Why the third email in your sequence often beats the first

The first email in a sequence gets all the attention when a campaign is built — the strongest hook, the most personalization, the most editing. The third or fourth often quietly outperforms it. That’s not a fluke of subject lines or better timing. It’s one of the oldest, most replicated findings in social psychology, and it explains something cold email data keeps showing without fully explaining why.

The experiment that started it

In 1968, psychologist Robert Zajonc ran a deceptively simple test. He showed American participants nonsense words and unfamiliar Chinese-like characters — stimuli with zero built-in meaning — anywhere from zero to twenty-five times each, then asked people to rate how positively they felt about each one. There was no reward attached to any exposure, no reason to prefer one nonsense word over another except having seen it before. People rated the more-seen items as more likeable anyway. Zajonc called it the mere exposure effect: repeated, unreinforced contact with something is sufficient, on its own, to make people feel warmer toward it.

The curve his data traced matters as much as the headline finding. Liking didn’t rise in a straight line with exposure count — it rose fastest across the first several exposures, then flattened. The tenth exposure added far less warmth than the second did. That decelerating shape is the part almost every popular retelling of “the mere exposure effect” drops, and it’s the part that actually matters for deciding how many follow-ups to send.

Advertising already fought this exact battle

Four years later, advertising researcher Herbert Krugman took the same underlying idea and turned it into a specific claim about how many exposures actually change a decision. His argument, laid out in “Why Three Exposures May Be Enough,” was that each exposure to an ad does a different cognitive job:

Krugman’s actual point, often lost in the popularized “rule of 7” that descended from it, was closer to the opposite of “more repetition is always better.” He explicitly dismissed the idea that audiences need twenty or thirty exposures to remember a message as a myth used to justify oversized ad budgets. The number that mattered to him was small — three, not seven — and the mechanism was cognitive stages, not brute familiarity.

Where the two findings meet in a cold email sequence

Put Zajonc’s decelerating exposure curve next to Krugman’s three-stage model and a sequence’s shape gets a real explanation, not just a data pattern. A first cold email does the work of Krugman’s first exposure: register that you and your offer exist. A recipient skimming it once, cold, has no basis yet to trust it — Zajonc’s data says liking hasn’t had room to build. The second touch is where evaluation starts, helped by the fact that the name and subject are no longer completely unfamiliar. By the third touch, two things have happened at once: enough repeated, harmless contact has accumulated for the mere exposure effect to do its quiet work, and the reader has enough information to actually decide. That’s a plausible psychological account for why 42% of replies in a full sequence come from follow-ups, not the first email, and why a single follow-up alone has been measured to lift replies by roughly two thirds over a one-and-done send.

This is also the reasoning behind why the follow-up curve’s own data tops out around 4–7 touches rather than climbing forever. Zajonc’s curve flattens. Krugman’s model treats exposures past the third as mostly repeating a decision already made. Both point at the same practical ceiling: value keeps accumulating for a while, then the marginal touch stops buying you anything, and past a certain point in an inbox specifically, an unwanted repeated message risks becoming actively irritating rather than merely familiar — the well-documented “wear-out” effect in advertising research, where the same curve that first built warmth can bend back down if it runs too long.

What this actually changes about how you write touch two and three

Knowing why the third touch performs well changes what it should say. If Krugman’s framework is right, touch one only needs to clearly answer “what is this” — it doesn’t need to close. Touch two’s job is answering “what of it,” which means it should add a genuinely new reason to care, not restate the first email more politely. Touch three is where the decision crystallizes, which makes it the right place for the clearest, lowest-effort ask in the whole sequence — not the most creative one. A sequence written as three variations on the same pitch wastes the structure the psychology is handing you; a sequence written as orient, then evaluate, then decide uses it.

Put concretely, a three-touch arc built on this framework looks different from three restatements of the same pitch:

• Touch one (orient): a short, specific line establishing who you are and why you’re writing this particular person — nothing more. The job is registration, not persuasion.
• Touch two (evaluate): new information the first email didn’t have room for — a relevant result, a specific detail about their situation, the “so what” the reader needed to start actually weighing the offer.
• Touch three (decide): the lowest-friction ask in the sequence — “worth 15 minutes?” rather than a fresh pitch — because by this point the reader either already has enough to decide or needs one small nudge, not more information.

That structure also explains a pattern experienced SDRs notice but rarely articulate: rewriting touch one to be cleverer usually helps less than rewriting touch two or three to actually say something new. Zajonc’s data says touch one’s job is mostly done just by existing and being read once; Krugman’s model says the persuasion work happens later. Effort spent polishing the part of the sequence that’s structurally not where the decision gets made is effort spent in the wrong place.

Isn’t this just “spam them until they cave”?

It’s a fair objection, and the research itself draws the line. Zajonc’s curve is decelerating and eventually flat, not endlessly rising — the mere exposure effect describes diminishing returns from more contact, not unlimited returns. Advertising’s own wear-out research is explicit that repetition past the effective range can flip warmth into irritation, which is a worse outcome than never sending the follow-up at all. The claim in this article isn’t “more touches are always better.” It’s narrower and more specific: the first few touches are doing real, underrated cognitive work — recognition, then evaluation — that a single email can’t do by itself, and stopping after touch one abandons a sequence right where the curve is steepest and least conclusive, not after it’s flattened.

How Norbelys builds the sequence around this

Getting the shape of a sequence right by hand, across every contact and every send window, isn’t realistic to track manually — so Norbelys builds the structure in. You write each step of a sequence once, and Norbelys handles the spacing and pacing automatically, keeping the touches close enough together that exposure compounds instead of resetting, without ever sending fast enough to look automated. You can A/B test any individual step — including testing a direct, low-effort ask at touch three against a softer one — and the winner is decided by human-verified reply rate, not opens most of which are robots. And the built-in circuit breakers watch bounce and complaint rates across the whole sequence, so chasing the third-touch effect never risks tipping into the wear-out zone the same research warns about.

Zajonc’s nonsense words had nothing to offer and still got warmer with repetition. Your third email has an actual reason for someone to reply. Give it the chance the psychology says it’s earned — start a sequence on Norbelys and see where your own reply curve actually peaks.

Follow-up psychology, quickly

Does this mean I should always send at least three touches?+

The research suggests three is where a decision often crystallizes, not that fewer than three always fails or more than three always helps — Krugman's own point was that the number needed is smaller than most marketers assume, not a strict minimum. The safer takeaway is structural: give the first touch room to just register, the second room to add real new information, and don't expect the sequence's persuasion to be finished after one email.

What's the difference between the mere exposure effect and just being annoying?+

Zajonc's effect is about unreinforced, low-friction exposure — seeing a name again without it costing the reader anything. Annoyance shows up when repetition stops feeling low-cost, usually because the messages are too frequent, too similar, or arrive after the reader has already made a decision. That's the same 'wear-out' pattern advertising researchers document past the effective exposure range, and it's exactly why a sequence should stop around 4-7 touches rather than running indefinitely.

Does the three-exposure idea apply to subject lines too, or just the sequence as a whole?+

Krugman's framework was about exposure to a message overall, not a specific email element, so it applies most directly to the sequence's arc rather than to, say, reusing the same subject line three times. A repeated, recognizable sender name and consistent thread context across touches is closer to what the research describes than repeating identical subject-line wording would be.