Personalisation at scale without the creepiness

Personalisation at scale means personalising the segment, not the individual. Relevance comes from writing one precise message to two hundred firms that share the same problem — not from scraping personal details about each recipient and stitching them into a template. Done properly, every email feels personal because it is accurate; done badly, it feels like surveillance with a mail-merge.

Why doesn't individual personalisation scale — and why doesn't shallow personalisation work?

The two standard approaches both fail, in opposite directions.

Hand-written personalisation — ten minutes of research per prospect, a genuinely individual opener — produces good emails at a rate of maybe 20 a day from a full-time human, which is why a BDR doing it costs £35k+ a year and still starves the pipeline. It fails on arithmetic.

Token personalisation — "Hi {FirstName}, I see {Company} is doing well in {Industry}" — scales beautifully and convinces nobody. Readers have seen a decade of mail-merge; a first name and a company name signal automation, not attention. It fails on credibility.

The resolution, as set out in The Complete UK B2B Outbound Playbook, is to move personalisation up a level: from the person to the segment.

What is segment-level personalisation?

It is the observation that relevance is a property of the problem, not the name. A recruitment firm with 12 staff and a founder still closing every deal has, with high probability, the same pipeline problem as the next recruitment firm with 12 staff. One email that names that problem precisely will read as personal to both — more personal, in fact, than a clumsy reference to either founder's LinkedIn activity.

So the craft moves into the list. When your segments are tight — one sub-vertical, one headcount band, one buying trigger — the "personalisation" is baked in before a word of copy exists. When your list is a broad scrape, no token can rescue it. Good copy cannot save bad targeting; a tight segment saves average copy daily.

What actually reads as creepy?

Anything that crosses from their business to their person. Referencing someone's children, home town, gym, or holiday photos; quoting a post from a personal account; mentioning where they studied fifteen years ago. It demonstrates effort and destroys trust simultaneously — the reader's question shifts from "is this relevant?" to "what else does this stranger know about me?"

The line I hold: public, professional, and recent is usable — a funding announcement, a hiring spree, an office move, a new service line. Personal, private, or scraped-from-socials is not, even when it is technically public. There is also a compliance angle: under UK GDPR, prospect data processed on legitimate interest grounds should be proportionate and business-relevant, and hoovering up personal detail for flavour sits poorly with that (not legal advice — just the safe side of the line).

What is the mechanism for personalising at volume?

It runs through the database, not the copywriter. When you build the list, you capture the fields that change the message: sub-vertical, headcount band, region, and one trigger field such as "hiring for sales" or "opened second office". When the list is segmented on those fields, you write one email per segment — the first line names the segment's situation, the body names the problem that situation predicts, and the proof point matches the sub-vertical. When a token appears in the copy, it carries information rather than decoration: "{headcount} consultants" earns its place; "{FirstName}" merely fails to. When the campaign runs, every reply is read as a test of the segment definition — a cluster of "not relevant to us" replies means the segment leaks and gets re-cut.

AI-assisted enrichment fits here as a research clerk, not an author: models are decent at classifying firms into sub-verticals and spotting trigger events at volume, and unreliable at writing "personal" openers that do not sound uncanny. Use them to fill fields, and keep a human on the sentences.

How do you know your personalisation is working?

The same arbiter as everything in outbound: positive replies, tracked per segment. Around 4% is healthy; below 3% you fix the campaign — and personalisation depth is the suspect after targeting. A segment out-replying another by double is telling you where relevance lives, which is exactly the sort of number that should surface on a weekly dashboard rather than in a quarterly archaeology dig — the MD Dashboard Blueprint covers that layer.

One caution as you scale: when a segment's message works, resist the urge to endlessly rework it for novelty. Personalisation is a targeting discipline, not a creativity contest — add new segments alongside, never rotate the working ones out. And when nobody in the firm has time to run any of this by hand, that is a systems problem with its own answer.

---

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