When is the best time to post on Threads: how to find your golden hours

Posting regularly but getting three hundred views? Most likely you’re posting at the wrong hours — when your audience isn’t in the feed. I built Treden, a platform for Threads auto-posting and analytics, and golden hours there are calculated separately for each account. Below are two things: a method you can repeat by hand in a spreadsheet, and how Treden does the same calculation itself, with a demo of what it looks like inside.

Averaged advice from the guides is useful exactly once: when the account is under two weeks old and there’s no statistics of your own yet. After that it starts to hurt, because you post at hours that are “right according to the internet” while your followers aren’t there.

What the averaged studies say and why they aren’t about you

The averaged data contradicts itself at the level of sources.

Buffer analysed 2.5 million Threads posts and named Thursday 9:00 as the best time, with weekday mornings from 6:00 to 11:00 as the optimal window. In their data, evenings from 18:00 to 23:00 performed worst. Meanwhile Postpone, on its own sample, recommends Wednesday, Friday and Saturday in the 23:00–01:00 interval. Russian-language authors offer a third opinion: 8–10 in the morning, 13–14 at midday and 18–20 in the evening, while another walkthrough names 9:00, 13:00 and 19:00.

Four sources, four different pictures. That’s normal, because each sample is computed over its own audience: Buffer’s global English-language base, a different base at Postpone, the personal observation of the post’s author. PostFast honestly clarifies that weekday mornings 8:00–11:00 in your audience’s time zone are strongest, but that’s still an average.

Someone else’s hours can be taken as a starting hypothesis for the first two weeks of an account. After that, only your own statistics decide.

Why the first 60 minutes decide the reach

The Threads algorithm looks at early interactions with a post and uses them to decide whether to show the thread wider.

The logic is this. Your post is first seen by part of your followers and a few people from recommendations. If answers, reposts and likes appear in the first hour, the platform expands the impressions. If it’s silence, the thread stops at a few hundred impressions, no matter how many good thoughts it holds. A post published while the audience is asleep collects its first reactions six hours later, when the train has left.

That’s why “I post quality content but get no views” is often cured not with new content but with shifting the publishing time by 3–4 hours.

How to calculate golden hours by hand: five steps

Collect the reach of your posts over 2–4 weeks, group it by publishing hour and find 2–3 windows where posts consistently gather more than usual.

Step 1. Export your posts’ statistics for the last 2–4 weeks. You need two columns: the hour of publication and the reach (in Threads that’s “Views” in each post’s stats). At least 15–20 posts, otherwise the sample is noisy.

Step 2. Split the posts by publishing hour and calculate the median reach for each hour. The median is more reliable than the average: one viral spike won’t drag the whole hour up. You’ll get something like this table (demo numbers for clarity, not my account):

In the demo table there are two candidates: 19:00 and 10:00. The 16:00 hour consistently fails.

Step 3. Take the top 10 posts by reach for the same period and look at which hours they were published in. If 6 of the top 10 fell into one or two windows, those are your golden hours. If they’re scattered chaotically, time isn’t yet the main factor in your reach, and you need to work on the content: the topic, the first paragraph, the format.

Step 4. Test the window with a series. For the next 5–7 posts, publish only in the found window and compare the median reach with the old one. One post proves nothing: any post can have a lucky and an unlucky day.

Step 5. Repeat every 3–4 weeks. The audience grows, the algorithm updates, the windows shift. A window that worked in September may sag by November.

How Treden does it: golden hours without spreadsheets

So I don’t have to calculate this in spreadsheets every month, I built the calculation into Treden: the platform takes your posts’ statistics and itself suggests hours for each day of the week. Here’s how it works.

• You connect a Threads account, and Treden pulls in the post history and their views.

• You open “Analytics”, choose the account and period. For the first calculation the last 30 days will do.

• In the “Reach by hours (TOP-10)” block you can see which hours the posts went out at and how much they gathered.

• In the “Recommendations by day of the week” block, Treden ranks hours by the average reach of your posts for each day of the week separately.

• Hours that hosted two or more posts come first: a single random post isn’t considered reliable for the calculation. If there’s no data for a day, the overall picture by hours is used.

There’s a nuance: Treden calculates the average, while in the manual method above I recommend the median. The average is more convenient for automation, and reliability is ensured by the “at least two posts per hour” rule. If you want to be extra safe, cross-check the recommendation against the top 10 posts, as in step 3.

What choosing golden hours looks like in Treden

This is what the “Recommendations by day of the week” block looks like: one row per day, hours in the row, the brightest ones higher in the rank. This is a demo account with placeholder data, but the interface is the one from the platform. Click the hours: an enabled hour lights up, a disabled one goes dark.

How to work with it, step by step:

• At the top there’s a switch: “TOP 1”, “TOP 5”, “TOP 10” or “Custom”. TOP N shows the N best hours for each day; in “Custom” mode you set yourself how many posts per day you want.

• Clicking an hour turns it on or off. Enabled hours are highlighted; the higher an hour sits in the day’s ranking, the brighter it is. A grey hour means you haven’t chosen it.

• The “Apply schedule” button saves the choice for the account.

• If there isn’t enough data, the day’s row says “Not enough data”. For a new account that’s normal — come back in a couple of weeks of posting.

Golden hours are available on all plans, including the free one.

What happens after choosing: golden hours in the planner

Saved hours don’t sit as a dead report — they become a working hint.

• In the planner they’re highlighted in the calendar. You can turn on the “golden hours only” mode and see just those slots in a day.

• When quickly adding a post, buttons with your golden hours appear next to the time field: press one, and the post is set to the right time.

• If you create drafts from ChatGPT or Claude via the Treden integration, the draft takes a free golden hour of the chosen date. You publish yourself.

Golden hours are a hint, not a guarantee. They put a good post at the time when it has the best chance of gathering its first reactions. They won’t save a bad post.

How much to post on Threads per day and how to spread posts across windows

A working norm for an expert: 1–3 posts a day, and the found windows should be filled first. That’s the “Posts per day” field in the schedule settings.

When there are two or three windows, everything becomes simple. The strongest post of the day goes into the best window, the second strongest into the second one. The third post, if there is one, you place at a new time and watch the data: this way you cheaply test an extra window without risking the main post.

Don’t dump 5–7 posts in a row in one evening. Each next post “eats” the reach of the previous one, because the audience is the same and the algorithm sees that you’re flooding the feed. One good post in a confirmed window is better than three merged in a row.

Why reach on Threads drops: a checklist beyond timing

Publishing time isn’t responsible for everything. Walk down the list from top to bottom:

• The publishing hour. The post goes out while the audience sleeps or works. Verified with the method above in one evening.

• A weak thread start. The first paragraph gives no reason to reply or read the thread to the end. The algorithm sees no early reactions and doesn’t boost the post.

• Zero activity in the first hour. You aren’t answering comments yourself. The author’s dialogue in the thread strengthens the signal for recommendations, while links in the first part of the thread, by authors’ observation, reduce the algorithm’s trust.

• A shadowban on Threads. The account is restricted, and posts don’t reach people who don’t follow you.

• A sharp change in behaviour. If you haven’t posted for a month and came back, or went into a daily marathon after a pause, the platform needs time to recalibrate the account. Give it 1–2 weeks before drawing conclusions.

How to check for a shadowban: ask someone who doesn’t follow you to find your fresh post via search by your username and text. If they can’t find it and the profile is barely visible from a stranger’s account, there’s a sign of a restriction. Review the rules, remove questionable content and pause for 2–3 days.

A ban is often confused with bad timing. A restriction suppresses all posts in all windows, while a bad hour suppresses only some. If all posts sank at once, it’s not about the hours.

What to do next

• Today: export the statistics of your last 20 posts (the “hour” and “views” columns) or connect an account in Treden and open “Recommendations by day of the week”.

• This week: 5–7 posts in one window, compared against the old median.

• Once a month: recalculate the windows and review the top 10 posts.

More about the platform I built is on the “Projects” page.