where the product depends on either. The ITU's statistics remain the standard reference.
- Trade and regulatory context, including the US government's country commercial guides, which are written for exporters and state entry conditions rather than opportunity sizes.
Three disciplines make a sizing defensible. State every assumption as a separate line so each can be challenged independently. Build a range rather than a point, because a single number implies a precision the inputs do not support. And identify which single assumption the answer is most sensitive to, since that assumption is what the primary research should be aimed at.
What Sizing Cannot Tell You
A market size describes how many people exist, not how many want your product. The gap between those is where entry failures live, and four specific things sit inside it.
- Whether the need exists in this market at all. Categories do not transfer. A problem that is acute in one market can be solved in another by something you have not heard of.
- What the real alternative is. The incumbent is frequently not a competing product but a habit, an informal provider, or doing nothing.
- Whether your price reads as expensive. Price is judged against local reference points, not against your home market's.
- Whether the channel works. Distribution, payment and trust mechanics vary more between markets than the product does.
None of those four are visible in any public dataset. All four are answerable by asking people, and all four have sunk entries where the sizing looked excellent.
How Do You Validate Demand With Real Respondents?
The sequence matters, because the cheapest studies come first and each one can stop the process.
| Stage | Question | Typical scale | What a bad result means |
|---|---|---|---|
| Category understanding | How is this need met today, and by whom? | Small, qualitative, saturation-led | The need is already solved; stop or reposition |
| Proposition reaction | Does this read as useful here? | Qualitative per segment | The proposition needs local reframing before anything else |
| Alternative mapping | What would they actually give up? | Qualitative, then quantified | The real competitor is not the one in your deck |
| Price context | What is this judged against locally? | Quantitative, trade-off based | Your price architecture does not survive the reference set |
| Channel and trust | How would they find and pay for this? | Qualitative, operational | The route to market is the constraint, not demand |
Best place to stop early: category understanding. It is the cheapest stage and it kills more bad entries than any other, which is precisely why teams under launch pressure skip it. A market where the need is already met adequately is not a small opportunity, it is not an opportunity.
For sample sizing on the qualitative stages, saturation logic applies. A published method for assessing thematic saturation found saturation at six interviews in two datasets and eight to nine in a third, using a threshold of no more than five percent new information, for relatively homogeneous groups with narrow aims. The critical caveat for entry work is that saturation inside one market says nothing about themes that only emerge when markets are compared, and cross-market work needs materially more than the sum of its parts.
That same analysis reports 70 percent of 114 themes surfacing in the first six interviews and 92 percent within twelve, which is why per-market samples stay small while the cross-market layer does not. Where that sample comes from in the first place is covered in survey panels and where respondents come from.
How Do You Reach Consumers in a Market You Do Not Operate In Yet?
This is the operational problem that decides whether entry research happens at all, and it is why so many entry decisions get made on desk research. Which platforms actually recruit outside established panels is covered in emerging-market research platforms.
You have no customer list, no local brand recognition to make recruitment easy, and frequently no panel coverage outside the largest cities. Panel supply thins sharply beyond metros in most markets, so the accessible sample is systematically the most urban, most affluent and most internationally-exposed part of a country. That group is also the least representative of a mass-market opportunity, and it is the group most likely to say yes to a foreign proposition.
The connectivity gradient makes this concrete. The ITU's connectivity series shows how unevenly access is distributed across and within countries, and even in high-income markets the pattern holds: Pew Research Center's mobile technology fact sheet reports 16 percent of US adults as smartphone-only internet users. Pew puts that at 34 percent among households under $30,000 a year.
The same fact sheet puts overall smartphone ownership at 91 percent, so the constraint is broadband and privacy rather than a handset. A browser-based study in an unfamiliar market compounds that skew with an unfamiliar-brand recruitment penalty.
Which Channels Reach Past the Metros?
Channel is what moves it. Alchemic runs interviews natively inside WhatsApp with no link and no app to install, and by outbound phone call for respondents reachable by voice rather than browser. Alchemic publishes 57+ languages including Spanish, Hindi, Tamil, Bangla, Arabic and Indonesian.
Fieldwork has run across the USA, the UK, India, the UAE, the Philippines, Nigeria, Indonesia, Uganda, Kenya, Sri Lanka, Malaysia, Thailand, Bangladesh and Pakistan, reaching from metros and Tier 1 through Tier 2 and Tier 3. Recruitment runs as managed fieldwork or bring your own, with incentives paid over bank transfer, global rails or local instruments where those are what respondents use.
Its new market entry work is built around that constraint: the sample frame is the hard part of entry research, not the discussion guide.
Whatever the channel, the professional expectations for consent, disclosure and respondent treatment travel with the study, and are set out in the ESOMAR code and guidelines and the AAPOR standards and ethics materials. The same selection effect is examined in sample validity and who you miss.
What Does a Go or No-Go Decision Actually Need?
Five things, and a sizing number is only one of them.
- A range, not a point, with the sensitivity assumption named.
- Primary evidence that the need exists locally, with respondents you could describe.
- The real alternative identified, including informal and do-nothing options.
- A price architecture tested against local reference points, not converted from home-market pricing.
- An explicit statement of what would falsify the case. If no realistic finding would stop the entry, the research is decoration and should be defunded rather than commissioned.
Where Entry Research Misleads
Four limits.
Early adopters in a new market are not the market. The people willing to talk to an unknown foreign brand are unrepresentative in exactly the direction that flatters your case.
Stated intent overstates behavior everywhere, and travels unevenly. Courtesy norms differ between markets, so the inflation is not a constant you can subtract.
Competitive response is invisible. Incumbents with local distribution can respond faster than an entrant can react, and no respondent can forecast that. Ask instead what an incumbent could do cheaply, and whether the case survives it.
Regulatory and operational reality can void a validated demand case. Demand you have proven is worth nothing if the route to market is closed. The trade and regulatory reading belongs early rather than as a compliance check at the end.
A fifth limit is about who is doing the interpreting. Entry research is commissioned by people who have usually already formed a view, and it is read in a room where someone's plan depends on the answer. That is not a reason to distrust it, but it does change what the design has to carry: the falsification criterion needs writing down before fieldwork, in the brief, with a named decision-maker attached. It will not survive being agreed afterwards.
Teams that skip this end up litigating the method whenever the finding is unwelcome, and the method is always attackable, since every sample has a limit and every qualitative base is small. Deciding in advance what evidence would be sufficient removes that argument before it has anywhere to go.
One practical note on sequencing. The regulatory and channel reading is cheap and fast, so it belongs before the fieldwork rather than after, even though it feels like a later-stage concern. Discovering that the route to market is closed is the one finding that makes every other finding irrelevant, and it can usually be established in days from public sources.
Teams whose next question is which method to run rather than whether to enter will get more from the concept testing overview and the qualitative research primer.