Analytics & improvement

A 15-minute weekly review for your X account: pick one post that worked, one that did not, and decide what to change

If you check your X numbers but never settle on what to change next, use this 15-minute weekly routine. Compare two posts, choose one thing to change in your next post, and keep a short record. You can put that change straight into your next draft and scheduled post.

Editorial responsibility: Ritomira teamPublished Updated

What you can do after reading

In 15 minutes a week, you can pick one post that did well and one that did not, compare them, choose one change for your next post and record it. The following week, open the record and check how the changed post did.

Scope & requirements · Features checked Nov 19, 2026

This guide covers X post data imported into Ritomira. The Post performance and Time and day screens are available on every plan. Notes requires Pro or higher. The metrics you can see depend on how data was imported (X sync or CSV). All numbers in this article are illustrative, not results from a real account. A review does not guarantee more engagement. Checked on November 19, 2026.

On this page

You look at your X numbers every week. Yet when you finish, you still have not decided what to change in your next post. This happens often when X is one of several jobs you handle.

It is not that you are bad at analytics. Usually, what to look at, what to change and what to record simply have not been decided. This guide turns that into a 15-minute routine: the same steps every week, two posts compared, and one change chosen. You do not need specialist knowledge of metrics to start, and a week where you cannot decide anything is fine too.

In Ritomira, Post performance, where you read the numbers, and Compose and scheduling, where you prepare the next post, are in the same place. You can put what you decide in the review straight into your next draft and schedule it.

The 15-minute routine

TimeWhat to doScreen
5 minChoose a period and posts. Pick one that did well and one that did notPost performance
5 minCompare the two and write down the differencesPost performance (check text and format)
5 minChoose one change for your next post and record itNotes (Pro or higher), or a notes app or spreadsheet you will keep using

The times are a guide to structure the session. Twenty minutes while you get used to it is fine. Following the same order each week matters more.

1. Choose a period and posts (5 minutes)

  1. Open Post performance and select a fixed period, such as the last four weeks. Using the same length each time makes weeks easier to compare.
  2. Narrow the post type (original, article, reply, quote) to what you want to review. Original posts alone are enough to start.
  3. Decide first what "did well" means. If you want more people to see a post, use impressions. If you care about reactions, use engagement rate. A different goal changes which post you pick.
  4. By that measure, pick one post that did well and one that did not.

Ritomira Post performance, filtered by period and post type to compare the numbers for each post

The screen example uses sample data.

Two things to watch when choosing. A very recent post has not collected its reactions yet, so pick posts with similar time since publishing. With very few impressions, a handful of reactions can swing the rate sharply, so do not call a post weak from the rate alone; check the impressions too. How to read X post analytics explains the metrics.

2. Compare the two posts (5 minutes)

Place the two posts side by side and write down the differences using these angles. At this stage, collect differences without trying to reach an answer.

AngleWhat to look at
OpeningWhat the first line conveys. Whether it leads with the point
TopicWhat the post is about, and how close it is to your usual topics
FormatImage or link, length and line breaks
Time postedDay and hour, and how long it has been live
Call to actionWhether it says what the reader should do

You may find three or four differences. Two posts alone cannot tell you which one mattered, so do not treat any of them as the cause. Keep them as candidates to test next.

3. Choose one change and record it (5 minutes)

From the differences, choose just one thing to change in your next post. If you change the opening, the format and the time together, you will not know next week what worked.

Four fields are enough for the record. The example below is illustrative.

FieldExample (illustrative)
Period reviewedLast 4 weeks, original posts
Two posts comparedDid well: a post showing how to use a product. Did not: an announcement
Difference noticedThe stronger post opened with the main point
One change nextIn the next announcement, put the key message in the first line

Next week, open this record first and look at how the post that tried the change performed. Do not conclude from one result. Watch several posts, then decide whether to continue.

A record template you can copy

Fill in the brackets and keep the record in the same place every week. On Pro and higher plans, you can also keep it in Ritomira's Notes.

Date of review: [e.g. Monday of the third week of November]
Period and post type: [e.g. last 4 weeks, original posts]
What "did well" means: [impressions / engagement rate, etc.]
Post that did well: [opening words or date]
Post that did not: [opening words or date]
Difference noticed: [e.g. opened with the main point]
One change next: [e.g. put the key message in the first line]
Result of last week's change: [e.g. only one post so far; keep watching]

Judgment calls to watch

  • While you have few posts, treat findings as tendencies. A gap between one or two posts does not make a form "the right one."
  • If posts differ in time since publishing, topic or format, record the conditions you could not align.
  • Treat posts boosted by ads separately rather than judging them by the same standard as regular posts.
  • If you could not try the change this week, just keep the record. You do not have to change something every week.

Making it last

Fix a day and time and put it in your calendar. "When I have time" is the first thing to slip. Keep the record in one place, so it can be handed over as it is when the person in charge changes.

If timing interests you, use the Time and day screen to see post counts next to responses. Do not label a slot "best" when it has few posts.

Ritomira Time and day analysis showing engagement and post counts by day and time

The screen example uses sample data.

Put the change into your next post

Once you have chosen the change, write it into a new post in Compose, save it as a draft and carry it through to scheduling. Schedule an X post explains the flow. Keep a one-person X account going shows how to combine the review with scheduling a week of posts in one session.

If you connect an AI app such as Claude to Ritomira, you can ask for "the usual review" and have it prepare the next drafts straight after. See Run your X account from Claude and Analyze your X posts and draft the next one from an AI app. The analytics feature page gives an overview.

Frequently asked questions

I only post a few times a week. Can I still do this review? Yes. Widen the period to the last four or eight weeks and pick two posts from it. While you have few posts, treat what you decide as a tendency and try it a few times before judging.

Which number should define "did well"? Decide by your goal: impressions if you want more people to see posts, engagement rate if reactions matter more. Write the measure in your record and keep it the same each week.

Can I keep this going on a plan without Notes? Yes. Notes requires Pro or higher, but a notes app or spreadsheet works for the record. Post performance and Time and day are available on every plan.

What if it takes longer than 15 minutes? Twenty minutes is fine while you get used to it. Spend less time collecting differences and make sure you get as far as choosing one change.

What was checked

Screen names and plan conditions (Post performance and Time and day on every plan, Notes on Pro or higher) were checked on November 19, 2026 against Ritomira's current labels and pricing table. The metrics you can see depend on how data was imported (X sync or CSV). Check the latest conditions on the pricing page and in Help.

This routine helps you organize how you read your data. It does not guarantee more reactions or results.

Editorial policy & corrections

Related guides

See the feature overviewCheck operation instructions and requirements