Personalization vs relevance: which one gets replies?

What is the difference between personalization and relevance?

Ask what the first line proves. A personal detail proves effort: you spent time on this person. A relevant line proves fit: you know what is going wrong in jobs like theirs. One line can prove both, but fit is the part a buyer can answer.

Swipe to see more →

Relevance answers three questions: what is going wrong, why now, and why it lands on this person's desk. If your first line covers none of them, it is decoration.

Is there proof that relevance beats personalization?

No, and here is what we found instead.

• Buyer survey, not a reply test. In a Gartner survey of 632 B2B buyers (August and September 2024, published June 2025), 73% said they actively avoid suppliers that send irrelevant outreach. That tells you what buyers say they dislike. It does not tell you how much a relevant email lifts replies, and it does not compare relevance with personalization.
• A real experiment, wrong setting. A Stanford Graduate School of Business working paper (2016) ran field experiments with three companies, reaching millions of people, and found, in its main experiment, that putting the recipient's name in the subject line of promotional emails made people about 20% more likely to open them (a relative lift). That is promotional email with name-only personalization, judged mainly on opens (the paper also reports more sales leads). It says nothing about B2B cold outreach to strangers.

Vendor blogs make bigger claims, but the ones we found give no sample size or method, so we don't repeat their numbers. We found no independent, published test that isolates a personal detail from a relevant problem in B2B cold email. So there is no honest "relevance gets X% more replies" number, and we aren't going to invent one. The next section gives you the way to find out for your own market.

How do you decide where your research time goes?

Decide by list, not by person. Our guidance, not a study:

Swipe to see more →

The line to stop writing is the one that cost research time and still names no problem: a compliment, a shared hometown, a mention of their weekend race. It spends the most effort for the least reason to reply.

The groups come from your ideal customer profile, and the time caps per tier are on our page about how much time to research a prospect.

Do the math on reuse. One problem statement written for a group of 60 is research you pay for once. A personal fact pays for one email, then it's spent.

What does this look like in real emails?

Same prospect, three openers. All invented for illustration.

• Personal only: "Saw you spoke on a panel last month. Great points." Nice. So what? Nothing in it connects to your offer.
• Relevant only: "Companies that bring in a new VP of sales usually rework territories within a quarter, and deals start colliding." Nothing about the person. It would fit any company that just hired a new VP of sales, and that is the point.
• Both: "Saw you just hired a new VP of sales. Companies that do usually rework territories within a quarter, and deals start colliding. How are you handling it?" The detail (the new VP) is the evidence for the problem.

The last one is the pattern. Start from a trigger you can see from outside, such as a hiring push or a new sales leader; our page on sales trigger events lists six. Then check the line from the reader's side: what could they reply to it? If the only honest reply is "thanks," the line names no problem. Rewrite it until it asks about something they would have to think about. For the full email structure around it, see cold emails that work. For what to do when the reader answers, see how to reply to "not interested" or "send me info" and what words get a prospect to put a time in the calendar.

How do you test it on your own list?

Run a three-arm test. Same segment, same offer, same send window, random split. Only the first line changes.

Swipe to see more →

Track replies and, separately, positive replies; our page on reply rate vs positive reply rate vs meeting rate explains why they differ. Do not judge arms by opens.

Our example, invented numbers: your current reply rate is 3%. You want to know if one arm can reach 5%. Our math, using the standard two-proportion test at 95% confidence and 80% power: about 1,500 sends per arm, or about 4,500 emails across three arms. To detect 3% to 4.5%, you need about 2,500 per arm; for 3% to 4%, about 5,300 per arm. These figures are for comparing two arms at a time; with three arms you make three comparisons, so treat a borderline winner with extra caution.

Three cautions:

1. Small samples lie. At 100 sends per arm, a 3% rate is three replies. A difference of two replies is noise. Don't call a winner early.
2. Change one thing. Don't change the subject line, the offer and the first line together.
3. A lift in one list isn't a law. Your result applies to your segment. Rerun it when your audience changes.

For the full method, see how to A/B test cold emails. If you can't reach the volume, test on a longer window rather than reading tiny differences.

When should you personalize anyway?

Personalize one-to-one when the account is worth the time, and when the detail is the proof of the problem. Skip it when:

• You are forcing it. If you have to stretch to connect the detail to your offer, the reader will feel the stretch.
• The list is uniform. If 300 contacts share a role and a trigger, write one strong segment email.
• The detail is private. Don't mention things from a prospect's personal social accounts. It reads as surveillance, and relevance never needs it.