Firmographic Segmentation for B2B Sales and Marketing

Choose attributes that change the conversation

Useful firmographics are the ones that change what you say or who you say it to: industry, employee size band, group structure, geography, and where relevant, regulatory status or technology footprint.

Attributes that do not change routing, message or price are reporting fields, not segmentation fields.

Fix classification before you segment

Industry codes assigned from self-reported sources drift badly at the edges. A parent code applied to a diversified group can put a specialist subsidiary in the wrong segment entirely.

Re-derive classification from what the company actually sells, and record the basis for the decision.

Size bands need a consistent source

Employee counts from mixed sources produce bands that are not comparable. Pick one primary source per region, define the fallback order, and apply it consistently across the base.

Watch for segments that hide their own variance

A segment with a wide spread of deal sizes and cycle lengths is not a segment; it is an average. Split it until the behaviour inside each group is genuinely similar.

Apply segments to real campaigns

Test a segmentation model by running two differentiated campaigns against it. If the response profiles look identical, the split was cosmetic.

Where classification goes wrong at scale

Two failure modes dominate. First, group codes applied to specialist subsidiaries, which puts a niche business in a generic segment. Second, historical codes that were accurate at registration and never revisited after the company pivoted.

Both are invisible in a dashboard and obvious the moment a seller reads the account list. Re-deriving classification from current activity, with the basis recorded, is the only durable fix.

Making size bands comparable across regions

Employee data availability differs sharply by country. Mixing a filing-based count in one market with an estimate in another produces bands that cannot be compared, which quietly distorts every cross-market report.

We set a primary source per region and a documented fallback order, then flag which method produced each value so analysis can exclude the weaker tier when precision matters.