Email Marketing

Email Frequency and Unsubscribes: What the Data Actually Shows

The best evidence on how often to email comes from one randomized experiment, and it does not say "send less". It says fewer emails cut unsubscribes by 59% and cost 7.6% of daily revenue. Here is what that, the vendor benchmarks and the law actually support.

oToK Team15 min read
Email Frequency and Unsubscribes: What the Data Actually Shows

Not legal advice: The experiment below ran at one large online retailer and is a working paper we could not confirm as peer-reviewed. The benchmark figures are vendor-reported and the survey figures are self-reported; none is an audited sample of Israeli small businesses. The legal points describe the text of the statute, not legal advice: for a specific campaign, ask a lawyer. Worked arithmetic uses invented rates and is labelled as such.

Search for how often to send marketing emails and you will meet the same few claims: once a week is the sweet spot, daily emails drive people away, and most people who unsubscribe say you simply sent too much. They are repeated as settled fact. Most of them come from sources that cannot show cause and effect: vendor dashboards that compare different senders, or surveys that ask people to explain their own behaviour afterwards.

We went to what can: a large randomized experiment that changed how often the same kind of customer was emailed (Baek, Chen, Ma and Mitrofanov, run at a North American online retailer), an older single-company model of the profit-maximising number of emails (Zhang, Kumar and Cosguner, Journal of Marketing Research, 2017), a 42,000-account vendor dataset, two consumer surveys, the Gmail and Yahoo sender rules, and the text of section 30א of Israel's Communications Law. The short version: more emails do cause more unsubscribes, but fewer emails also cost sales, the lifetime-value result depends on who you are emailing, vendor benchmarks across senders point the other way for reasons they cannot explain, and the thing you can rely on is your own list, measured per subscriber and not per send.

Key takeaways

  • The best causal evidence we found is a randomized experiment: 603,153 customers of one North American online retailer, one week in October 2024, daily coupon emails against one every other day. Every-other-day customers were 59.2% less likely to unsubscribe, and bought less: 7.6% lower daily revenue, 4.8% lower purchase probability, 10.0% fewer items. Price per item did not change.
  • The absolute numbers are small. The authors say the daily-email unsubscribe rate was "a few tenths of a percent per week". By our arithmetic from their coefficient, halving the frequency saved roughly one to two unsubscribes per 1,000 customers per week. It matters because it compounds and because of what an unsubscribe costs, not because any one week looks dramatic.
  • An unsubscribe is a strong signal, but part of it is a symptom. In the same retailer's 2024 history, monthly spending fell about 36% after a customer unsubscribed (instrumental-variable estimate; 44% in a simple before-and-after comparison), and the likelihood of buying in the next six months was more than 60% lower. The authors stress that unsubscribing is both a cause and a symptom, so do not multiply the drop by every unsubscribe you avoid.
  • Cutting frequency for everyone was roughly a wash on modelled lifetime value (+0.2%). Personalising frequency was not: +6.8% for a revenue-first policy and +8.3% for a retention-aware one, and newer customers did better on every-other-day while older ones did better on daily. These are model outputs built on the experiment, not further experiments.
  • Vendor data across senders does not show daily senders unsubscribing more. In MailerLite's 2025 data (more than 42,000 accounts), unsubscribe rates were 0.87% for senders mailing less than once a month, 0.38% weekly, 0.33% twice a week and 0.36% daily. Those are different senders with different lists, so the numbers describe who sends how often, not what frequency does to a list.
  • There is no universal best number. An older model on one home-improvement retailer found about 7 emails a month optimal (a range of 5 to 14 by customer), with 4 a month costing an estimated 32% of lifetime profit and 10 a month 16%. Too few was worse than too many there.
  • Two sets of rules carry real weight. Gmail's guidelines say to keep the spam rate below 0.1% and never reach 0.3%, and require one-click unsubscribe on marketing mail from bulk senders (about 5,000 or more messages a day); Google recommends honouring unsubscribes within 48 hours and Yahoo asks bulk senders to do so within 2 days. Israel's section 30א lets a recipient refuse advertising at any time, in writing or in the way it was sent, and sending after a refusal notice is a case where the advertiser has no defence against being presumed to have acted knowingly.

Does sending more emails cause more unsubscribes?

Yes, and this is the best test we found. Jackie Baek, Daniel Chen, Will Ma and Dmitry Mitrofanov (New York University, Boston College and Columbia) worked with a major North American online retailer. From 25 to 31 October 2024, 603,153 customers were randomly assigned to one of two groups: the control group received the retailer's usual daily coupon email, and the treatment group received one every two days. The paper is titled "Balancing Customer Engagement and Annoyance in Online Retail: Insights from a Field Experiment" and is posted on SSRN as a working paper. We could not confirm that it has been peer-reviewed.

Over the week, customers on the every-other-day schedule were 59.2% less likely to unsubscribe from the list. They also bought less. The probability of buying on a given day was 4.8% lower, the number of items bought was 10.0% lower and daily revenue per customer was 7.6% lower. The average price per item did not change, so the lost revenue came from fewer purchase occasions and slightly smaller baskets. The authors' own summary is that there is no free lunch: you cannot cut emails without losing some short-run sales, and you cannot push frequency to the maximum without wearing down people's willingness to stay subscribed.

Two cautions apply. The measure is unsubscribing, which the authors treat as a proxy for churn, and the test ran for one week, so it cannot show how the effect builds. And the emails were coupons whose codes were valid for a day or two, which is exactly the kind of message where an extra send can create an extra purchase. A weekly newsletter or a long-sales-cycle service is not the same case.

59.2%

lower chance of unsubscribing for customers emailed every other day instead of daily, in a one-week randomized test of 603,153 customers at one online retailer. The paper describes the daily-email unsubscribe rate as "a few tenths of a percent per week".

Source: Baek, Chen, Ma and Mitrofanov (SSRN working paper), Table 2

Every other day instead of daily: how much lower each outcome was
Probability of unsubscribing
59.2%
Items purchased
10%
Daily revenue per customer
7.6%
Probability of a purchase on a given day
4.8%
Source: Baek, Chen, Ma and Mitrofanov, "Balancing Customer Engagement and Annoyance in Online Retail" (SSRN working paper), randomized experiment on 603,153 customers, 25 to 31 October 2024. Relative differences against the daily-email group.

How big is that in real terms? The paper's coefficient is a fall of 0.0002 in the daily probability of unsubscribing, which is about 0.14 percentage points over a week: roughly 1 to 2 fewer unsubscribes per 1,000 customers per week. That is our arithmetic from a rounded coefficient, not a figure the paper states. The sales effects are statistically significant at the 5% level but imprecise (the revenue estimate comes with a wide margin), while the unsubscribe effect is much tighter.

What does an unsubscribe actually cost?

Quite a lot, but less than the headline suggests. The same team analysed a year of that retailer's records for about 200,000 customers in 2024. Customers received an average of 0.94 coupon emails a day and opened 23% of them, and 15.6% unsubscribed at some point during the year. After unsubscribing, a customer's daily revenue was 59% lower than that of customers who never unsubscribed, and 75% lower than the same customer's own spending before.

For monthly revenue the authors report two numbers. In a simple before-and-after comparison, a typical customer spending $22.47 a month spent $9.85 less after unsubscribing, a 44% drop. Their preferred instrumental-variable estimate, which tries to separate cause from selection by using the previous month's email frequency, puts it at 36%. They also found the likelihood of making any purchase over the six months after unsubscribing was more than 60% lower than for otherwise similar customers who stayed subscribed.

The authors are explicit that unsubscribing can be both a cause and a symptom: some customers were already drifting away before they clicked. So 36% is not the revenue you keep for every unsubscribe you prevent. It tells you that an opt-out marks someone who has mostly stopped buying, and that watching it by cohort is an early-warning signal.

Measure (one retailer, 2024)FigureHow to read it
Customers who unsubscribed at some point in the year15.6%Descriptive: share of about 200,000 customers
Daily revenue after unsubscribing, against customers who never unsubscribed59% lowerComparison of different people, so selection is built in
Daily revenue after unsubscribing, against the same customer before75% lowerBefore and after, so part of it is a drift that was already under way
Monthly revenue, simple before-and-after44% lower ($22.47 to $12.62)Same caveat
Monthly revenue, instrumental-variable estimate36% lowerThe authors' preferred figure; designed to reduce selection, not to remove it
Chance of buying in the six months after unsubscribingMore than 60% lowerStrong leading indicator of churn, cause and symptom both

Two things follow for any business. Count unsubscribes by cohort and segment, because a rising rate in new subscribers means something different from a rising rate in your oldest ones. And judge a frequency change on revenue per subscriber over months, not on the revenue of one send.

So does emailing less often win on lifetime value?

Not for everyone. The authors fed the experiment and the retailer's history into a customer lifetime value (CLV) model. If every customer moved to every other day, modelled CLV rose from $213.03 to $213.56, which is +0.2%: the retention gain roughly offset the lost sales. Personalising frequency by customer did much better. A policy that maximised short-run revenue lifted average CLV by 6.8% ($227.46), and one that also counted churn lifted it by 8.3% ($230.78) while sending fewer emails than the revenue-first policy (0.52 a day against 0.68). When coupon use was capped at the status-quo level, the gains were smaller, +1.8% and +4.5%.

The segment results are the practical part. Splitting customers at the median tenure (2,583 days, about seven years), the newer half did better on every other day ($248.30 against $241.80) and the older half did better on daily ($178.90 against $184.25). The same pattern appeared in the experiment itself: the cut in unsubscribes was larger for newer customers, while higher-volume customers showed little retention benefit and the biggest revenue losses. The paper's reading is that newer subscribers are more fragile, and established ones tolerate frequency and also respond to it.

The payoff also depends on how loyal your customers are. In the authors' sensitivity test, a uniform switch to every other day lowered modelled CLV (a ratio of 0.98 to the status quo) when baseline annual retention was set at 10%, and raised it by about 50% when it was set at 90%. The more repeat business a company has, the more an unsubscribe costs, which is why a one-time-purchase business and a membership should not copy each other.

Modelled customer lifetime value by tenure: daily against every other dayDaily emailsEvery other day

Newer customers

Daily emails: $241.80
Every other day: $248.30

Older customers

Daily emails: $184.25
Every other day: $178.90
Source: Baek, Chen, Ma and Mitrofanov, Table 12: model outputs, split at median tenure of 2,583 days. Dollars per customer; these are projections from the experiment and the retailer's history, not measured results.

These are model outputs for one retailer, resting on assumptions about how unsubscribing links to later spending. They support an idea (treat newer and older subscribers differently) and not a number to copy.

Is there an optimal number of emails per month?

One published model says yes, for one company, and it points at a middle. Xi (Alan) Zhang, V. Kumar and Koray Cosguner (Journal of Marketing Research, vol. 54, no. 6, 2017) modelled email opens and purchases for a U.S. home-improvement retailer. They found that the number of emails has a non-linear effect on profit: too few or too many both reduce it. Their optimum averaged about 7 emails a month, ranging from 5 to 14 depending on each customer's history, and they estimated that sending 4 a month instead of the best number could cost about 32% of lifetime profit, and 10 a month about 16%. They conclude that too few emails was the worse mistake. The authors caution that the results come from a single firm.

A third study asks a neighbouring question. Wang, Lu and Shi (Manufacturing & Service Operations Management, 2022) ran a field experiment at a large U.S. car-wash chain that sells tiered subscriptions. Two engagement emails over a month raised retention by 7.4% five months after the start and cut the odds of churn by 26.3%, but the effect faded and showed fatigue after the second email, and profit rose for mid-level infrequent users and top-level subscribers and fell for mid-level frequent users and basic-level subscribers, because engaged subscribers also use more of the service. Their advice is to email only the subscribers for whom it pays.

StudySettingWhat it found
Baek, Chen, Ma and Mitrofanov (SSRN working paper)Online retailer, randomized, 603,153 customers, one weekEvery other day instead of daily: 59.2% fewer unsubscribes, 7.6% lower daily revenue; effect depends on tenure and engagement
Zhang, Kumar and Cosguner (JMR, 2017)U.S. home-improvement retailer, statistical model, a single firmAbout 7 a month optimal (5 to 14 by customer); 4 a month costs an estimated 32% of lifetime profit, 10 a month 16%
Wang, Lu and Shi (MSOM, 2022)U.S. car-wash subscriptions, field experimentTwo engagement emails raised retention but faded after the second; profit effect depends on subscriber tier and how often they already use the service
Estimated lifetime profit lost against the best frequency (one retailer, model)
Sending 4 emails a month
32%
Sending 10 emails a month
16%
Source: Zhang, Kumar and Cosguner, Journal of Marketing Research 54(6), 2017. The model-based optimum is about 7 emails a month; figures are the authors' illustrative estimates for a single firm.

Read the three together: neither the experiment nor the model says that less is always better. Both say frequency is a trade between sales now and a list that still listens later, and that the right point depends on the customer.

Why don't vendor benchmarks show daily senders unsubscribing more?

Because they compare different senders, not the same subscribers under different schedules. MailerLite, an email-marketing platform, analysed more than 42,000 accounts that sent in 2025 (over 1.4 million campaigns and 12 billion emails), grouped by how often they sent campaigns. Its unsubscribe rates (the page does not define the calculation) were 0.87% for senders mailing less than once a month, 0.54% for one to three a month, 0.38% weekly, 0.33% twice a week and 0.36% daily. Open rates fell slightly with frequency (35.11% to 30.04%) while click rates rose from 3.74% to 5.31% at twice a week and eased to 4.97% daily.

The vendor itself says it does not know why the rarest senders do worst; it guesses that recipients forget they signed up or no longer need the content. There are other candidates. Businesses that send daily tend to be ones whose subscribers expect it (a deals list, a news-style newsletter), subscribers who dislike daily mail leave early and the remaining list is a tolerant one, and a rate per campaign is not a rate per subscriber, so a list receiving 200 emails a year is not comparable to one receiving 12. The page also excludes automations, and it warns that open rates are less accurate since Apple's Mail Privacy Protection. None of it contradicts the experiment, because the experiment held the customers constant and changed only the schedule.

Benchmarks also move with how easy it is to leave. MailerLite's separate 2026 benchmark report (2025 data, 3.6 million campaigns from 181,000 accounts) gives an average unsubscribe rate of 0.22%, more than double 2024's 0.08%, and the report says the rise is likely due to Gmail changes that make it easier to unsubscribe from promotional emails. A higher number can therefore mean people can finally leave, not that they are more annoyed.

Unsubscribe rate by how often the sender mails (different senders)
Less than once a month
0.87%
1 to 3 a month
0.54%
Once a week
0.38%
Daily
0.36%
Twice a week
0.33%
Source: MailerLite, email cadence and frequency report, 2025 data from more than 42,000 accounts (campaigns only, automations excluded). Different senders and lists, so not causal.

The takeaway is the opposite of "daily is safe". A sender who emails rarely and then suddenly is the one people forget. The benchmark is a reason to be regular, not permission to be frequent.

Why a per-send rate hides the cost of frequency

A rate per campaign looks small, and so it feels harmless to add one more send. But each subscriber faces every send, and the losses accumulate. As illustrative arithmetic, assume a constant, invented 0.2% of the remaining list unsubscribing on every send. At about 12 sends a year you lose roughly 2.4% of the list; at 48 (weekly) 9.2%; at 96 (twice a week) 17.5%; and at 192 (nearly four a week) 31.9%. The formula is one minus (1 minus the per-send rate) to the power of the number of sends.

Real lists do not behave at a constant rate. The people who dislike frequency leave first, the remaining ones are more tolerant, and in the retailer's experiment unsubscribes fell by 59% while emails fell by about half, so the per-email rate also fell. What the arithmetic shows is only why the per-send figure on a dashboard is the wrong number to manage: track unsubscribes per subscriber over a month or a quarter.

Illustration: share of a list gone after a year at an invented 0.2% per send
About 4 a week (192 sends)
31.9%
About 2 a week (96 sends)
17.5%
Weekly (48 sends)
9.2%
Monthly (12 sends)
2.4%
Source: Our arithmetic: 1 - (1 - 0.002)^n. The 0.2% rate is invented and assumed constant; it is not a benchmark and not data from any study.

What do people say about email frequency?

In surveys, too many emails is consistently near the top of the reasons given, and the size of the number depends on how the question is asked. Litmus, from Validity, commissioned Dynata to survey 1,000 U.S. adults in April 2025: 67% named emails sent too frequently as a reason for unsubscribing from a retailer's emails, 46% irrelevant or non-personalised content and 42% excessive promotions. Asked how often they would like to hear from a brand, 32% said weekly and 20% two to three times a week. Mailgun's report "Email and the customer experience" (the page does not give its sample size or date) lists three primary reasons: too many emails from the brand (19.8%), no longer interested in the offerings (17.9%) and irrelevant content (17.3%).

The gap from 67% to 19.8% probably reflects differences in question format, population and wording; neither page lets us pin it down. The Litmus figures add up to 155%, so respondents evidently could give several reasons. Mailgun's three add up to 55%, and the page appears to show only its top three. Neither tells you what frequency would keep a particular list. People also explain their behaviour badly after the fact, which is why the randomized test is worth more than any survey percentage.

One company's own report is a useful sanity check. In July 2015 LinkedIn announced that it was sending fewer emails, "for every 10 emails we used to send, we've removed 4", and said member complaints had dropped by half. These are the company's claims, with no supporting data, and members could still adjust which emails they got and how often.

SourceSampleFigure
Litmus (Validity) with Dynata, April 20251,000 U.S. adults, designed to represent U.S. adults67% give too-frequent emails as a reason to unsubscribe (46% irrelevant content, 42% too many promotions); respondents evidently could give several reasons
Mailgun, Email and the customer experienceNot stated on the page we read19.8% too many emails (17.9% no longer interested, 17.3% irrelevant); the three sum to 55% and appear to be only the top three
LinkedIn, July 2015Company's own announcement40% fewer emails; complaints "cut in half"; no data supplied

Use surveys to understand what annoys people, and the experiment to learn what a schedule does. The two measure different things.

What do Gmail and Yahoo require, and does frequency matter to them?

Neither publishes a rule on how often you may send. What they publish are complaint-rate and unsubscribe rules, and frequency matters to them only through the complaints it creates. We found no public dataset that links send frequency to spam-complaint rate, so we cannot put a number on that link.

RequirementGmailYahoo
Who it applies toBulk senders: close to 5,000 or more messages a day to personal Gmail accounts, counted across a primary domain; once classified, always a bulk sender. Requirements started in February 2024Bulk-sender rules (the page does not define a threshold) and a separate list for all senders; started February 2024 and were rolled out gradually
Spam-complaint rateKeep it below 0.1% and prevent it from ever reaching 0.3% or higher (0.3% applies to all senders); above 0.3% a sender is ineligible for mitigation until it stays under 0.3% for 7 consecutive daysKeep it below 0.3% (all senders); calculated against mail delivered to the inbox
One-click unsubscribeRequired for marketing and promotional mail, not transactionalRequired for bulk senders for marketing and subscribed messages; the POST method (RFC 8058) is highly recommended
Speed of honouring itGoogle recommends fulfilling unsubscribe requests within 48 hours (FAQ wording; its requirements table also lists failing to do so as a compliance issue)Honour unsubscribes within 2 days (bulk senders)

Gmail's page says enforcement on non-compliant traffic began ramping up in November 2025. A small business that sends a few thousand messages a day in total is below the bulk threshold, but a complaint rate still counts toward reputation, and the one-click header is cheap to add.

What does Israeli law say about stopping the emails?

The Hebrew text of section 30א of the Communications (Telecommunications and Broadcasting) Law, 1982 does not set a frequency limit, but it governs the other half of this article: what must happen when someone wants fewer emails or none. We read the statute text (a copy that records amendments up to תשפ״ב) and paraphrase it here. This is not legal advice, and our article on Israel's spam law covers consent in full.

ProvisionWhat it says (our paraphrase)What it means for your send schedule
s.30א(ד)(1)A recipient who agreed, or did not refuse, may at any time give the advertiser a refusal notice, covering advertising in general or of a particular type, and withdraw consent. It costs nothing beyond the cost of sending the message, and the recipient may give it in writing or in the way the advertisement was sent"Fewer of these" is a legitimate refusal: it can be narrow (a type of message) or total. A reply to the email counts, so a monitored reply address or a working link matters
s.30א(ה)(1)(ג)Each advertisement must state the recipient's right to refuse at any time and give a simple, reasonable way to do it; for an electronic message, a valid internet address for the noticeEvery marketing email needs a working opt-out, in the message
s.30א(ד)(2)In an ongoing purchase relationship, the recipient is treated as having refused when the relationship ends, including by valid cancellationA customer who cancels should leave the marketing list on the date they leave, unless they opt in again
s.30א(י)(1) and (5)(א)For advertising knowingly sent in breach, a court may award compensation without proof of damage, up to NIS 1,000 per advertisement received; a business is presumed to have acted knowingly, and sending after a refusal notice is one of the cases where it has no defenceAfter an opt-out, the next send is where the exposure sits. Suppress immediately, across every list and automation that sends the same kind of message

The statute gives no number of days for acting on a refusal, so Yahoo's 2-day rule (for bulk senders) and Google's 48-hour recommendation are the practical outer limit, and the statute points to immediate suppression. We found nothing in the statute about how often an agreeing recipient may be sent advertising.

How should a small business set its email frequency?

The evidence does not hand you a number, but it does hand you a method. Every step below is built from the sources above.

  1. 1Measure unsubscribes per subscriber, not per send. Divide opt-outs in a month by active subscribers, by cohort, so a change in frequency shows up as a change in list loss.
  2. 2Treat new subscribers more gently. The experiment's retention gains were largest for newer customers, and the modelled lifetime value favoured a lighter touch for them and daily for established, responsive ones. A welcome sequence followed by a slower rhythm is a reasonable default.
  3. 3Test with a holdout, but size it honestly. The base rates are tiny, so a frequency test on a small list will not show a difference (see the table below). If you cannot reach the sample size, make a bigger change and watch a longer window rather than a modest change over a week.
  4. 4Watch revenue per subscriber, not revenue per send. Halving the schedule cost the retailer 7.6% of daily revenue, and a send-level report shows that loss at once while the retention gain arrives later.
  5. 5Do not drift into sending too rarely. Both the vendor data and the single-firm model suggest it hurts: senders mailing less than once a month had the highest unsubscribe rates, and the model's too-few penalty was larger than the too-many one. MailerLite guesses that irregular timing is part of it but does not test that.
  6. 6Stop mailing people who have gone quiet, judging by clicks and not opens. Apple's Mail Privacy Protection makes opens unreliable, and a last re-engagement email gives people a chance to say yes before you stop.
  7. 7Offer a narrower exit before a full one. The law explicitly allows refusal of one type of message. We found no controlled study showing that preference centres reduce total unsubscribes, so treat it as a courtesy and a legal convenience and measure it yourself.
Baseline unsubscribe rate in the test windowDrop you want to detectSubscribers needed in each group
0.25% a week (the experiment's scale)59% (what the experiment found)About 12,700
0.25% a week30%About 59,200
1% over the test window59%About 3,200
1% over the test window30%About 14,700

Our calculation: standard two-proportion test, two-sided 5% significance, 80% power. The 0.25% is our rounding of the paper's "a few tenths of a percent per week" and the 1% is illustrative. A list of a few thousand cannot prove a modest effect, which is a reason to rely on the larger published tests and on your own trend over months.

How does oToK help with email frequency?

oToK is a CRM with email marketing included on every plan, so the unsubscribe, the preference choice and the inactivity rule sit on the same contact record as the rest of the customer's history. What it does not do is pick your frequency: there is no setting that tells you how often to send, and we would not trust one.

  • Every marketing email carries an unsubscribe link and the one-click unsubscribe header that Gmail and Yahoo ask for, and a marketing email cannot be sent without a working opt-out. Transactional email is exempt from both, as the sender rules intend
  • A one-click unsubscribe is recorded straight away, and unsubscribed contacts are skipped automatically on every future send
  • A public preference page lets a contact opt out of specific topics instead of leaving altogether. Topics are yours to define: campaigns and automation emails that you assign to a topic skip contacts who opted out of it, and every newsletter gets its own topic automatically
  • Every newsletter issue carries an "unsubscribe from this newsletter" link ahead of the full unsubscribe, so the narrow exit is the first one a tired reader reaches
  • An optional sunset policy in the email settings, which you switch on and give a re-engagement template, stops mail to people who have gone quiet. It counts clicks, not opens. It starts at 120 days with no click, when the subscriber gets one re-engagement email in the template you choose, and unsubscribes them if they still have not clicked 14 days after that email. Both numbers are editable
  • Campaign reports show opens, clicks and unsubscribes per send, and a deliverability dashboard tracks complaint and bounce rates, with an automatic pause when they climb too high
  • Audiences let you send a campaign to a subset, which is how you give a newer segment a lighter schedule than an established one

None of this raises your results by itself. It makes the opt-out work the way the law and the mailbox providers expect, and it makes list loss countable, so you can run the per-subscriber measurement above on your own list.

Share this article

Frequently asked questions

How often should I send marketing emails?
There is no universal answer. The best experiment we found (603,153 customers of one online retailer) showed that emailing every other day instead of daily cut unsubscribes by 59.2% and cost 7.6% of daily revenue, and a single-firm model put the best rate at about 7 emails a month, with too few worse than too many. Start with a regular rhythm you can sustain, lighter for new subscribers, and measure unsubscribes and revenue per subscriber over months.
Does sending more emails increase unsubscribes?
Yes, in the one randomized test we found. Customers who received daily coupon emails were more likely to unsubscribe than those who received one every other day (the every-other-day group had a 59.2% lower chance), but they also bought more. The absolute difference was small, roughly 1 to 2 unsubscribes per 1,000 customers per week by our arithmetic, and it came from one retailer and one week.
What is a normal email unsubscribe rate?
MailerLite's 2026 report puts the 2025 average at 0.22% of recipients, up from 0.08% in 2024, and says the rise is likely due partly to Gmail making unsubscribing easier. Its cadence data show 0.38% for weekly senders and 0.87% for senders who mail less than once a month. The two MailerLite datasets use different samples and the pages do not define the rate, so the figures are not directly comparable. They are vendor figures across different senders, so use your own previous quarter, counted the same way, as the benchmark.
Is it true that most people unsubscribe because of too many emails?
Surveys disagree. A Litmus and Dynata survey of 1,000 U.S. adults in April 2025 found 67% gave emails sent too frequently as a reason (respondents evidently could give several), while Mailgun's report found 19.8% named too many emails as a primary reason. Both are self-reported, and neither shows what a given frequency does to your list.
What spam-complaint rate do Gmail and Yahoo allow?
Gmail says to keep the spam rate below 0.1% and prevent it from ever reaching 0.3% or higher. Yahoo says to keep it below 0.3%, calculated against mail delivered to the inbox. Gmail also requires one-click unsubscribe on marketing mail for bulk senders (close to 5,000 messages a day or more) and asks that unsubscribes be fulfilled within 48 hours; Yahoo asks bulk senders to honour them within 2 days.
How quickly must I stop emailing someone in Israel after they unsubscribe?
Section 30א of the Communications Law lets a recipient refuse advertising at any time and gives no number of days, but it lists sending after a refusal notice as a case with no defence against a presumption that the breach was knowing, and a court may award up to NIS 1,000 per advertisement without proof of damage. In practice, suppress immediately and treat Google's 48-hour recommendation and Yahoo's 2-day rule as the outer limit. This is not legal advice.
Should new subscribers get fewer emails than long-time customers?
The one experiment we found suggests so. Reducing frequency cut unsubscribes more among newer customers, and established customers gained less from it; the largest revenue losses were among high-volume buyers. The lifetime-value model favoured every other day for newer customers and daily for older ones. It is one retailer's coupon list, so test it on yours.
How big does my list need to be to test email frequency?
Larger than most small lists. Our calculation (two-sided 5% significance, 80% power) shows that detecting a 59% drop in a 0.25% weekly unsubscribe rate needs about 12,700 subscribers per group, and a 30% drop about 59,200. If your list is smaller, make a bigger change, watch a longer window and look at trends rather than a single comparison.
Should I remove subscribers who never open or click?
Usually, after one last attempt. Judge by clicks, since Apple's Mail Privacy Protection makes opens less reliable, send a final re-engagement email, and then stop mailing people who still do not respond. We found no controlled test of this step, so treat it as sound list hygiene and measure the effect on your own list.
Does oToK handle unsubscribes and inactive subscribers automatically?
Yes on unsubscribes, on every plan: every marketing email carries an unsubscribe link and a one-click header, unsubscribed contacts are skipped on future sends, and a preference page lets contacts opt out of specific topics. Inactive-subscriber clean-up is an optional sunset policy: turn it on in the email settings, pick a re-engagement template, and it sends one email to subscribers with no click for 120 days (the default) and unsubscribes those who still have not clicked 14 days later (also the default). oToK does not choose your sending frequency for you.

Ready to upgrade your customer communication?

Get started with oToK — connecting a WhatsApp number takes minutes.

No credit card required — start right away