# How to Test the Best Time to Post Instagram Reels

To choose a useful time to upload Instagram Reels, test two or three posting windows on your own account. Rotate comparable Reels across those windows. Then compare the same public count at the same time since publication. Use the result to choose your next schedule. The supplied evidence does not establish one best hour for every account.

Start with your own publication records and a simple worksheet. If you can access Instagram Insights, your private account reports, you can use them to help choose practical windows. Keep those reports separate from public counts. A public metric row does not show when your followers are online.

Readers ask both [whether upload time still matters](https://www.reddit.com/r/InstagramMarketing/comments/1ufrd9g/whats_the_best_time_to_upload_reels_on_instagram/) and how to [find basic patterns between counts and posting times](https://www.reddit.com/r/InstagramMarketing/comments/1uwsily/please_recommend_me_a_basic_but_decent_analytics/). Those questions explain the need for this guide. They do not supply a controlled timing result.

The method below produces two useful records: a planned posting schedule and a result table. It also gives you a clear next decision: continue, revise the test, or report no conclusion.

## 1. Define one account, one count, and one decision

A Reel is an Instagram video post. A metric is a measured value, such as a public view count. Choose one metric before you publish the test posts.

For example, you could choose public `views` at 24 hours after publication. That is a proposed test rule. It is not a claim that 24 hours is the best observation age.

Elapsed age means the time since publication. A count at 24 hours and a count at two days cover different elapsed ages. Keep them out of the same comparison.

Write these rules before you collect results:

1. Name the account you will test.
2. State the Reel format and topic limits.
3. Choose one public count and keep its definition unchanged.
4. Choose the elapsed age for each reading.
5. State the schedule decision you need to make.
6. Write how you will treat missing, late, and uncertain readings.

Keep `views`, `plays`, `playsInstagram`, `likes`, and `comments` separate. Do not replace a missing view count with plays. Do not add likes to views. A view count is not a count of unique people.

A useful decision might be to choose between two posting windows that you can maintain. This test does not establish Instagram ranking rules. It does not measure how long people watch, private reach, or separate paid and unpaid results.

## 2. Choose practical windows and one time rule

Choose two or three windows that fit your work. Start with two if a third would leave too few comparable posts in each group.

Define each window clearly. Give its start, end, and time zone. State whether the end time belongs to the window. For example, you can include the start and exclude the end. That rule prevents one publication time from belonging to two adjacent windows.

Write the account time zone and its daylight-saving rule. Daylight saving changes the relationship between some local clocks and UTC. UTC is a common time standard. Keep the local publication time for the schedule. Also save an exact UTC time when you have one. Use exact times to calculate elapsed age.

If the clocks change during the test, follow the rule you wrote. Do not quietly change a window from a local-clock schedule to a fixed UTC schedule.

Owner Insights can provide planning context. They do not replace the test. Keep any private audience information in your owner records. Do not claim that a public view count identifies an audience-online window.

You can use [the format and cadence guide](/en/articles/ig-format-cadence-map/) to describe your current posting pattern. Cadence means how often you publish. That guide counts publication activity. Here, you choose future posting times and compare later results.

## 3. Rotate comparable Reels across windows and days

Assignment means the window you choose for each Reel. Write the assignment before you see its results.

Keep the planned posts as comparable as your work allows. Record the topic, format, and other differences that could affect the comparison. Do not put every strong topic in one window and every uncertain topic in another.

Use rotation to spread those differences. Rotation means changing which window receives each type of planned post. You can also use random assignment, which uses chance to choose a window. Neither approach removes all uncertainty from a small test.

Follow this sequence:

1. List the Reels you plan to publish.
2. Group similar topics and formats.
3. Spread those groups across the chosen windows.
4. Spread each window across the days you intend to test.
5. Save the assignment before publication.
6. Record changes and the reason for each change.

Avoid giving one window only weekday posts and another only weekend posts unless that combined schedule is the question you intend to test. Otherwise, the day and time remain mixed together.

Do not move the strongest remaining Reels into a window after one large result. Keep the original plan. If you need a new plan, label it as a new test period.

Also record promotion, holidays, and collaborators. A collaborator is another account involved in a post. These differences can remain important even when you rotate assignments.

## 4. Save the schedule and observation records

An observation is a saved reading of a count at a known time. Keep the planned reading time separate from the actual reading time.

The table below lists useful worksheet columns. Some use Actor field names. An Actor is a hosted data collection program. Other columns are records you create yourself.

| Record | What to save |
| --- | --- |
| Account and test period | Account identity, test start, test end, and decision |
| Window assignment | Planned window, planned day, and assignment changes |
| Content notes | Topic, format, promotion, holiday, and collaborator notes |
| Post identity | Known direct URL, `postId`, and `postUrl` when returned |
| Publication evidence | Owner publication record, `publishedAt`, time zone, and time precision |
| Observation plan | Chosen metric, target elapsed age, and planned reading time |
| Actual observation | `scrapedAt` for an Actor reading, or the recorded time of a manual reading |
| Saved count | Metric field, returned value, source, and known precision |
| Collection state | `status`, `statusReason`, and `postMetricStatus` when returned |
| Comparison state | Included, late, failed, missing count, or unresolved time |
| Review note | Reason for exclusion, conflict, or other uncertainty |

Check publication-time precision. A date-only value tells you the day. It does not establish the publication hour. Do not turn a date into a midnight timestamp.

Keep a separate owner publication record when you need an exact hour. If that record conflicts with the public publication field, preserve both and resolve the conflict before you use the elapsed age.

For a manual reading, save the source and whether the displayed count is exact or rounded. Do not mix a rounded screen label with exact saved counts as if they had the same precision.

## 5. Collect the chosen public count

You can maintain the test with manual records if you can save the same public count at the planned age. Keep the source consistent across windows. Record the actual reading time each time.

The [Instagram Metrics Scraper](https://apify.com/chronometrica/instagram-metrics-scraper) offers a documented route for collecting current public metric rows from known content URLs. Its public build 1.0.35 accepts `mode` set to `contentUrls` and direct URLs in `startUrls`. The supplied public contract was checked on October 25, 2026. No paid run was made for this guide. This supports the documented input and field descriptions, rather than a current runtime result. See the [public default-build record](https://api.apify.com/v2/acts/Y7Ib4ARpUHw5Ue8op/builds/default).

A run is one execution of the Actor. The following small input was checked against the public input schema, which defines accepted input fields. It was not run:

```json
{
  "maxItems": 1,
  "mode": "contentUrls",
  "startUrls": ["https://www.instagram.com/reel/DIwcCZgxVZP/"]
}
```

5. Collect the chosen public count

Replace the sample URL with one known public Reel URL from your test. Replace the sample value completely. Keeping it beside your own URL would add an unrelated target.

`maxItems` limits the total number of returned metric rows. A limit of 1 suits this one-URL check. For a larger collection, review the limit so it does not cut your planned list. Still match every planned URL to a result or a gap.

The [Instagram Reels Performance Tracker](https://apify.com/chronometrica/instagram-metrics-scraper/examples/instagram-reels-performance-tracker) is a public Task. A Task is a saved input preset. Its page and preset identity were checked. Its saved settings were not read through an authenticated account, and it was not run for this guide. Inspect those settings and replace sample inputs before use.

The Actor supplies current public observations. You supply the posting assignment, reading plan, comparison, and decision. Do not assume the preset schedules readings or completes this experiment.

## 6. Check every planned reading before comparison

Use `scrapedAt` as the actual collection time for an Actor reading. Do not use the run start or your planned reminder time in its place. `publishedAt` is publication evidence. These fields serve different purposes.

Calculate the elapsed age from checked publication and observation times. A reminder to collect at 24 hours is only a plan. It does not prove that collection occurred at 24 hours.

For the exact 24-hour method shown below, only exact 24-hour observations enter the summary. Keep early or late readings in separate rows. If your workflow needs an allowed timing margin, define it before the test. Report the resulting comparison as an age band and show the actual ages. Do not describe it as exact 24-hour data.

Review these failure states:

| State | Required treatment |
| --- | --- |
| Count missing | Keep it unknown. Do not enter zero. |
| Returned count is zero | Preserve zero if the source and quality checks support it. |
| Reading is late | Save its actual age. Keep it outside the exact-age summary. |
| Publication time is date-only | Keep exact elapsed age unresolved unless separate evidence supplies the time. |
| Collection fails | Save the failure and any returned reason. Keep the planned post visible. |
| Metric differs | Keep the reading outside the chosen metric comparison. |
| Count precision is uncertain | Keep it outside an exact-count summary until resolved. |
| Target has no matching result | Keep a gap and investigate. Do not assume no activity. |

Read `status`, `statusReason`, and `postMetricStatus` where returned. Preserve their values. Do not assume that the presence of a row proves that every metric is usable.

A new run cannot recover a missed earlier reading. The [monthly saved-count guide](/en/articles/instagram-publication-cohort-view-contribution/) explains the same limit for past reporting boundaries. This test applies that limit to each post's planned elapsed age.

## 7. Compare the eligible posts and show the gaps

An eligible reading meets the rules you wrote before the test. Show every post, including those that do not qualify. Then summarize only eligible readings within each planned group.

For each posting window, report:

1. The number of planned posts.
2. The number of eligible readings.
3. The number and reasons for missing or excluded readings.
4. Every eligible count.
5. The median count.
6. The lowest and highest eligible counts.

The median is the middle value after you sort the counts. With an even number of counts, take the mean of the two middle values. The mean here is their sum divided by two.

The range shows the lowest and highest values. It helps the reader see whether a small median difference sits within a wide spread of post results. Do not report a median without its post count and gaps.

Review the assignment again. Did one window receive more promoted posts? Did one topic dominate? Did failed readings remove a whole type of post from one group? Keep those limits beside the result.

A difference describes what happened in the eligible sample. It does not establish a universal effect of publication time.

## 8. Work through an example in original fiction

Everything in this example is original fiction. These are invented teaching records. They are not Actor output or results from a real account.

The fictional account tests two practical posting windows, A and B. It uses exact public `views` at an elapsed age of 24 hours. The fictional owner has exact publication records. Comparable topics rotate across the two windows and test days.

Nine Reels were planned. Eight have eligible exact 24-hour readings. The ninth has only a late reading at 25 hours. Its count does not enter the 24-hour summary.

| Reel | Assigned window | Actual reading age | Exact 24-hour views | Comparison state |
| --- | --- | --- | --- | --- |
| A1 | A | 24 hours | 100 | Eligible |
| A2 | A | 24 hours | 120 | Eligible |
| A3 | A | 24 hours | 140 | Eligible |
| A4 | A | 24 hours | 160 | Eligible |
| B1 | B | 24 hours | 110 | Eligible |
| B2 | B | 24 hours | 130 | Eligible |
| B3 | B | 24 hours | 150 | Eligible |
| B4 | B | 24 hours | 170 | Eligible |
| B5 | B | 25 hours | Missing | Late reading kept separately |

B5's missing value is the unobserved exact 24-hour count. It is not zero. The late reading does not fill that gap.

Window A's sorted eligible counts are 100, 120, 140, and 160 views. Its middle values are 120 and 140 views:

Median A = (120 + 140) / 2 = 130 views.

Window B's sorted eligible counts are 110, 130, 150, and 170 views. Its middle values are 130 and 150 views:

Median B = (130 + 150) / 2 = 140 views.

The observed median difference is:

140 views − 130 views = 10 views.

Using A as the reference, the relative difference is:

10 views / 130 views × 100 = 7.69%.

The view units cancel in that division. This percentage compares two medians. It is not a percentage of people, a measure of certainty, or a promised increase.

| Window | Planned posts | Eligible readings | Missing exact-age readings | Median | Lowest to highest |
| --- | --- | --- | --- | --- | --- |
| A | 4 | 4 | 0 | 130 views | 100 to 160 views |
| B | 5 | 4 | 1, due to late reading | 140 views | 110 to 170 views |

B has the higher observed median among eligible readings. The groups overlap, and B has a missing exact-age reading. This small fictional sample does not prove a universal best time. A careful next decision is to repeat the planned comparison before adopting B as a stable preference.

## Choose the next schedule

Continue with a window when the pattern repeats enough for the decision you stated, and the schedule remains practical. Keep the conclusion local to the account, metric, elapsed age, and test period. The supplied evidence gives no universal number of posts that makes a timing result certain.

Revise the test when topics, days, promotion, or missing readings prevent a useful comparison. Report no clear difference when the available records do not support a preference. Preserve the original assignment and gap table in either case.

Start by writing two practical windows and assigning your next comparable Reels. Set the observation plan before publication. If you use the [Instagram Metrics Scraper](https://apify.com/chronometrica/instagram-metrics-scraper), check one known public URL with the small input above. Inspect the returned identity, timestamps, count, and status before you rely on the collection route.

## Sources

- [Instagram Metrics Scraper public default build 1.0.35 — checked October 25, 2026; public contract inspection, no paid runtime test](https://api.apify.com/v2/acts/Y7Ib4ARpUHw5Ue8op/builds/default)
- [Instagram Reels Performance Tracker — checked October 25, 2026; public page and preset presence only, saved settings and runtime not tested](https://apify.com/chronometrica/instagram-metrics-scraper/examples/instagram-reels-performance-tracker)
- [Reader question about the best Reel upload time — published June 26, 2026; observed August 11, 2026; one question, no controlled timing result](https://www.reddit.com/r/InstagramMarketing/comments/1ufrd9g/whats_the_best_time_to_upload_reels_on_instagram/)
- [Reader question about basic counts and posting-time patterns — published July 15, 2026; observed August 11, 2026; one question, no product-use or performance evidence](https://www.reddit.com/r/InstagramMarketing/comments/1uwsily/please_recommend_me_a_basic_but_decent_analytics/)