Personalisation has a strange reputation. Buyers say they want it, marketers say it works, and yet most people can remember an email or ad that made them feel uncomfortably watched.
Both reactions are real. The difference between them is not how much personalisation you use. It is what you base it on, and how you express it.
People expect relevance
McKinsey’s research found 71% of consumers expect companies to deliver personalised interactions, and 76% get frustrated when that does not happen. Personalisation most often drives a 10–15% revenue lift (McKinsey).
In B2B, irrelevance is punished directly. Gartner’s 2025 research found 73% of B2B buyers actively avoid suppliers who send irrelevant outreach (Gartner).
So the question is not whether to personalise. It is how to do it without crossing the line.
The expectation test
Before using a piece of data, ask: would this person expect us to know this?
Usually fine
- Their company’s industry, size and public business context
- Their role, as they told you or as publicly listed
- Topics they chose to subscribe to
- Content they downloaded or events they attended with you
- Products they use or asked about
Usually uncomfortable
- Personal life details found elsewhere
- Precise tracking of their browsing, stated back to them (“we saw you looked at pricing three times yesterday”)
- Inferences about sensitive topics
- Anything you are guessing but presenting as fact
Personalise the substance, not the surveillance
The best personalisation often does not mention the data at all. It simply makes the message more useful.
- Instead of “We noticed you visited our pricing page”, send a clear, helpful comparison of plans.
- Instead of “As a logistics company in Rotterdam…”, use an example from a similar logistics company.
- Instead of listing what you know about them, show that you understand their problem.
Relevance feels like help. Displayed data feels like surveillance.
Respect consent and control
- Collect data transparently and use it for what you said you would.
- Let people choose topics and frequency, not just subscribe or unsubscribe.
- Follow the privacy laws that apply to your audience, and keep consent records.
- Make opting out easy; it protects both trust and deliverability.
AI raises the stakes
AI makes personalisation cheap. It can draft a unique message for every contact, every account and every moment. That is powerful, and it removes the natural limit that manual effort used to provide.
With that power come new risks:
- Confident errors. AI can state wrong facts about a company as if they were true.
- Over-personalisation. Messages packed with details can feel intrusive.
- Scale. A small mistake in a prompt can be repeated across thousands of messages.
The answer is clear rules and human review:
- Define which data AI may use for personalisation, and which it may not.
- Require verification for any specific fact about a company.
- Review samples of every personalised campaign before sending, and every one-to-one ABM message.
- Monitor replies and complaints for signs that something feels off.
The takeaway
The best personalisation makes people think “that is exactly what I needed”, not “how do they know that?”. Use data people expect you to have, personalise the value rather than the surveillance, respect consent and keep humans in charge of what AI produces.
Responsible personalisation is built into our AI-powered marketing and email marketing services. For ABM specifically, read how AI helps BDRs open better conversations.
