How often to post on each network: one answer per account type

The cadence that survives a bad month

Almost every content calendar is designed in a good week. Somebody sits down on a Monday with energy, counts the gaps in the diary and writes five posts a week onto a sheet. Then March arrives with the flu, an urgent client request and two bank holidays, and the sheet breaks. What is left behind is not a lower frequency: it is the feeling of being permanently behind, which is what makes people abandon the channel altogether.

So the useful question is not how much you should post. It is how much you can post in the worst month of the quarter. That number is your floor, and posting above it when you have the room is a bonus nobody is owed. Setting the number high and cutting it every few weeks does the opposite: fewer pieces published, and a channel that looks abandoned in every gap.

There is a second, less obvious reason to build upwards from the floor. These platforms do not reward a peak week, they reward there being something new when somebody comes back. Three pieces a week sustained over twelve weeks beat ten pieces in January and silence until April, and your own data shows this with uncomfortable clarity.

What each network forces you to produce

Before you settle on a number, look at what each network asks for per piece, because the cost is not the same everywhere and neither is the sustainable cadence. This is what each connected network actually accepts, with its caption limit and how many files fit in a single post.

That table explains at a glance why a cadence of five pieces a week means very different things depending on where you publish. A TikTok photo post takes up to 35 files, which is an afternoon of editing; a Bluesky post is 300 characters and a spare moment. And some networks are rigid about the unit: YouTube, Pinterest and Snapchat accept exactly one file per piece, so there is no option there to stretch one shoot into a carousel and fill the week's gap with it.

What actually makes a piece expensive

Shooting is rarely the expensive part. The expensive part is everything around it that never appears on the calendar, and there are three specific costs worth counting before you fix a number.

The first is the caption. Writing 2,200 characters that say something costs more than filming a shot, and adapting them to five networks with five different limits costs five times as much if you leave it until the end. Working the other way round, writing for the shortest limit first and expanding afterwards, turns that multiplication into an addition.

The second is the automatic first comment. The habit of pulling hashtags or a link out of the caption and into the first comment is only automated on Instagram, Facebook, LinkedIn and YouTube. On the other ten networks a person does that step by hand after publishing, which is fine at three posts a week and stops being fine at fifteen.

The third is alt text. Up to 1,000 characters per image, 500 on Pinterest, and only six of the fourteen networks actually store it while the rest ignore it silently. It is the first step to fall away when frequency goes up and the hardest to recover, because nobody goes back to edit a three week old post to add it.

Once those three costs are counted, the number stops being an opinion. If a finished piece takes you 90 minutes and you have six hours a week for content, your floor is four pieces and not seven, however often you read another figure. How that floor gets divided between topics and formats is exactly the job of your content pillars.

How to tell when your frequency is too low

This is where most people go wrong, because they look at reach and reach answers a different question. There are four readings that do work, and all four come from data you already have.

• Average engagement by cadence. The dashboard crosses your weeks against the average engagement each one produced and breaks it down by network. If your five piece weeks come out worse than your three piece weeks, you are not underposting, you are diluting.
• Engagement build up. What percentage of its final total a post has gathered in each window since it went out. If your pieces are still collecting engagement at 72 hours, posting daily has them stepping on each other.
• Best times. Average engagement by weekday and hour slot across your whole history, in Madrid time. It is worth placing the pieces you already have better before you consider adding another one.
• The period's engagement rate. It is recalculated from the totals, not averaged from the daily rates, which gives a different and worse number. The formula and its traps are in the engagement rate guide.

Two warnings about metrics that get misread in this context. Reach cannot be summed across days: somebody who saw you on Monday and on Tuesday is one person, not two, so a monthly reach obtained by adding thirty daily values will always make the frequency look like it is working. And follows attributed to a specific post exist on Instagram only, only on feed posts and stories, and are zero on reels by platform design: if your strategy is reels, that column is not bad news, it is an empty column.

Three starting points, by account type

A creator working alone does well to start below what they think they can manage. Two pieces a week on a single network for six weeks, with the alt text written and the caption looked after, teach more about what works than fifteen pieces spread across four networks nobody has time to watch. When those two go out without effort, move to three.

An in house marketing team has a different bottleneck: review. A piece is not ready when it is filmed, it is ready when somebody approves it, and that step needs its own days on the calendar. A cadence of four pieces a week with internal review needs production running at least one week ahead of publication, and a content calendar is where you find out whether that buffer exists or is wishful thinking.

An agency has the same problem multiplied by the number of brands, with the client in the loop. Between a piece being finished and going out there is often an approval, from the team or from the client, and if the cadence does not allow for that round trip, the frequency promised in the proposal turns into posts that go out late. The practical way to hold it is to plan brand by brand and publish from one piece to every network that brand has connected, which is what the publishing calendar is built for.

What to cut when the month goes wrong

There will be a month when it cannot be done. The difference between the accounts that survive that month and the ones that do not is having decided in advance what falls away, instead of improvising it on the 12th.

The first thing to cut is the number of networks, not the number of pieces. Holding three posts a week on one network beats letting one a week slip on five, because a network with gaps recovers worse than a network with less volume. The second is the expensive formats: a ten image carousel becomes one good image, and an edited video becomes a single unbroken take. The last thing you touch is the caption, because it is the only item on the list that costs no money and does decide whether anybody stays.

What never gets cut is the reply. An unanswered comment costs more than a week without posting: the person who asked had already given you their attention and you are handing it back empty. If the month looks bad, post less and answer everything, in that order. How that queue is organised when it arrives from fourteen places at once is covered on the calendar and publishing page.