Sub-vertical targeting: broad firm, narrow campaigns

You can serve a broad market and still never send a broad campaign. Sub-vertical targeting means the firm stays generalist while every outbound campaign speaks to one narrow group — recruiters, MSPs, architects — in that group's own language. Positioning is decided per campaign, not per firm, so you get the reply rates of a specialist without turning work away.

Why do broad campaigns underperform?

Because the reader decides in seconds whether an email is about them, and "businesses like yours" fails that test for everyone. A generic message has to gesture at problems; a sub-vertical message can name them — the contractor who went direct, the retainer that did not renew, the compliance deadline in April. On a well-targeted list, around 4% positive replies is a reasonable expectation and below 3% means the campaign needs fixing; broad campaigns live below that line because the copy cannot be specific to an audience that is not.

This is a targeting decision before it is a copy decision, which is why it sits inside The B2B Database Building Guide: the segment is chosen at list-build time, not at writing time.

What counts as a sub-vertical?

Not "professional services". A sub-vertical is a group narrow enough to share three things: a vocabulary, a pipeline pattern, and a set of objections. "Independent recruitment agencies, UK, 5–50 staff" qualifies — they all think in placements, splits and client terms. "Agencies" does not; a PR agency and a PPC agency barely share a problem. The test: could you write one email that every firm in the group would recognise as being about their business? If yes, it is a sub-vertical.

Note the order of operations. First define the ICP — precision before size — then partition it into sub-verticals. Sub-verticals are slices of the ICP, not a replacement for it.

How do you run narrow campaigns without narrowing the firm?

The mechanism:

1. Partition the ICP into sub-verticals — three to six is typical for a firm serving the UK 5–50 staff market.
2. When the partition is set, build one list per sub-vertical. Each gets its own build: SIC-code filters against the Companies House register, then people, then verified addresses. Never one mixed list with a "segment" column that copy ignores.
3. When each list exists, write one sequence per sub-vertical — 4 emails over 14 days, in the buyer's vocabulary, naming their pipeline pattern.
4. When sequences are written, run the campaigns in parallel. Each mailbox sends 25–40 cold emails a day; when you want to add a sub-vertical, you add a campaign and, if volume requires it, a mailbox. Add campaigns; never rotate one mailbox across audiences mid-flight.

The firm's website can keep saying "UK B2B service firms". The inbox of a recruitment MD only ever sees the recruiter campaign.

How do you pick the first sub-vertical?

Where you have proof. The first campaign should go to the group where you hold your best client stories, your most fluent vocabulary and your clearest before-and-after numbers — because that campaign will also teach you the mechanics (deliverability, reply handling, booking) while the message risk is lowest. Expand one sub-vertical at a time, and let reply quality — not gut feel — decide which segment gets the next build.

What does this change in results?

Three things, in order of appearance. Reply rates rise, because relevance rose. Reply quality rises further — the replies you get come from people who recognised themselves, so discovery calls start halfway done. And learning compounds: every send teaches you about a defined group, so the second campaign to a sub-vertical is measurably sharper than the first.

None of this is automatable away. AI will happily draft six sequences and a tool will happily send them, but choosing the partition, knowing which group you have proof for, and hearing what the replies are telling you — that is judgement applied through a system, which is the difference between owning a pipeline and owning a subscription. AI won't fill your pipeline; a system using AI will, and sub-vertical targeting is one of the load-bearing parts of that system.

---

Next step: the Growth System Audit — £450, seven days, credited against any build — maps where your growth system leaks and what to build first.