The unsubscribe you should be glad about

How often should you email your list? It is the first question anyone asks about email, and everyone wants a number back — weekly, twice a month, daily if you are brave. The number is almost never what decides whether a list survives. A product can send every weekday for a year and end with a bigger, warmer list than it started with. Another can send four times in a quarter and lose a third of its subscribers. Frequency is the variable everyone argues about because it is the one that is easy to see. It is not the one doing the damage.

What actually burns a list is the share of your sends that were worth opening. Every email spends a small amount of goodwill up front — a person stops, reads a subject line, decides. If what follows is useful, the goodwill comes back with interest and the next one gets opened faster. If it is filler written because it was Tuesday and Tuesday is send day, the goodwill is simply gone, and the cost is charged not to that send but to the next three. High frequency with a high hit rate is fine. Low frequency with a low hit rate still ends badly, only slower.

The arithmetic is worth doing once, with made-up but plausible numbers, because it changes how the question feels. Take five thousand subscribers and a weekly send: fifty-two emails a year. Suppose a good send costs you two unsubscribes per thousand — ten people, the ordinary price of appearing in an inbox. Over a year that is roughly five hundred people leaving, ten percent attrition, comfortably covered by even modest growth. Now suppose a quarter of your sends are filler and filler triples the rate. Those thirteen sends alone shed close to four hundred more. Same cadence, same list, and the difference between healthy and shrinking sat entirely in whether you had something to say.

Consent is the other half, and it is the half people underestimate because it happens once, at signup, and is never revisited. Whatever you said you would send is the deal. Someone who subscribed for a monthly product update and starts receiving three promotions a week did not get more email — they got different email than the one they agreed to, which is the thing that actually feels like spam. Changing the deal quietly is worse than changing it loudly. If the cadence or the content is going to shift, the honest move is to say so in a send and let people re-decide, and to accept that some of them will leave right then.

an unsubscribe is not churn. it is a person doing your delivery a favor and telling you the truth for free.

Which brings us to the exit, and to the reframe worth internalizing: the unsubscribe is the good outcome. The bad outcome is the spam button, and the two are substitutes — someone who cannot find the exit will reach for the complaint instead. That trade is brutally asymmetric. An unsubscribe removes one person who was never going to buy. A complaint is a signal to the mailbox provider about your sender reputation, and providers publish thresholds on the order of one complaint per thousand messages before delivery starts degrading for everyone else on your list. One person's frustration, priced across five thousand inboxes. Make the exit obvious, make it one click, and do not put it in six-point gray at the bottom out of some instinct to hide it. You want that door to be easy.

The practical move that lowers both numbers at once is to stop sending everything to everybody. Of five thousand subscribers, suppose twelve hundred have opened something in the last ninety days. Those twelve hundred are your actual list; the rest is history. Sending the weekly to all five thousand does not reach more people in any meaningful sense — the other thirty-eight hundred mostly contribute delivery risk, and a few of them contribute complaints. Send the regular cadence to the engaged group, drop the dormant group to something rare, and after a last-chance email that says plainly that you will stop, stop. A smaller list you can trust beats a big one you are afraid to mail, for the same reason an empty chart you believe beats a full one you do not.

A cadence should also be allowed to lose an argument with the day it lands on. A send with nothing in it should not go out just because the calendar says so, and neither should one aimed at an audience that is empty after filters and removals, or one pointing at a page that is currently broken. We have written before about a scheduler whose most valuable ability is declining to fire — the same logic applies to your own hand on the send button. A skipped week is invisible to subscribers. A hollow week is not, and it is charged forward.

This is roughly the shape we build toward in SiteOps: the system drafts the email, checks it against real numbers rather than invented ones, and holds it for your approval instead of shipping on its own confidence. Removals are respected per person, so an old exported list cannot quietly resurrect someone who left. None of that decides what you have to say — that part is still yours — but it removes the two ways a small team usually burns a list by accident, which are sending to people who already left and sending because it was scheduled.

Sustainability is the last piece and the least glamorous. The best cadence is the one you can still run in month eight, in a bad week, with thirty minutes to spare. Monthly kept is worth more than weekly abandoned, because a list that hears from you rarely but reliably stays warm, while a list that goes quiet for four months is genuinely hard to restart — reappearing in an inbox you have not visited since spring reads as a mistake. Pick the interval you would keep on your worst week and let good weeks be a bonus.

So the rule of thumb we would hand you: before any send, write in one plain sentence what the reader gets out of it. If you can write that sentence, send it, whatever the calendar says. If you cannot, you do not have an email — you have a slot. Skip it, and spend the hour on the next one that earns its place.