It is the question that comes up in every call, usually right after installation: "how many pushes can I send per day?". Whoever asks is hoping for a number. Three, five, one. A number to write on a sticky note and stick to.
That number does not exist, and anyone who gives you one without knowing your site is selling you a certainty they do not have. What does exist is a way to find yours, and a set of brakes that let you send quite a lot without readers leaving. This article covers both: why there is no magic number, what really makes people unsubscribe, how frequency changes depending on the kind of site, which tools to use and how to tell from the numbers whether you are overdoing it.
Why there is no magic number
Picture two readers. The first subscribed to the pushes of a local news site because they want to know whether schools are closed tomorrow, whether there is a train strike, whether the road to work is blocked. The second subscribed to the pushes of a gardening blog because they liked a guide about roses.
To the first, six pushes on a day of severe weather feel like a service. To the second, two pushes in the same week feel like an invasion. The same number, two opposite reactions.
The right frequency depends on three things: how often you really have something new to say, how urgent it is, and what the reader expected when they pressed "Allow". A site that publishes fifty articles a day and one that publishes two a week cannot follow the same rule. Not even two newsrooms with the same volume, if one covers breaking crime and the other publishes in-depth economics.
There is a fourth variable few people consider: who receives the push. Sending five pushes to every subscriber is one thing. Sending five pushes, each only to the people who follow that topic, is another: each individual reader may get one or two. We will come back to this, because it is the most powerful lever you have.
What really makes people unsubscribe
Quantity matters, but almost never on its own. A reader rarely leaves because they got "one push too many". They leave because they got a push they did not need, then another, then another. The number becomes a problem when it multiplies the disappointment.
Relevance comes before quantity. Every push is a small interruption: the phone buzzes, the reader looks. If what they find is worth the interruption, next time they look again. If it is not, they learn that your notifications are noise, and noise gets switched off. Ten relevant pushes are less tiring than three useless ones.
Off-topic pushes. Someone who subscribed for school news and gets transfer market gossip does not think "what a content-rich site". They think "this is not for me". It is the most common mistake on sites that cover several topics, and also the easiest to fix, as we will see.
The wrong times. A push at six in the morning for an article that could have waited until nine is not just useless: it is rude. The same goes for late at night. The reader does not distinguish between "the automatic feed fired on its own" and "these people wake me up": they only see the result.
Clickbait titles. Promising one thing and delivering another works once. The second time the reader does not click, the third time they unsubscribe. We covered this at length in how to write push titles people click: an honest title is also a brake on unsubscribes.
Bursts. Three pushes in ten minutes, even if all relevant, feel like being buried. It is much better to space them out, or merge them: if three articles come out on the same story, one push to the most complete one is worth more than three.
Notice one thing: none of these reasons is "too many pushes" in the strict sense. They are all ways of wasting the reader's attention. Quantity is the multiplier.
How it changes by type of site
Below are four typical profiles. They are not rules: they are reasonable starting points, to be corrected by looking at your own numbers.
The newsroom with breaking news
This is the case where you send the most, and where you can send the most. A news reader subscribed precisely to be alerted, and news has an expiry date: a push that arrives three hours after the fact is worth little.
The risk here is not the number itself but the publishing volume: if dozens of articles go out each day, "new article equals push" is a bombardment. The strategy that works is layered: a few fixed pushes during the day to the latest news page, plus the truly important stories sent by hand, plus pushes by category to the people who follow them. The guide for publishers has a full example of how a newsroom organises it.
The niche site
Think of a site about schools, tax, jobs, a specific trade. It publishes less than a general news site, but its readers care a lot about a few topics. A deadline, a ranking, a call for applications: these are pushes the reader is glad to receive.
Here frequency can be high on busy days (a new decree, a list being published) and almost zero on quiet ones. The secret is to let the news calendar decide, not a fixed rhythm. And above all to split by category: someone who follows one topic does not want all the others.
The blog
A blog publishes little and its content does not expire. There is no urgency that justifies more than one push per article, and often even that one does not need to go out immediately: an article published at 11pm can wait until the next morning.
For a blog the opposite risk is more common: sending too little, so little that readers forget they subscribed and at the first push think "who are these people?". One push per new article, at a sensible time, is almost always the right rhythm.
The online shop
A shop does not have "news": it has deals, products back in stock, carts left halfway. Readers tolerate pushes that save them money or concern something they wanted, much less generic promotional ones.
Here frequency should stay low for promotions sent to everyone and can be freer for targeted pushes. A general promotion every single day wears people out quickly; a "the product you wanted is back" alert never does. We go into detail in push for e-commerce and, for WooCommerce users, in web push for WooCommerce.
Categories chosen by the reader: sending more without wearing people out
This is the lever that really changes the maths. If your site covers several topics, you can ask readers what they care about. In Natom you do it with the preferences panel: after subscribing, the reader sees the list of your categories and ticks the ones they want. Each category becomes an audience.
How it works is simple, and worth explaining properly:
- Readers who pick one or more topics receive the pushes for those topics, plus the ones you explicitly decide to send to all subscribers. They do not end up in another category just because they read an article from that section: their choice stays theirs.
- Readers who do not pick receive everything, as before.
The result is that you can send many more pushes in total without any single reader getting more. A newsroom with four sections can send eight pushes in a day, two per section, and someone who follows only one section sees two. For them, the site is perfectly calibrated.
A push can go to one or more audiences at once, up to five: an article about both "Teachers" and "Exams" reaches people who follow either, and people who follow both get it only once. If you use WordPress, the plugin already suggests the audiences of the article's categories when you prepare the push from the editor.
One caution: if you send only category pushes, readers who picked nothing do not get them, because they are in no audience. Important pushes, the ones that must reach everyone, should go to everyone.
Daily cap, quiet hours and drafts
If you connect an RSS feed with automatic sending, every new article can become a push without anyone pressing a button. It is convenient, and without brakes it is dangerous. That is why automatic feeds have three limits:
- Max pushes per day. Once the cap is reached, the feed stays silent for the rest of the day. The next day it starts again from the most recent article. It guarantees that a day of heavy publishing does not turn into a day of nonstop notifications.
- Quiet hours. For example from 22:00 to 07:00, in the feed's time zone. Nothing goes out in that window, and when it resumes, the latest article published goes out, not the whole backlog.
- Only these categories. The feed notifies only the sections you list and ignores the rest.
There is also a natural brake: on each check the automatic feed only looks at the newest article. Ten articles in fifteen minutes do not make ten pushes.
If instead you want to decide by hand, use the feed in draft mode. Every new article becomes a draft, already routed to its audience, waiting in the "To send" queue. The newsroom opens the queue, sends the ones that deserve it and deletes the rest. Drafts do not count toward the cap and never go out on their own. It is the best way to keep frequency under control when you publish a lot: the machine prepares, a person decides.
Recurring pushes as an appointment
There is a way of sending pushes that tires people very little, even when there are many: make them predictable. A push that arrives every day at the same time becomes an appointment, like the evening news. The reader knows that at 7 the stories to start the day arrive and at 6pm the ones to read in the evening, and they wait for them instead of putting up with them.
Recurring pushes are built for this: you write the notification once, choose the times, days and time zone, and it goes out on its own. For a newsroom, two or three recurring pushes a day to the latest news page cover most of the need without sending a push for every article.
Two tips. First, a different title for each slot: "Today's news" four times a day is something nobody reads after a week. Second, consider whether you need them every day: for many sites Sunday is a day when one recurring push fewer is a favour to the reader.
How to tell if you are overdoing it
Gut feelings count for little: numbers count. The Natom dashboard has four to look at together.
Impressions. Notifications actually displayed on readers' screens. On their own they say nothing, but they are the denominator of everything else.
Clicks. Readers who tapped the notification.
CTR. Clicks divided by impressions. It is the thermometer of relevance. The absolute value depends a lot on the industry and the topic, so do not compare it with other sites: compare it with yourself.
Subscriber trend. In Analytics you find the subscriber chart over time and the new subscribers of the last seven days. Subscriptions get rechecked and those no longer valid are removed, so the line tells you how many reachable readers you really have. If you keep collecting sign-ups but the line does not rise, something is leaking along the way.
The main warning sign is one: a CTR that drops week after week. A single push that performs badly means nothing: the title was weak, the timing unlucky, the story not very interesting. But if the weekly average CTR falls for three or four weeks in a row while the number of pushes goes up, readers are telling you they have stopped looking. Before unsubscribing, people stop clicking.
Also look at CTR by time slot and by audience. Often the problem is not the total but one slot (late-night pushes) or one kind of push (generic promotions) dragging the average down.
A four-step method to find your frequency
1. Start from content, not from the number. List what you really have worth notifying in a typical week: urgent news, new articles, deals, fixed appointments. That is your natural ceiling. If you have three things worth a push in a week, sending ten means sending seven fillers.
2. Split before you increase. If you cover several topics, turn on preferences and create audiences by category before raising frequency. Each extra push aimed at the people who want it costs far less than an extra push aimed at everyone.
3. Put brakes on automation. Daily cap and quiet hours on automatic feeds, or draft mode if you want to decide yourself. Recurring pushes at fixed times for the ones that repeat every day. That way you decide frequency once, with a cool head, and not in the moment an article comes out.
4. Change one thing at a time and watch four weeks. Add one recurring push, or raise the cap by one, and let a month go by. If CTR and the subscriber trend hold, readers have accepted the new rhythm. If CTR drops steadily, go back. Changing three things together leaves you not knowing which one worked.
Repeat the cycle. The right frequency is not found once and for all: it changes with the site, with the seasons, with the kind of news of the period.
Frequently asked questions
Is there a maximum number of pushes beyond which everyone unsubscribes? No. Some sites send many a day with happy readers, and some send two a week and lose subscribers. What counts is the match between what you send and what the reader expected.
Is one push per article better than a digest? It depends on volume. With few articles a day, one push per article is fine. With many, better a few recurring pushes to a summary page, plus the important stories sent separately.
Do category pushes count toward the total? For you yes, for the reader no. They only see the ones for the categories they picked, plus the ones sent to everyone. That is why splitting lets you send more.
If a reader is in two audiences, do they get the same push twice? No. A push sent to several audiences (up to five) reaches each device only once.
Do quiet hours apply to pushes I send by hand? No: they apply to feeds with automatic sending. A manual push goes out when you send it, or at the time you schedule. Common sense about timing is yours there.
How do I know if one extra push did damage? A single push is rarely visible in the numbers. Look at the weekly average CTR and the subscriber trend in the weeks after the change of rhythm.
Do pushes take readers away from the newsletter? They are different channels with different rhythms: push lives on timeliness, email on depth. We cover it in push vs email.
If you are still evaluating which tool to use, frequency is one of the things to check before choosing: caps, quiet hours, drafts and categories are not details, they are what lets you grow without losing readers. You will find a complete checklist in how to choose a web push platform.