Semantic clusters: find targets in the language of their trade
To identify targets that databases cannot see, the engine does not start from an industry code. It starts from the vocabulary companies in your market actually use. The method has three steps, stays readable end to end and works in the language of each country.
From industry codes to the real vocabulary of companies
An industry code files a company in an administrative box. The vocabulary it uses on its website describes what it really does: its processes, materials, certifications, references and markets. That vocabulary is what separates two competitors a classification would file side by side.
From your criteria and the benchmark companies you already know, the engine builds your target’s semantic cluster: the full universe of language that characterises it, with its variants, trade synonyms and the nuances that separate two seemingly identical businesses.
Three steps, from acquisition thesis to validated list
- Step 1 · Framing the thesis
- what you are looking for, why you cannot build it yourself, what a good target looks like. You add the benchmark companies you already know.
- Step 2 · Semantic cluster
- processes, materials, certifications, references and trade synonyms, in the language of the target market. The cluster defines the scope, not the industry code.
- Step 3 · Scan and list
- every company whose language matches the cluster is identified, whether or not it publishes its accounts. You validate the list before any action is taken.
A search scope written in plain language, open to debate
The business scope is not an administrative code but a written definition, validated with you. In cleaning services, for example: “companies specialising in B2B cleaning and related services, working mainly in routine building cleaning, industrial cleaning, cleaning of logistics and distribution sites, and sensitive environments”.
The definition comes with visible keywords. For a search in France, that example used fourteen terms, in French and English: nettoyage, cleaning, nettoyage industriel, propreté, service de nettoyage, facility management, cleaning services, nettoyage des bâtiments, entreprise de propreté, services de propreté, facility services, maintenance des locaux, entretien des bâtiments, nettoyage de bureaux.
For each company identified, an indicator summarises its alignment on five criteria: target activity, applications, keywords, reference URL and activities to exclude. Each criterion can be opened and edited during the working session.
This transparency changes how you work. Most sourcing tools present a list without showing the rule that produced it. Here the rule is readable: if the list looks poorly calibrated, you change three keywords and the list rebuilds itself.
Searching Germany in German
The cluster is built in the language of the target market. Looking for a component manufacturer in Germany means working with German technical vocabulary, not a translation of French or English terms. The words a German company uses to describe its process are not a literal translation of ours.
The structural consequence: the same acquisition thesis can be run in several countries at once, with a consistent scope. That is the difference between genuine international coverage and coverage on paper.
What each line of the list tells you
The list a campaign produces, the deal set, is read line by line. Location, headcount and revenue immediately rule out whatever falls outside your thesis. A relevance score, between 69 and 96 on the extract shown in our documentation, measures how closely each company aligns with the cluster and the validated criteria.
Trend indicators flag changes in headcount. They matter more than they seem: a shrinking company and a fast-growing one are not approached the same way, or at the same time. Next comes the approach to the owner, described in Direct approach.
If the list looks poorly calibrated, you don’t change tools: you change three keywords, and the list rebuilds itself.