A tweet's median half-life is about 80 minutes. That is the moment when half of every impression the post will ever earn has already been counted, and for roughly 95% of tweets the impression count stops moving in any meaningful way inside 24 hours. Both numbers are real. Neither one is yours.
What you get from this walkthrough.A plain breakdown of the three separate numbers people all call the lifespan of a tweet.A repeatable method for reading your own account's impression decay instead of borrowing a global median.A repost interval you can defend with your own data rather than habit.
Circleboom's Twitter Post Analytics charts impressions, engagements, reposts, and replies day by day for your connected X account, with every count delivered through X's own Enterprise API rather than a scraped estimate. Start where the data lives and measure the average lifetime of a tweet.
Why the average lifetime of a tweet is really three numbers
The confusion in every article on this topic comes from one thing: three different clocks get reported under the same name.
The first clock is the peak window. Impressions per second reach their maximum about 72 seconds after a tweet goes out, according to the half-life measurement study published by researchers at the Technical University of Munich.
That is the number behind the widely repeated "a tweet lives 15 to 20 minutes" claim.
The second clock is the half-life, and the same study puts the median at roughly 80 minutes. Half of the post's total impressions land before that mark, half land after it.
The third clock is the tail. Impressions keep dripping in for hours, sometimes days, and then flatten. For most posts, the drip is gone within a day.
These three numbers do not contradict each other. They answer different questions:
- Peak window tells you when to be present in replies.
- Half-life tells you when a second exposure would land in fresh territory.
- Tail tells you when a post has finished earning and is safe to resurface.
The problem is that all three are medians pulled across the entire platform. A breaking-news account and a B2B software account do not share a decay curve, and neither of them shares yours.
If you have ever wondered why the published figures never match what you see in your total impressions on Twitter, that gap is the reason.
Work the arithmetic on a following spread across three continents and the mismatch stops being mysterious.
Suppose a third of your audience is asleep when you publish. That third cannot contribute an impression until it wakes, which pushes real activity into hour eight or hour ten while the platform median has already filed the post as finished.
The median is not wrong. It is describing a different account.
Circleboom charts impressions, engagements, and repost counts for your connected X account against the calendar, which is what turns a platform-wide statistic into a line you can check.
That reading is only as good as its source.
Circleboom is an official X Enterprise Developer company, so every number behind these charts arrives through sanctioned API access rather than a scraper guessing at a public page.
A decay curve built from estimated data produces a repost schedule that is confidently wrong, which is worse than having no schedule at all. Twitter Post Analytics is where the disagreement between the median and your own account becomes visible.
What a measured tweet lifespan changes about your posting on X
Knowing your own decay curve converts three habits from guesswork into decisions you can defend.
The first is posting frequency. If your bars flatten at hour six, a second post at hour four competes with the first one for the same feed slot. If they flatten at hour 16, that same second post lands in genuinely fresh territory.
That is the same underlying data behind best time to post on Twitter recommendations, read from the output side instead of the audience side.
The second is repost timing, which matters more than most accounts treat it. A repost fired before your curve flattens buys you very little, because the followers who were going to see the post are still seeing the original.
The third habit it changes is reporting. A flat month stops being an argument with a client and becomes a chart with a comparison window attached, which is a much shorter conversation than defending a strategy on instinct.
A fourth effect only shows up after a few weeks of reading the same chart. You stop treating every dip as a signal.
A decline that sits inside your account's normal range is weather rather than a trend, and the only way to tell the two apart is to have watched the range long enough to recognize it.
One boundary applies to all four. X serves analytics only inside its own retention window, so posts far enough back come back blank or half-filled even though they genuinely earned engagement on the day.
That is a platform limit rather than a gap in the tool. It is why measuring the average lifetime of a tweet on your own account works best on recent posts rather than archive material. If you have hit a wall pulling this data natively, the notes on how to see tweet analytics explain where the platform stops.
All four of those decisions run off the same input, and the input takes about ten minutes to produce.
How to measure the average lifetime of a tweet on your X account
Watch first: the impression pattern that shows up when an already-published post gets a second distribution window.
https://www.youtube.com/watch?v=XiItTNcpchM
The process runs in two phases: pull the raw impression trend for a known window, then read the shape of the decline instead of the totals.
Connect your X account and open the analytics view
- Log in to Circleboom Twitter and connect your X account through official OAuth authorization.
- Open the X Post Planner menu from the dashboard, where all post creation and performance views live together.
- Open Post Analytics and set the range selector to 4W. Shorter windows like 7D react violently to a single strong post, so a four-week base gives you a stable reference before you narrow anything down.
Read the decline, not the totals
- Choose Impressions in the first Insights dropdown and Engagements in the second. The chart renders both metrics as overlapping series across the date range, which is the comparison that shows whether reach and interaction decay at the same rate.
- Switch the chart type from Line Chart to Bar Chart and use the horizontal scroll slider to isolate the days after a post you remember publishing. Bars make a day-over-day drop easier to read than a smoothed line.
- Read the summary cards below the chart for the period-over-period change. Every card compares the selected window against the immediately preceding window of equal length, so a 4W view is always measured against the four weeks before it.
That order matters because each step removes a source of noise from the one before it. The four-week base cancels single-post distortion, the paired metrics separate reach decay from engagement decay, and the period-over-period card tells you whether what you are seeing is a real shift or a normal week.
Skip the base window and every conclusion afterwards is built on one lucky tweet.
Quick recap of the sequence:
- Connect the account with official OAuth.
- Open X Post Planner, then Post Analytics.
- Set a 4W range before narrowing.
- Pair Impressions with Engagements and switch to bars.
- Read the summary cards for period-over-period movement.
The number you are looking for is the day the bars flatten.
That is your account's practical tail. It usually sits further out than the global 24-hour figure for accounts with international audiences, and closer in for accounts publishing into a single time zone.
One honest limit on this method: the overview is an aggregate view, so what you are reading is the account's typical decline, not one specific post's curve.
On a day when you published once, the two are nearly identical. On a day when you published four times, the chart blends all four together, and the flattening point you read is an average of the set rather than a clean single-post measurement.
That is usually fine, because the number you want is the typical case.
When you do need a single post isolated, the move is to pick a quiet day, publish once, and read the following four days on their own. It takes a week of patience and returns a cleaner number than any published median will ever give you.
Every metric on that screen has a precise definition. The walkthrough of tweet activity and Twitter analytics metrics covers the ones this section assumes you already know.
Does reposting actually extend the lifetime of a tweet?
Reposting extends distribution, not the original post's decay curve. The first publication finishes decaying on its own schedule regardless of what you do afterwards; a repost simply starts a second curve for the same content in front of a partly different audience.
Circleboom builds that second window into the post editor itself. Inside X Post Planner, RePost after decides how long after publication the repost fires, Un-RePost after optionally pulls it back down again, and Repeat this cycle controls how many times that pair repeats.
Only the first of the three is a number your measurement produces. The other two are hygiene.
Set that delay to the point where your bars flattened, not to a figure you read somewhere. That is the whole payoff of the measurement.
One caution comes from the feature's own limits. A repost is a real action on your account, and X reads the same post recycled again and again as spam.
Un-Repost also removes only the repost. The original tweet stays, and so does every like or reply the repost collected while it was live. X's own public metrics documentation defines which counters move when, which is worth reading before you assume a second window doubles anything.
Evergreen posts are the right candidates here. Anything tied to a live moment reads as broken when it resurfaces 12 hours later, and the seam is more memorable than the point being made.
There is ranking context behind why a second window reaches people the first one missed. The explainer on how the Twitter algorithm really works covers those distribution mechanics.
Pair the timing decision with Twitter key performance metrics. That is how you tell a genuine second wave from a flat repeat that only moved the repost counter.
What guessing your tweet lifespan costs your X account
Without your own number, every posting decision inherits somebody else's audience. You post again too early and split your own reach. You repost too late and land in a feed that has already moved on. You judge a quiet month against a benchmark that was never measured on an account shaped like yours.
With your own number, all three of those turn into settings. The repost delay has a value. The gap between posts has a floor. A quiet month has a comparison window that tells you whether it is unusual.
The measurement takes about ten minutes the first time and roughly two afterwards, which is a poor excuse for running an account on a median pulled from millions of strangers.
What people ask
How long does a tweet stay visible on my profile?
Indefinitely, unless you delete it. Visibility on your profile and visibility in the feed are separate things: the post remains permanently on your timeline and in search, while its feed distribution ends within hours. That difference is why deletion and reach are separate decisions.
Can I compare two non-adjacent periods in Post Analytics?
Not in the overview chart. The summary cards always compare your selected window against the period of equal length immediately before it, with no manual control over the comparison window. For arbitrary period pairs you need post-level analysis rather than the trend view.
Why do my 7-day numbers swing so wildly?
Small sample size. A single strong post inside a seven-day window can produce a percentage change that looks dramatic against a quiet previous week. Four-week and three-month ranges give far more stable readings, which is why the measurement above starts at 4W.
Should I measure every tweet or just the important ones?
The account-level trend is the number you act on, so measure the account first and isolate single posts only when one of them behaves strangely. A post that outran your usual curve by a wide margin is worth pulling apart, because whatever caused it may be repeatable. A post that died early usually tells you about the hour it went out rather than about the writing.
Does deleting an old tweet change my analytics history?
Deleting removes the post and its future impressions, and the historical rows for that post drop out of the aggregate view along with it. Measure before you clean, not after, or your comparison window will move for reasons that have nothing to do with your content or your posting rhythm.
