The data points worth personalising on

The data points worth personalising on are the ones that change what you would actually say: the prospect's role, their sub-vertical, their headcount band, and any visible trigger event. Everything else — alma mater, hobbies, a recent LinkedIn post — is decoration, and readers increasingly recognise it as automated flattery. If a data point cannot change a sentence in the email, it is not worth collecting.

Why does most personalisation fail?

Because it optimises for familiarity instead of relevance. "Congratulations on eight years at the firm" proves you ran a lookup; it does not prove you understand the reader's problem. B2B buyers commonly skim a cold email in a few seconds, and in that window a line about their operational reality outperforms a line about their biography.

Which fields you can even use is decided upstream, when the list is built — the full pipeline is laid out in the B2B Database Building Guide. Personalisation is a data problem before it is a writing problem: good copy cannot save bad targeting, and it cannot reference fields the list never held.

Which data points actually change the message?

Four categories earn their place, in rough order of value:

• Role and seniority. An MD of a 12-person firm who still does all the selling has a different problem from a sales director managing a team of five. The pitch, the pain named, and the proof offered all change.
• Sub-vertical. A recruiter, an MSP, and a consultancy each have a distinct pipeline pattern. Naming it — contingency feast-and-famine, contract renewal cycles, partner-led referral dependence — signals you have seen their world before.
• Headcount band. A 6-person firm and a 45-person firm sit at different stages of systems maturity. Size determines whether you talk about getting a first outbound engine running or replacing a failing one.
• Trigger events. Hiring for a sales role, a new director, an office move. These are readiness signals — facts that suggest timing, not just fit — and they justify a sentence that references them directly.

Which data points are decoration?

University, sports team, company founding year, and generic praise of a recent post. The tell is simple: none of them changes what you are offering or why it is relevant now. AI-generated first lines built on this material have flooded inboxes, and readers have adapted; a compliment that could have been generated commonly gets treated as if it was.

How do you decide what to collect?

Run every candidate field through one test before it earns a place in your enrichment spec. The mechanism:

1. When a data point changes which segment a contact belongs to, keep it as a segmentation field.
2. When it changes a sentence inside the email, keep it as a merge field.
3. When it does neither, drop it.

Collection is not free. Every field must be found by the enrichment waterfall, every field can be wrong, and a wrong merge field is worse than none — "as a recruiter" sent to an accountant ends the conversation. Fewer fields, held to a higher accuracy standard, beat a wide spreadsheet of half-right guesses.

Should you personalise every email, or every segment?

For most of the list, personalise the campaign, not the email. Write one message per segment that names that sub-vertical's specific problem, and let the shared context do the work — segmentation done properly delivers most of personalisation's value at a fraction of its cost. At 25–40 cold emails a day per inbox, hand-crafting each one is arithmetic nobody survives; reserve individual research for a short priority list where the deal size justifies it.

One caveat on AI-written first lines: grounded in a verifiable, message-relevant fact, they can read fine. Left to freewheel, they produce the same confident vagueness for everyone, and the reader has seen it a thousand times. The data point rule still governs — the model can only personalise on what the list holds.

What has to be true before any of this matters?

The address has to work. Personalisation spent on a contact that bounces is worse than wasted, because the bounce damages sender reputation for every email that follows. Verify before you send — then let the surviving records carry the smaller, sharper set of fields that actually change the message.

---

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