Most email that fails is not rejected. It is not marked as spam, it does not get unsubscribed from, and it does not even get read and dismissed. It simply stops being seen by a brain that learned to skip it without anybody consciously deciding anything.

Which is a problem, because your reporting cannot see that happening. An open is measurable, a click is measurable, an unsubscribe and a complaint are both events, with timestamps and rows in a table. A person whose visual system has quietly stopped registering your sender name generates nothing at all, and the space where the data should be looks identical to somebody who is busy this week.

So the most consequential thing happening to your programme is invisible to the tools you use to judge it, and the way to understand it is not to look at your ESP. It is to look at what is known about attention, repetition and memory, most of which has been settled science for decades and almost none of which shows up in email advice.

What follows is that science, and what it means for how often you send, whether repeating yourself works, and why the person who never unsubscribes is not necessarily the person who wants to hear from you.

The mechanism

Habituation, and why sending more eventually stops working

The single most useful idea in this whole subject comes from behavioural neuroscience rather than marketing, and it explains almost everything people misdiagnose as content getting stale.

Habituation is a reduction in response to a stimulus that has been repeated, where the reduction is not caused by tiredness of the senses or of the muscles. It was characterised by Thompson and Spencer in 1966 and revised by Rankin and colleagues in 2009, who set out nine defining behavioural characteristics of it. Habituation is a form of learning, arguably the simplest one there is, and it shows up in virtually every organism that has been tested for it.

Worth being precise about what that means for your subscribers. When somebody stops noticing your emails, they have not decided anything, they are not annoyed with you, and they have not evaluated your content and found it wanting. Their nervous system has learned that a particular signal reliably carries no consequence, and has stopped allocating attention to it. Exactly the same process by which you stop hearing your own fridge.

Four of the nine characteristics matter enormously for email, and I have never seen any of them written about in an email marketing context, which is a shame because each one is a lever.

One: frequency accelerates it

The more frequently a stimulus is repeated, the faster and more pronounced the habituation. Sending more often does not simply produce diminishing returns, it actively speeds up the process by which you become invisible. Which is the mechanism underneath every programme that increased frequency, saw a short lift, and then watched everything decline to a level below where it started.

Two: it is specific to the stimulus

Habituation attaches to the thing that was repeated, and the response still occurs to other stimuli. Somebody can be completely habituated to your weekly promotional email and fully responsive to a different kind of message from you on the same day. Silence in one stream does not mean silence in all of them, and treating a habituated segment as a dead segment throws away people who would answer a different question.

Three: it generalises to similar things

Habituation to one stimulus develops to other stimuli that resemble it. Changing your subject line while keeping the same sender, the same shape, the same rhythm and the same kind of ask does not present a new stimulus, it presents a variation on the old one, and the response stays suppressed.

Which is the uncomfortable news for anybody whose answer to declining engagement has been better copy. Copy is a small variation inside a much larger pattern, and the pattern is what got learned.

Four: rest restores the response

Spontaneous recovery is the characteristic almost nobody uses, and it is a gift. When the stimulus stops for a period, the response returns. More striking still, recovery tends to be more rapid where the habituation was produced by frequent stimulation in the first place.

Read that again with a segment in mind. The group you have habituated fastest, by emailing them most, is the group most likely to come back to you if you leave them alone for a while. Reducing frequency to a quiet segment is not giving up on them, it is the documented mechanism for getting their attention back, and it is the exact opposite of what most programmes do when numbers fall.

Two problems

There are two ways of not being seen, and they need different fixes

Web research made a distinction that transfers cleanly to the inbox and that almost every email team collapses into one problem.

The important part is the distinction it produces.

• Blindness operates at the format level. The brain has built a mental model of what a certain type of thing looks like, and filters anything matching it before conscious attention engages. In an inbox, your sender name is the format, and once it has been learned as skippable the eye passes over it without any event happening at all.

• Fatigue operates at the creative level. The brain ignores a specific thing it has seen too many times. In an inbox, that is your campaign, your offer, your recurring subject line structure.

Most teams work exclusively on the second problem while suffering from the first. New creative, new templates, a rebrand of the newsletter, all of it operating inside a sender identity that has already been categorised and filed. Fixing fatigue when you have blindness is redecorating a room nobody walks into.

And if you doubt how completely attention can fail, Simons and Chabris demonstrated in 1999 that fully half of observers counting basketball passes missed a person in a gorilla suit walking into the middle of the shot, stopping, beating their chest and walking off. Nine seconds, in plain view, unmissable to anybody who was not busy. Attention is not passive reception, it is an allocation decision, and the inbox is a place where people are extremely busy.

Repetition

Does repeating yourself work? Yes and no, and the two answers point in opposite directions

The question comes up constantly and gets answered badly in both directions, usually because the person answering has only got hold of one half of the evidence.

The case for repetition

Zajonc published the mere exposure effect in 1968, showing that simply being exposed to something repeatedly improves your attitude towards it. In his experiments people saw nonsense words zero, one, two, five, ten or twenty five times, and the words seen five, ten and twenty five times were rated more positively than the ones seen barely at all. The effect has since been demonstrated across sight, sound, smell and taste, and Kunst-Wilson and Zajonc showed it works even when the exposure is subliminal, meaning it does not require you to consciously notice anything.

The explanation most favoured is processing fluency. Familiar things are easier for the brain to handle, ease feels good, and we misattribute that pleasant ease to the thing itself rather than to our own familiarity with it. Which is a slightly deflating account of how brand preference works, and also a very useful one, because it means consistency is doing work that cleverness cannot.

The case against

Even in Zajonc's own data the effect is not linear. Preference rose with exposure frequency up to a point, and then stayed flat or declined. Meanwhile habituation is running in parallel, reducing the attention that any of it receives.

So you have two processes operating on the same person at the same time, in opposite directions, from the same input.

What breaks habituation is consequence, not creativity

A signal that sometimes matters keeps earning attention. A signal that never matters gets filtered regardless of how well it is designed, how good the copy is, or how much the template cost.

Which moves relevance out of the category of nice-to-have and into the category of the thing keeping you visible at all. Not because being relevant is virtuous, but because consequence is the only known reliable defence against a process that is otherwise automatic and inevitable.

Relevance

Why relevance beats cleverness, with a mechanism attached

Most advice about standing out in the inbox assumes attention works one way: make something loud enough or odd enough and it will be noticed. Attention research has been telling us otherwise since the early nineties, and the finding is the most useful one in this whole field for anybody who sells to busy people.

Folk, Remington and Johnston published work in 1992 showing that involuntary attention shifts are contingent on what the person is currently set up to look for. In their experiments a cue only captured attention when it shared the property the participant was searching for. An abrupt onset cue pulled attention when the target was defined by abrupt onset, and produced nothing when the target was defined by colour. Salience on its own did not do it. Matching what somebody was already looking for did.

Which reframes the whole business of standing out. Your reader arrives at their inbox with an attentional set already loaded, made up of whatever they are trying to get done that morning. Anything matching that set gets attention involuntarily and almost instantly. Anything not matching it can be as bright, as clever and as well-crafted as you like and it will be filtered before conscious thought engages.

So being noticed is not a property of your email. It is a relationship between your email and what the reader is currently looking for, which is why the same message performs differently for two people on the same morning, and why context beats best practice every single time.

What personalisation is supposed to be doing

Rogers, Kuiper and Kirker demonstrated in 1977 that information encoded in relation to the self is remembered considerably better than the same information processed for its structure, its sound or even its meaning. Their conclusion was that the self operates as a superordinate schema, deeply involved in how personal information gets processed, interpreted and stored. A meta-analysis by Symons and Johnson in 1997 confirmed the effect holds up.

Notice what that does and does not license. Self-relevance is a genuine and powerful encoding advantage, so making an email about the reader's situation is not a soft marketing preference, it is a memory mechanism. But putting somebody's first name in a subject line is not self-reference in any meaningful sense, because nothing about the content has been encoded in relation to them. It is a token attached to a message that remains about you.

First name tokens did work once, for the same reason your name cuts through a noisy room, and they have been repeated so universally and so emptily that they now sit squarely inside the habituation problem described earlier. The cue survived; the consequence behind it did not.

Awareness 

The real case for sending regularly

If repetition erodes attention, the obvious conclusion is to send less, and that conclusion is wrong for a reason worth understanding properly.

Professor John Dawes of the Ehrenberg-Bass Institute set out what became known as the 95:5 rule in 2021, showing that up to 95% of buyers in a category are not in the market at any given moment, and may not be for months or years. His conclusion about what that means for advertising is the part that matters here: if most of the people you reach will not buy any time soon, then advertising is mainly working by building and refreshing memory links to the brand, which activate later when those buyers do come into the market.

Applied to email, you are not sending in order to convert this week's list. You are buying presence at a moment you cannot predict and will not be told about, for a person whose need has not arrived yet. Judging a send purely on what it converted that week is measuring the smallest and least interesting thing it did.

Being remembered is weak, being linked to the moment is strong

Ehrenberg-Bass measure mental availability through category entry points, meaning the specific situations in which a category might be bought or used, an approach set out by Romaniuk and Sharp in How Brands Grow Part 2. The goal is not general recall, it is the number of buying situations that bring your brand to mind.

The scale of the task is worth knowing, because it is humbling. Dawes notes that many well established brands achieve no more than 20 to 30% of respondents linking them to a category entry point, and even market leaders often reach only around 50%. Mental availability is a multi-year project, not a campaign.

For email the implication is a content one. Writing endlessly about your product builds a link between your brand and your product, which is a link people already have. Writing about the situations that create the need builds links to the moments when somebody will think of you, and those moments are where purchases start.

Why the gaps matter as much as the sends

Ebbinghaus established two things in 1885 that email people should know and mostly do not. The first is the forgetting curve, showing that memory for something new decays sharply and then more slowly, unless it gets reinforced. The second is the spacing effect, showing that the same total amount of exposure produces substantially better retention when it is distributed over time than when it is crammed together. Over a century of subsequent work has held both up.

Which gives you the argument against the campaign burst, and it is a stronger argument than the usual one about annoying people. Six emails in a fortnight and then silence for two months is the massed condition, and the massed condition is the one that produces worse retention from identical effort. Regular, spaced contact is not merely more polite, it is a better use of the same resource.

And it explains why not sending enough is its own problem rather than a safe default. Go quiet for long enough and the forgetting curve does its work, so when you return you are not resuming a relationship, you are introducing yourself to somebody who now has no memory of agreeing to any of this. Which is precisely the point at which people press the spam button.

Trust

How trust builds, and how much faster it comes apart

Email is a trust channel before it is anything else, and there is a body of evidence explaining why trust behaves so asymmetrically in it.

Baumeister, Bratslavsky, Finkenauer and Vohs reviewed the literature in 2001 under the title Bad Is Stronger Than Good, finding the pattern across everyday events, major life events, close relationships, social networks and interpersonal interactions. Negative events reliably have a greater psychological impact than equivalent positive ones, and the asymmetry is consistent enough that they treat it as a general principle rather than a curiosity.

Applied to an inbox, a kept promise moves the needle a little and a broken one moves it a lot. The subject line that oversold, the discount that turned out to have conditions, the unsubscribe that did not work, the frequency that quietly tripled after signup. None of those produces a complaint on the day, and each one makes the next email slightly less likely to be believed.

So trust does not accrue at the rate you are spending it, and the arithmetic is not in your favour. Which is the case for saying less and delivering it, rather than promising more and hoping the average holds.

Why giving first works, and why it is not a trick

Gouldner set out the norm of reciprocity in 1960 as something close to a human universal: people feel obligated to return benefits they have received. Regan demonstrated it experimentally in 1971, finding that participants who had been given an unsolicited drink by another participant subsequently bought considerably more raffle tickets from him, and notably that the effect did not depend on whether they liked him.

The part worth sitting with is that last detail. Reciprocity operated independently of liking, which means the mechanism is not charm, it is the debt. Giving something of real value creates a small obligation, and obligations accumulate quietly across a relationship.

Which is the evidence behind giving more than you ask, and it also explains why the fake version fails. A gift that costs you nothing and serves you obviously, such as a discount that requires a purchase or a guide that is a brochure, does not register as a benefit received and creates no obligation at all. The debt only exists if the value was real.

The filter

Why people filter you out instead of unsubscribing

Every mechanism a person uses to protect their attention is cheaper than unsubscribing, and the whole of this section follows from that one fact.

• Effort asymmetry. Deleting is a single gesture, often a thumb swipe. Unsubscribing is a decision, then a page load, then a click, then sometimes a confirmation and a preference screen. Behaviour follows the path of least resistance and the path of least resistance is to do nothing twice a week forever.

• Status quo bias. Samuelson and Zeckhauser demonstrated in 1988 that people show a pronounced preference for the current state of affairs across a range of decision scenarios, and the finding has been replicated since. Staying subscribed is the current state, and changing it requires a reason strong enough to overcome the bias.

• Option value. People keep subscriptions they never read because unsubscribing forecloses a discount they might want one day, which makes staying and ignoring the rational choice rather than a lazy one.

• Deferral rather than refusal. Almost nobody decides never. They decide not now, and not now repeats indefinitely because nothing ever forces the decision.

• Mental filtering, which costs nothing at all. Once a sender has been categorised as skippable, the filtering happens before conscious attention engages. No irritation, no verdict, no moment where a decision could have gone the other way.

And the reason a hard unsubscribe backfires twice

Brehm described psychological reactance in 1966: when people perceive a freedom being restricted, they experience a motivational push to reassert it, and often act more strongly than they otherwise would have.

Somebody who wanted to leave quietly and finds the exit obstructed is no longer mildly disengaged, they have had a freedom threatened, and the response is sharper than the feeling that preceded it. A dark pattern on your preference centre does not retain a subscriber, it converts a passive departure into an active one and hands them the most damaging button on the screen.

When the complaint becomes the cheaper action

The exception to all of the above, and the reason a hidden unsubscribe link is a false economy. On a phone, the junk control sits within very easy reach while your unsubscribe link sits at the bottom of a long email behind a page load. When the complaint becomes the lower effort option, people take it, and that choice costs you reputation rather than one subscriber.

So make unsubscribing easy, not as an act of generosity but because you want the verdict. An unsubscribe is information. Filtering is not.

What to do

What all of this changes about how you send

• Hold identity constant and vary everything else. Same sender name, same voice, same promise, forever. Different shape, format, length, ask and value. Consistency in who you are is what mere exposure rewards; variation in what you send is what slows habituation.

• Use rest as a deliberate tool. Spontaneous recovery means a habituated group regains responsiveness when you stop for a period, and recovers faster where the habituation was caused by high frequency. Reducing sends to a quiet segment is an intervention, not a surrender.

• Stop resending to non-openers by default. Stimulus generalisation says a near-identical email a few days later is not a new stimulus, it is the same one again, so the practice is a habituation machine with a respectable name.

• Treat novelty as a depreciating asset. Every format trick loses power with repetition, so spend them deliberately rather than making them your regular voice.

• Change the shape, not only the words. A different length, a different structure, a plain text send, an email that asks nothing at all. Shape is a bigger variable than copy and it is the one people leave alone.

• Make the unsubscribe obvious. You want people who are done to leave through the door that gives you information and costs you nothing, rather than the one that damages your reputation.

• Write to the trigger rather than the product. Category entry points are built by connecting your brand to the situations that create need, and situations are far more interesting to write about than features.

• Trust behaviour over what people tell you. Sheeran's 2002 meta-analysis of the intention-behaviour gap found that stated intentions explain only a modest share of what people go on to do. Preference centres and surveys are useful, and they are weaker evidence than what somebody clicked.

• Measure decay rather than performance. Filtering leaves no trace in per-send metrics, so watch how each joining cohort behaves over its lifetime. A cohort that declines steadily while your campaign averages hold steady is the signature of habituation hiding behind new subscribers.

Where I would leave it

The instinct when email stops working is to make the emails better, and it is a reasonable instinct that happens to be aimed at the wrong target. Better copy, better design and better offers all operate at the level of the individual message, while the thing suppressing your results operates at the level of the pattern, and it has been learning that pattern for years.

Your competition is not the other brands in the inbox. Your competition is a filtering system that is free, automatic, faster than conscious thought, impossible to see in your reporting and already extremely good at its job. You will not out-write it. Nobody out-writes it.

What you can do is be a sender whose emails sometimes matter, because consequence is the one thing the mechanism has no defence against. Everything else in this post is a variation on that single idea.

Sources: Thompson & Spencer (1966) and Rankin et al. (2009), Habituation Revisited, Neurobiology of Learning and Memory · Zajonc (1968), Attitudinal Effects of Mere Exposure, Journal of Personality and Social Psychology · Bornstein & D'Agostino (1994) on processing fluency · Benway & Lane (1998), Banner Blindness, Rice University · Simons & Chabris (1999), sustained inattentional blindness · Samuelson & Zeckhauser (1988), status quo bias · Dawes (2021), Ehrenberg-Bass Institute, the 95:5 rule · Romaniuk & Sharp (2016), How Brands Grow Part 2, on category entry points · Folk, Remington & Johnston (1992) on contingent attentional capture · Rogers, Kuiper & Kirker (1977) and Symons & Johnson (1997) on the self-reference effect · Ebbinghaus (1885) on the forgetting curve and the spacing effect · Baumeister et al. (2001), Bad Is Stronger Than Good · Gouldner (1960) and Regan (1971) on reciprocity · Brehm (1966) on psychological reactance · Sheeran (2002) on the intention-behaviour gap.