The direct answer
A new-country market map is not a country-filtered export. It is a documented estimate of which operating accounts could fit, why each is in scope, and where the estimate is uncertain. Separate the market-sizing model from the reviewed account queue: the first explains the denominator; the second contains entities a team has actually reviewed. If the account rule itself is still vague, define it first with the ICP-to-account-list workflow.
Start with a country, activity, operating footprint, scale range, and exclusion policy. Decide whether the commercial unit is a legal entity, an operating site, a group, or a local buying centre. Then use classifications and registers to find candidates, but require current company evidence before calling a row targetable.
1. Define the unit before counting
One group can create five different rows: a parent, national subsidiary, local sales entity, factory, and branch. Your market-map number changes materially depending on which one is counted.
Write the rule before opening a database. Otherwise the map becomes a list of companies that happened to be easy to find.
2. Use classifications as a starting hypothesis
For European economic statistics, NACE is the common activity classification; Eurostat notes that NACE Rev. 2.1 applies to European statistics from 2025 onward. A code can help define a candidate universe, but it is not proof of a company’s current commercial relevance. Eurostat: NACE
Use a two-layer model:
- Coverage layer: sector codes, registers, association lists, exhibitor lists, and known accounts generate candidates and a defensible estimate.
- Qualification layer: current company pages, local operating evidence, and the written ICP decide whether an entity can enter the reviewed account queue.
The EU business-register system can help users locate register information across participating countries, but availability and fields differ by jurisdiction. Treat register data as an evidence source with a date and local limitation, not a complete current market feed. EU e-Justice: business registers
3. Build a coverage ledger
Record the discovery path for every candidate. This prevents the loudest source from being mistaken for the best source.
Calculate the map in stages: candidates found, duplicates merged, entities with enough evidence, accounts accepted, and accounts requiring review. Do not publish the first number as “TAM” if it merely counts search results.
4. Localise the evidence, not only the country field
Country expansion fails when a global category is assumed to mean the same thing locally. Add a local-language vocabulary list, legal-form variants, regional site patterns, and exclusions that would otherwise slip through. Preserve original local terms rather than replacing them with an English summary only.
Run a small negative-control review: companies from a neighbouring country, national distributors, entities registered locally but operating elsewhere, and companies using the same category terms in another context. The aim is not a perfect market count; it is to reveal whether the account rule survives local reality.
5. Prove the mechanism on 20 accounts, then expand to 50
Use the first 20 deliberately mixed accounts to prove that another reviewer can apply the entity boundary, evidence rule, and negative controls. If that mechanism holds, expand the same contract to a 50-account coverage pilot with a fixed source mix and review window. Do not change the account unit between those stages. For each account, retain:
- canonical entity and domain;
- country/site evidence and date;
- activity evidence and original local wording;
- scale evidence or unknown state;
- inclusion/exclusion rule applied;
- duplicate/group relationship;
- decision, reviewer, and next action.
Measure accepted accounts per review hour, review rate, duplicate rate, and recurring rejection reasons by source. If a sector code repeatedly produces distributors, refine the rule before searching for more rows. If the team cannot agree whether a local branch is one account or part of its parent, solve that model before assigning territory or buying contacts.
Use the country-market-map ledger
Download the country-market-map ledger. Before filling rows, write the local entity boundary and negative controls at the top of the working brief. For example, a French operating site may be in scope while a French-registered holding company is not; that is an illustrative rule, not a fact about any named market. Keep the original local term and the source next to each decision so a later reviewer can tell a translated label from the underlying evidence. For the independent quality check, also record the second reviewer, review date, agreement result, disagreement reason, and whether the disagreement was reconciled.
Plan the first country-market review
Estimate the human review capacity needed to turn a country-market candidate set into a bounded, reviewed first sample.
The sample is capped at the candidate count. The accepted count is a planning assumption for the reviewed sample, not a market-size or coverage estimate.
Review 60 of 600 candidates at a capacity of 60 accounts per week. The provisional accepted outcome is 21 accounts.
| Market-map measure | Result |
|---|---|
| Candidate accounts | 600 |
| First reviewed sample | 60 |
| Provisionally accepted in sample | 21 |
| Total review hours | 12 |
| Weekly review capacity | 60 |
The chart keeps the candidate universe, reviewed sample, and provisional accepted outcome visibly separate.
Weekly review capacity equals reviewers multiplied by weekly review hours and divided by minutes per account. Weeks equal the first sample divided by that capacity, rounded up. Provisionally accepted accounts apply the entered share only to reviewed rows.
This plans review work; it does not prove market coverage, official company counts, current operating status, or the number of sales-ready accounts in a country.
6. Resolve entities before you calculate coverage
Most market-map errors are identity errors disguised as research progress. A legal name can represent a parent, a local subsidiary, a branch, or a company that has changed name. A brand site can cover several legal entities. Conversely, two different companies can share a trading name. Do not let a search result decide which relationship is commercially relevant.
Use a small entity-resolution protocol for every row that moves from candidate to accepted account:
- Keep the source identity. Record the name exactly as the source writes it, its country or register context, and the URL or identifier. Never replace the source name with a guessed English canonical name without retaining the original.
- Find the operating identity. Look for the domain, local address, local site, and activity evidence that support the particular unit in your account rule. A parent-domain match alone does not prove a local operating account.
- Record the relationship. Mark parent, subsidiary, branch, brand, distributor, successor, or unknown. “Unknown” is a valid relationship state; it is better than silently merging rows.
- Apply one duplicate rule. For example: merge only when the same commercial unit has the same canonical domain and the same local operating evidence; otherwise keep the rows linked and require review. The rule is an operating choice, not a universal identity standard.
- Preserve the losing row. When two records are merged, retain the merged identifier, reason, reviewer, and date. That allows a territory or analyst to reverse a wrong merge rather than rediscovering the account.
The important distinction is between an identity conclusion and a commercial scope conclusion. A registry may establish that two names belong to one group. Your team still has to decide whether that group should receive one account owner, two local motions, or no outreach. Do not allow an entity-resolution tool or a classification code to make that commercial decision implicitly.
A practical duplicate decision tree
Ask these questions in order. Stop once the evidence is insufficient and place the record in review.
This protects the denominator as well as the sales queue. If a map first says 500 candidates and later discovers that 80 are duplicate group records, the correct finding is not “we lost 80 accounts.” It is that the first estimate used a different unit. Version the denominator rather than hiding the change.
7. Measure coverage as a claim with a denominator
“We found the market” is not a measurable result. Coverage needs a named denominator, a date, a source mix, and a definition of what remains outside the map. This does not require pretending that a private-market universe is perfectly observable. It requires saying what the number means.
Start with a coverage statement such as: “As of the review date, this ledger contains candidates discovered through the listed sources that meet the written country and activity rule. It is not a census of all companies in the country.” Put that statement in the work brief and in any slide that uses the number.
Then separate these four measures:
Avoid averaging these measures into a fake precision score. A register may give broad identity coverage but low product context. A local event list may give high relevance but intentionally omit non-members. Their strengths are complementary. For source attribution, record one primary discovery path before review; label multi-source accounts as multi-source and exclude them from single-source rates rather than crediting every source with the same accepted account.
Use a stratified quality check
Reviewing only the easy records creates a flattering acceptance rate. Draw a small sample from each important discovery path, activity group, company-size band, and border case. For each sampled row, have a reviewer who did not add it decide whether the documented evidence supports the outcome. Record agreement, disagreement, and the reason for disagreement.
For a small first map, an operationally useful sample can be modest: for example, review a fixed number from every source stratum rather than claiming a statistically representative national survey. State that it is an operational quality check, not a population estimate. If one stratum creates repeated disputes, revise the inclusion rule and recheck records from that stratum before expanding it.
8. Work a hypothetical country-entry example
The following is deliberately hypothetical. It illustrates a method; it does not describe a real market, company, or expected Leadbase outcome.
Imagine a B2B equipment supplier evaluating a move into a target country. The commercial hypothesis is: “Prioritise local operating sites that publicly describe maintenance of installed industrial equipment, exclude pure distributors, and treat a local branch as a separate account only when it has its own service operation.”
The team writes three rules before research:
- Positive evidence: a current local service, maintenance, repair, commissioning, or field-support page tied to the target country/site.
- Negative evidence: a company describing itself only as import, resale, representation, or distribution without the qualifying operation.
- Review-needed evidence: a group-level page that mentions service but does not establish which local entity performs it.
It discovers 120 candidates from a classification-based source, 35 from a local exhibitor list, and 20 from a known-account expansion. After explicit duplicate linking, 145 unique candidates remain. A reviewer checks a balanced first 50. Thirty-one have positive local evidence, nine are distributors, six belong to the same parent but lack local-operating proof, and four are unclear. The result is not “there are 31 accounts in the country.” The defensible result is: “Under version 1 of the rule, 31 of the first 50 reviewed candidates entered the queue; 10 records need an entity/role decision before the source mix is expanded.”
That distinction changes the next action. The team should not search for 1,000 more rows. It should clarify whether group-level service claims count, then revisit the six uncertain group records. If the commercial owner says each local service site needs a separate motion, the ledger version changes. If the owner says group-level coverage is sufficient, linked sites may be rolled up. Both paths are valid only when they are documented.
9. Define the handoff gate before contact research
A market map becomes expensive when contacts are researched for accounts that later fail a basic entity or activity check. Add a clear handoff state to the ledger:
Do not convert “accepted account” into “good customer,” “reachable buyer,” or “lawful outreach.” Those are separate decisions. Keeping this gate explicit makes a Leadbase Sheet useful: the shared record holds the evidence and the human decision that justify moving a selected row to the next narrowly scoped enrichment task.
10. Retest the map when a premise changes
Every map needs a retest trigger. Dates alone are not enough. Retest when the country definition changes, a product line changes, the commercial unit changes, a major discovery source changes its coverage, a repeated rejection reason appears, or a territory owner finds a material false inclusion. Keep the former version and its results so the team can distinguish a market change from a rule change.
Set three owners: one for the ICP/account rule, one for entity-resolution decisions, and one for the commercial handoff. A single researcher can prepare evidence, but the person who benefits from a broader or narrower universe should not silently change the rule. This separation is a practical control against list-building bias.
What this guide does and does not establish
This workflow helps create a transparent account-research estimate. It does not establish market demand, addressable spend, legal permission to contact people, a forecast, or the completeness of any register. Use formal market research, local advice, and your own commercial data when those claims matter. The value of the ledger is simpler: someone can inspect why a candidate entered, left, merged, or remained unresolved.
A short first-run review protocol
Before work starts, schedule a 30-minute review with research, territory, and sales. Do not read every row. Sample five accepted, five excluded, and five unresolved candidates from different sources. Check only four questions together: does the evidence support the account unit, country condition, activity rule, and exclusion or merge reason? Turn every disagreement into a more precise rule, a review state, or a documented exception.
Then version the map and add a short change log: what changed, which already-reviewed rows it affects, and who decides whether they must be revisited. Give the map a visible validity window as well. That is not a promise that every row stays true until that date; it is the date or defined product, market, or organisational trigger that requires the team to recheck sources, criteria, and unresolved cases before relying on the map for a budget, territory, or campaign decision.
Where Leadbase changes the market-entry workflow
Market research and sales execution normally live on opposite sides of a handoff. The market model sits in a memo or deck; the sales database starts again with a country filter. The assumptions that created the market disappear just before the team begins spending time and contact credits on it.
Leadbase connects those two pieces by executing the market definition. Semantic discovery can find candidates that match what the segment actually does; custom research fields can then test the local operating, service, language, certification, or exclusion rules that a country filter cannot express. The map therefore does not end as a number on a slide. It becomes a qualified target universe built from the same business definition.
The operating bridge matters: the same evidence that defines the market decides which accounts are allowed into execution.
Once an account passes review, the team can select it for the next focused research task. If one field genuinely changes quickly, that research column can run on a schedule without pretending the entire country map is permanently current. The result is a market-entry workflow that preserves its assumptions all the way to handoff.
Leadbase cannot make an estimated market count exact and does not replace official registers, industry research, local expertise, or counsel. Its advantage begins after those inputs define the market: it can discover, research, qualify, and prepare the matching companies for contact work in one operating path. Build the first country pilot, or pressure-test the market definition with us.
Market-map decision record
Before an expansion budget or campaign depends on the result, record the country, source date, account unit, NACE or other classification version, inclusion/exclusion rules, discovery sources, duplicate policy, sampled evidence-support rate, unresolved share, and retest trigger. Define evidence-support rate as independently reviewed sampled rows whose documented evidence supports the recorded outcome ÷ independently reviewed sampled rows. That record is more valuable than a large but unexplainable country list.
