Best time to post on X (Twitter) (2026)

Buffer’s 2026 analysis of 8.7 million posts on X puts the three best slots at Tuesday 9 a.m., Wednesday 10 a.m., and Wednesday 9 a.m., with mid-morning (9–11 a.m.) the most reliable window on every weekday. Wednesday is their best day overall; Saturday is the worst, with Friday close behind, and evenings from 6 to 11 p.m. underperform across the whole week.

Hootsuite’s analysis — 1M+ posts across 118 countries, run with the data-science agency Critical Truth and published in late 2025 — agrees almost exactly: a tight 9–11 a.m. window from Wednesday to Friday, with Monday and Tuesday peaking around 10 a.m. SocialPilot’s study of 50,000+ accounts and roughly 700,000 posts joins the morning camp too, favoring 8–11 a.m. on weekdays plus a smaller mid-afternoon bump around 3 p.m.

Sprout Social is the outlier. Their 2026 report — nearly 2 billion engagements across ~307,000 profiles between November 2025 and February 2026 — lands on 12–6 p.m. for Tuesday through Friday, with Tuesday–Thursday strongest and Saturday the worst day on the platform.

The split is methodological, not random. Sprout maps when engagement volume happens — and afternoons are when the most people are scrolling and reacting. The morning camp measures which publish hour earns each post the most engagement — and a post published at 9 a.m. is still fresh when that midday scroll peak arrives, while a post published at 4 p.m. ages into the evening lull. Read together, the studies agree more than they appear to: publish late morning, harvest the afternoon.

X no longer shows a chronological feed by default, but recency still weighs heavier here than on any other major network — the For You feed is built around live conversation, and a post’s window of opportunity is short. Sprout Social’s 2026 report describes the mechanism as engagement velocity: the platform treats early likes, reposts, and replies as a quality stamp, and posts that earn them quickly get shown to far more people.

Buffer’s 2026 analysis adds the honest caveat that timing matters less than it did in Twitter’s chronological era — the algorithm values engagement signals above the clock. But those two statements compound rather than conflict: the algorithm decides how far your post travels based largely on what happens in its first hour, and timing decides how many people are around for that first hour.

The weekly rhythm follows the news cycle. X engagement tracks workdays — commute scrolls, coffee-break checks, live reactions to whatever is happening — and Sprout Social attributes the Saturday collapse to users disengaging from the news cycle altogether. When the news pauses, so does X.

None of the four studies cited on this page publish format-level timing data for X — Buffer’s 2026 report recommends experimenting with text, images, video, and polls, but does not rank their hours. So treat what follows as reasoning, not measurement.

Threads (the multi-post kind) are the format most dependent on timing: they spread through early replies, reposts, and bookmarks, so launching one into a dead zone wastes its strongest distribution mechanic. The late-morning midweek windows above are the natural home for them. Live commentary is the opposite case — when you are reacting to a launch, a match, or breaking news, the event is the timing, and no heatmap applies; this is where X engagement compounds fastest. Evergreen one-liners and links are the most forgiving format and the easiest to schedule into the blended windows above, freeing your live attention for the conversations that cannot be scheduled.

X tolerates — and rewards — a higher posting frequency than slower feeds like Instagram or LinkedIn, because content cycles out of view within hours. Buffer’s 2026 guidance is to post daily, or multiple times a day if you have the bandwidth, and to treat consistency as the growth lever rather than any single perfectly-timed post.

A practical default from the blended heatmap: one post each weekday morning (Tuesday–Thursday 9–11 a.m. as your anchors), a second afternoon slot on your strongest days if you can sustain it, and a weekly experiment slot you rotate. Replies count too — on X, answering people in your niche is distribution, not admin. Batch-write the scheduled layer in one sitting so your live time goes to conversation, and hold the cadence for a month before judging any slot.

Every dataset above averages wildly different accounts — newsrooms, SaaS founders, sports fans, and meme accounts mixed together. Your followers are not average. A developer audience may scroll X late at night; a B2B audience checks it with morning coffee; a sports audience shows up when the game does. Even the studies disagree with each other, which is the strongest argument for trusting your own numbers. (Hootsuite’s data even finds a modest Saturday 8–10 a.m. window — a minority view worth testing only after the weekday slots.)

The reliable method is a simple loop. Pick two candidate windows — one from the morning camp (say Tuesday 9 a.m., per Buffer) and one from Sprout Social’s afternoon range (say Wednesday 2 p.m.) — and alternate between them for three to four weeks while keeping format roughly constant. Then compare impressions and engagement per post, keep the winner, and test a new challenger.

SocialKit handles the boring half of that experiment: queue the alternating calendar once, let posts go out at the same times every week across all 11 platforms, then compare impressions per slot in the built-in analytics until your own winner is obvious.