How Often to Post on Social Media? Frequency and Half-Life

Why frequency depends on half-life

The question of how often to post is usually treated as a discipline problem. It is a distribution problem. On the large platforms every post is tested against the entire content pool instead of simply being delivered to your followers. How long that test runs, and how long a post keeps collecting reach afterwards, differs between channels by orders of magnitude. How delivery works in principle is set out in the overview of social media algorithms and distribution.

Half-life here means the span of time in which a post collects half of its total reach. A short half-life means the feed has moved on and visibility only comes from posting again. A long half-life means a post keeps working and additional frequency buys you little. Running both channel types at the same weekly count burns capacity in one place and stays below the visibility threshold in the other.

The second driver is the type of distribution. Feed-driven channels hand out most of the reach shortly after publication. Search- and recommendation-driven channels keep delivering long afterwards, because a post there surfaces again and again as an answer to a query or as a recommendation. How search works on the individual platforms is covered in Social SEO on TikTok, Instagram and LinkedIn.

How often to post on social media: the reliable data

Little of it is reliable, and most of what is comes from Instagram. Buffer analysed 2.1 million posts from around 102,000 Instagram accounts: accounts posting 3 to 5 times per week recorded more than double the follower growth rate compared with accounts posting 1 to 2 times, at 6 to 9 posts the factor stood at around 3.7, and at 10 or more posts at around 5.5. Reach per individual post did not fall in the process, it rose: plus 12 percent at 3 to 5 posts per week, plus 18 percent at 6 to 9, plus 24 percent from 10 upwards.

The biggest jump sits between 1 to 2 and 3 to 5 posts. After that, every additional post costs more than it brings in. For a team without a content department of its own, that is the practical limit.

Two caveats belong with this. The analysis is correlational, not causal: accounts that post more often usually also have more resources, established processes and more practice. Frequency explains part of the effect, not all of it. And the data base is the global Buffer user base, not a DACH sample.

The second documented finding concerns the lower limit. In an analysis of 4.8 million channel weeks from around 161,000 profiles on Facebook, Instagram and X, accounts stayed consistently below their own baseline in weeks without a post. A break therefore costs you more than the posts you skipped: it lowers the level you start again from.

Taken together, the two findings produce an uncomfortable order of priorities. What counts first is that something appears regularly at all, and only then how much. A week with two posts beats a week with none, and the difference between none and two weighs more than the one between five and ten.

For the remaining platforms the recommendations openly contradict each other. Buffer's frequency guide names 1 to 2 posts per day for Facebook. HubSpot considers daily posting there unnecessary and reports, as of August 2025, that only 19.7 percent of marketers publish several times a day. Anyone looking for a number will find one for every position. That is why half-life is the better starting point than a second-hand benchmark table.

Half-life per platform: the classification

The table below sorts channels by mechanics, it measures nothing. Platforms do not publish half-lives. What is documented is the distribution mechanics and Buffer's Instagram frequency figure; sorting channels into short and long is a judgement based on experience.

Three points the table does not show. First, search extends half-life noticeably. The eMarketer Social Search Report 2025 puts TikTok at 73 percent of users who search in the app at least once a day and almost 40 percent who do so several times a day; the figures are not DACH-specific. On YouTube, suggested videos drive the largest share of traffic. A post that feeds into a recurring search or interest pattern keeps collecting views long after publication.

Second, LinkedIn has actively moved that boundary. VP Product Gyanda Sachdeva has confirmed tests that show older but still relevant posts at the top of the feed again. The visibility of a LinkedIn post therefore depends less strictly on the time of publication than it used to.

Third, feed speed depends on how the target group consumes. In Germany, daily usage time sits at around 75 minutes for TikTok against 23 minutes for Facebook and 21.5 minutes for Instagram (Agorapulse, 2025). Where users scroll for long stretches, the feed moves on faster and the individual post has a shorter window.

What this means for editorial planning in a small team

Work out your cadence from your capacity, not from the picture in your head. A channel with a short half-life demands ongoing output, otherwise it effectively does not exist. A channel with a long half-life tolerates fewer posts but demands more quality per piece, because that piece works for a long time. That is the trade-off: volume where the feed forgets, depth where search remembers.

For a two to five person team in DACH B2B, this split holds up in practice:

• One core channel with 3 to 5 posts per week: the range where the Buffer data shows the best return and which a small team can hold without permanent stress. In DACH B2B that is usually LinkedIn, while XING has become effectively irrelevant as a content platform; the details are in LinkedIn vs XING in the DACH region.
• One secondary channel with a long half-life: a video or a long-form piece at wider intervals that keeps running through search and recommendations and doubles as a source for AI answers.
• One owned channel: newsletter or WhatsApp channel, independent of ranking changes and therefore the only hedge against a bad algorithm month.
• One hero asset per topic: produce it once, then re-edit it natively per platform instead of producing separately for every channel. Subtitles, aspect ratio and hook are part of the re-edit, not of the export.

Work the effort out in hours before you fix a cadence, and do it including idea, production, approval and replies in the comments. That last item gets forgotten in planning on a regular basis and in small teams it is usually the one that breaks the cadence. If those hours are not blocked in the calendar, you are planning a frequency that collapses after a few weeks. On the Buffer data, a collapsed cadence costs more than a lower one from the start, because the weeks off sit below your own baseline.

The most expensive planning mistake is spreading across too many channels. Five channels with one post per week each sit below the threshold at which frequency has any measurable effect anywhere. Two channels with a clean cadence beat that clearly, at the same effort.

The first 60 minutes: what is documented and what is folklore

The start phase is real. New posts are tested on a small audience, and early signal quality decides the further delivery. In the published X code, early engagement velocity is visible as a lever. For LinkedIn, the industry analysis by Richard van der Blom (Just Connecting, 1.8 million posts, 2025) shows around 5.2 times amplified reach with three or more comments in the first 60 minutes. That is not a LinkedIn figure, so read it directionally.

Folklore is the golden hour with fixed thresholds: the idea that a post has to collect y interactions within x minutes or it is dead. The same applies to the staged test pools on TikTok, first 500 views, then the next tier. Neither is backed by a primary source. Something practical follows from it anyway: post when your target group is active, and schedule community time directly after publishing. Not because a clock is running, but because early comments create signal quality. Why drops in reach are rarely a secret punishment is covered in What is a shadowban.

Typical mistakes in frequency planning

• Cadence set by channel count instead of capacity: four channels mean more than four times the work of one channel, because native re-editing, separate community care and separate reporting come on top.
• Sprint instead of cadence: posting daily for weeks, then nothing for weeks. On the Buffer analysis the weeks off push down your own baseline, and the sprint does not make up for it.
• Volume without a quality floor: from around five posts per week the marginal return drops clearly. Thinner posts additionally weaken the signals that ranking responds to, meaning watch time, sends and qualitative comments.
• Cross-posting as a frequency lever: the same clip everywhere at once raises the post count but not the reach. TikTok deprioritises visibly third-party watermarked content, and the tightened originality policy has applied since 15 September 2025.
• A single channel with no hedge: major ranking changes come several times a year. Without an owned channel, every one of them is a direct loss of reach with no safety net.

How to tell whether your frequency is right

Measure reach per post and engagement by reach, not by followers. If reach per post stays stable or rises as frequency goes up, you still have room. If it drops clearly over several weeks, you have reached your team's quality limit and not the limit of the algorithm. Which values are realistic and in what order of magnitude is covered in What is a good engagement rate.

The second indicator is how reach is distributed over time. If most of the views of a YouTube video land in the first few days and almost nothing follows afterwards, the video is working as feed content and not as a search asset. In that case the topic choice is the problem, not the frequency. Which signals Instagram weighs in detail, and why sends count for so much there, is covered in Instagram algorithm 2026.

As a rule of thumb: do not change a new cadence after a couple of weeks, but only once enough posts have been measured to average out outliers and cold-start variance. As of August 2026, what holds for every figure here is that the principles are more stable than the individual values. And the principle is consistency before volume.