How-To Guide

How do you measure content decay on social media?

Read the decay curve rather than a lifetime total, set half-life expectations by platform, and decide whether to repost, refresh or retire a post.

10 min read

Measure decay as a curve, not a total. Pick the one metric that matches the post's job, keep timestamped readings at even checkpoints, and compute what share of the post's lifetime total had arrived at each checkpoint. The point where the rise becomes a fall is the peak; the checkpoint where the post crossed half of its total is its half-life. Set that curve beside the published averages, which in the 2026 Graffius update are 18.27 hours for Instagram, 1.43 hours for Facebook and 23.22 hours for LinkedIn, and beside your own earlier posts on the same surface. Repost, refresh or retire only after the tail has stopped adding reach.

  • One post, one curve: readings taken at different times, never a single end total.
  • One definition per comparison: platform help pages define impressions, views and engagement differently.
  • Half-life is the comparison unit: it turns a raw decline into a figure you can rank and revisit.

The short answer

Content decay is the fall in a post's reach and engagement after its peak. To measure it, keep timestamped readings of one metric for the same post, find the point where the rise turns into a fall, and work out how long the post took to collect half of its lifetime total. That half-life figure is the comparison unit: it tells you whether the post is still earning distribution or has settled into a tail.

Start from the platform's own definition, because the numbers are not interchangeable. YouTube counts a view when a video starts to play, and counts a thumbnail impression only when the thumbnail is shown for more than one second and at least half of it is visible. TikTok's video analytics report watch duration, total views and average watch time. Instagram's ranking explanation says it runs a variety of algorithms and processes, each with its own purpose, so a post can decay at different speeds on Feed, Reels and Explore.

  • Record the metric at more than one checkpoint. An end total cannot show where the fall began.
  • Hold the definition constant. An impression on one platform is not an impression on another.
  • Compute the half-life from the post's own totals, then compare it with published averages.
  • Decide after the tail is flat, not at the first quiet day.
A post's reach curve rises to a peak, then falls; the halfway point on the vertical axis is labelled half-life, and the flat section at the end is labelled tail.

Original diagram of the reading described in this article: peak, half-life and tail on one post's reach curve. The shape is illustrative; the definitions come from the Graffius half-life explanation and the platform analytics pages read 17 November 2026.

Superdirector

Make it work for your brand

Find it. Remake it. Post it. Superdirector turns the viral formats in your niche into your brand's carousels, X posts and video scripts, and keeps the calendar full.

Read each post as a curve, not a total

A total is an end state. A curve shows where the audience stopped arriving, and that is the part you can act on. Each platform gives you a different clock for the same idea. YouTube's help for creators says plainly that videos on YouTube generally taper off during the playback period. TikTok's Content tab lists the account's videos from the last seven days and surfaces the top nine with the fastest growth in views over that window. Instagram's ranking explanation describes predictions for each surface, and the actions it predicts differ by surface.

SurfaceWhat the platform reportsSource
YouTubeImpressions counted after one second with half the thumbnail visible; views counted when playback starts; key metrics within a few hours of publishingYouTube Help
TikTokWatch duration, total views and average watch time per video; a seven-day Content windowTikTok Newsroom
InstagramPredicted actions: Feed weighs seconds watched, comments, likes, shares and profile taps; Reels weighs reshares, full watches, likes and audio page visitsInstagram ranking explanation

The practical consequence is that you should compare a post with itself before comparing it with anything else. The shape is what you are reading: a fast rise and a short plateau is a different outcome from a slow rise and a long tail, even when both posts finish with the same total.

YouTube's retention report applies the same logic inside a single video. Dips highlight moments that were either skipped or moments where viewers stopped watching the video completely, and the intro figure tells you what percentage of your audience still watched after the first 30 seconds. If retention falls early, the decay began inside the video, before any distribution question.

Build the per-post reading

This is a five-step reading you can repeat for every post. It needs no tool beyond the platform's own analytics and a record you keep.

  1. Choose one metric that matches the post's job. Use reach or impressions for discovery, and watch time for video. Avoid engagement rate as the first choice, because Hootsuite notes that some platforms and tools calculate it against followers while others calculate it against reach or impressions.
  2. Record the metric at even checkpoints and keep the checkpoints identical across posts. Take the first reading after the platform has processed the day's data, since YouTube says its key metrics are available within a few hours of publishing.
  3. Turn each reading into a share of the post's lifetime total, using the final checkpoint you keep as the denominator.
  4. Mark the peak interval, then mark the checkpoint where the post crossed half of its total. That crossing is your measured half-life.
  5. Compare the curve with the published ranges in the next section, and with the account's own earlier posts on the same surface.

Two cautions keep the reading honest. First, a metric that keeps climbing has no settled total yet, so a half-life computed early will move as the tail grows. Second, a post that was amplified by a paid placement or a share from a large account is not a clean organic reading; keep it in a separate set so it does not reset your baseline.

When the numbers disagree with your memory of the post, believe the timestamped record. The impression that a post did well is usually formed in the first day, which is exactly the period the curve is designed to test.

Use published half-lives as a range, not a rule

The most complete public dataset is the 2026 Graffius update, which defines half-life as the time it takes for a post to receive half of its total engagement, such as likes, shares, and comments. It reports an average Instagram half-life of 1,096 minutes, or 18.27 hours, a Facebook half-life of 86 minutes, or 1.43 hours, and a LinkedIn half-life of 1,393 minutes, or 23.22 hours. Frederic Gonzalo reports that the update is built on 37 different data sources and restates the definition in plain language: the time it takes for a post to reach at least half of the engagement it will ever generate.

A second set of figures is often quoted beside it, and the two are not the same measurement. MtoM reports lifespans, a different measure: around 48 hours for an Instagram post, with the first few hours most critical; around 5 hours for Facebook, with most engagement in the first 2 hours; 15 to 30 minutes for an X post; and 24 to 72 hours for LinkedIn, with engagement sometimes continuing for a week or more. A lifespan is a working window, so it will naturally read longer than a half-life.

PlatformPublished half-life (Graffius, 2026)Published lifespan (MtoM, 2024)
Instagram18.27 hoursaround 48 hours
Facebook1.43 hoursaround 5 hours
LinkedIn23.22 hours24 to 72 hours, sometimes a week or more
Xnot cited in the facts used here15 to 30 minutes

Graffius also warns that algorithms and other factors on platforms can change over time, potentially affecting the average half-life. That warning is the reason to use the table as a starting range and the account's own history as the deciding evidence.

Horizontal bars compare published average half-lives: LinkedIn 23.22 hours, Instagram 18.27 hours and Facebook 1.43 hours, with a note that these are averages and not per-post rules.

Original chart of the average half-lives in the 2026 Graffius update, redrawn for comparison. Figures are averages from that dataset, not platform-published limits; X is omitted because the facts used here only carry a lifespan range for it.

Fix the metric definitions before you compare

Most false decay readings come from comparing two numbers that share a name but not a definition. Hootsuite's metrics guide puts the problem directly: some platforms and tools calculate engagement rate against followers, others against reach or impressions, so the same post can show two different engagement rates in two different tools. The guide's advice is to confirm how a platform defines a metric before you benchmark or report on it.

The same caution applies to reach and views. YouTube counts a view when playback starts, not when a viewer finishes, and it separates thumbnail impressions from the views they produced. TikTok's analytics show how long viewers watched, the total views the video received, and its average watch time, which are three different readings of the same post. Instagram's ranking page describes predictions about what a viewer will do, which is not a reach count at all.

TikTok's own guidance is a useful discipline for the whole exercise. It recommends considering the big picture of a video's performance over time, and not focusing on a single metric such as likes or comments from video to video. A curve built on one well-defined metric beats a dashboard of mismatched ones.

Decide whether to repost, refresh or retire

Once the curve is read, the decision is narrower than it looks. Ripl frames a content audit as four questions: what is worth continuing, what needs to change, which ideas deserve another try, and what should stop. Sprout Social adds a useful move for a post that worked: repurpose a high-performing short post on another platform to extend its lifespan and reach.

What the curve showsDecisionWhat must be true
Still rising, or half of the lifetime total not reachedKeep readingThe metric is still moving, so the total is not settled
Flat tail, topic still relevantRefresh and publish as a new postThe new version adds something material, not a repetition of the old one
Flat tail, topic no longer relevantRetireNothing current depends on the post
Fast rise and sharp fallStudy the hook before repeating the formatThe early data shows where viewers left or engagement stopped

Keep the repost rule honest. A repost is a new post with a new audience and a new start time; none of the sources read here promises that the old distribution returns. There is also a platform rule to check before blaming decay. Instagram says that if an account repeatedly posts content that goes against its Recommendation Guidelines or Community Standards, none of its posts will be eligible for recommendation for a period of time. That is a visibility limit with a stated cause, not a half-life.

A decision tree sorts a post by its curve: still rising means keep reading; flat with a relevant topic means refresh; flat with a stale topic means retire; a sharp fall means study the hook first.

Original decision guide drawn from Ripl's four audit questions and the repost caution in this article. It is a working guide, not measured data, and the labels mark the branch each decision needs.

What the numbers cannot tell you

A per-post curve is more honest than a total, but it still has blind spots, and the sources read here are explicit about some of them. Instagram says most of what a viewer sees in Explore comes from accounts they do not follow, which means an Explore post is being tested on a different audience from a Feed post. A single account-wide decay figure blends those audiences together.

Three gaps are worth naming. No stored source separates paid amplification from organic decay on the same post. No stored source gives a numeric cutoff for calling a post dead. And the published averages are averages, not readings of your account. Treat any one of those gaps as a reason to say less, not as a reason to fill it with a guess.

Make the reading a routine, not a one-off audit

A content audit and a decay reading answer different questions. The audit asks what to repeat; the decay reading asks whether one post is still working. Ripl suggests that for most active accounts, looking back three to six months is enough for the audit. The per-post reading is lighter and should happen while the post is still moving, so the record exists when the decision arrives.

The routine is the point. A founder who keeps one metric per post and three checkpoints has a defensible answer to the repost question. A founder who reads the end total has an opinion.

Superdirector's Growth plan at $199 a month runs one AI-operated social account: it researches what is working in the niche, builds carousels and X posts, schedules them and publishes to supported Instagram and X accounts under account-level standing posting permission. The Starter plan at $29 a month is the Claude or ChatGPT lane, where the same data, ideas, scripts and downloads are available but the account is not run for you.

Either way, the measurement advice on this page stands on its own: one metric, kept at even checkpoints, compared with a published range and the account's own history.

Frequently asked questions

How long does a social media post keep earning reach?

There is no single answer, and the published figures measure different things. The 2026 Graffius update reports average half-lives of 1,096 minutes (18.27 hours) for Instagram, 86 minutes (1.43 hours) for Facebook, and 1,393 minutes (23.22 hours) for LinkedIn. MtoM reports lifespans of around 48 hours on Instagram, around 5 hours on Facebook, and 15 to 30 minutes on X. A half-life is the time to collect half of a lifetime total; a lifespan is a working window. Use both as a starting range, then read your own post's curve.

What is half-life in social media measurement?

Graffius defines it plainly: the half-life is the time it takes for a post to receive half of its total engagement, such as likes, shares, and comments. It is a share of a final total, so it can only be computed after the total has settled. That is why a post needs a kept record at more than one checkpoint.

Is content decay the same as a drop in followers?

No. Follower count is an account-level number; decay is a post-level curve. YouTube separates thumbnail impressions from views and reports both, and TikTok reports watch duration, total views and average watch time for each video. None of those is a follower count, and a flat follower line can hide a post whose reach has already collapsed.

Should I delete a post whose reach has stopped?

Retire it only after the other options are checked. Ripl frames the audit as four questions: what is worth continuing, what needs to change, which ideas deserve another try, and what should stop. A post that answered a current question can be refreshed, and deleting it is not the only option; a post that still earns a slow trickle can be left alone. Deleting removes the record you would use to compare next time.

How often should I audit content decay?

Ripl suggests that for most active accounts, looking back three to six months is enough for a content audit. Treat that as a starting window, not a schedule for every account. Read the per-post curve while a post is still moving, and run the wider audit when you are deciding what to repeat.

Can I compare decay between platforms?

Only after matching definitions. Hootsuite notes that some platforms and tools calculate engagement rate against followers, others against reach or impressions, which is why the same post can show two different engagement rates in two different tools. Compare a post with earlier posts on the same surface first, and compare across platforms only on a metric each platform defines the same way.

Related practical guides

Record one post before you repost anything

Open the post's analytics, choose the one metric that matches its job, and write down the platform's definition beside your first reading. Add a second reading at the same interval you will use next time. When the tail flattens, compare the crossing with the published range and decide.