Twitter Analytics Guide: How to Read X Analytics and Turn the Numbers Into Better Posts
Which X metrics matter, how to calculate engagement rate, and a simple weekly routine for turning your Twitter analytics into better posts.
By Shahriar Ahmed ยท
What Twitter analytics are for
Most people open Twitter analytics to check whether yesterday's post did well, feel good or bad about the view count, and close the tab. That is the least useful way to use them. Analytics on X are worth reading for one reason: to decide what to post next. Every number in this guide is judged by that test.
You can read your numbers in several places: under each post on X, in X's own analytics screens (what they show depends on your account), or in a tool such as GrowX, a Chrome extension that collects your post numbers into one Insights dashboard. The ideas below work wherever your numbers come from.
The X analytics metrics that matter
- Views: how many times a post was shown. This is reach, not interest. A big view count with few interactions means many people scrolled past.
- Likes: the lightest signal of approval. Useful in volume, but cheap to give.
- Reposts: people putting your post in front of their own followers. Reposts are how posts travel beyond your audience.
- Replies: conversation. A post with real replies is starting discussions, and answering them keeps the thread alive.
- Bookmarks: people saving a post to come back to. Many bookmarks usually mean the post was useful: a checklist, a how-to or a resource.
- Engagement rate: interactions per view. The one number that lets you compare posts with very different reach.
Follower numbers sit next to these: new followers, unfollowers and growth over time. They answer a different question, which is whether your posts are building an audience or only entertaining the one you already have.
How to calculate engagement rate
Engagement rate is interactions divided by views. Add up the likes, reposts, replies and bookmarks on a post, divide by its views and multiply by 100 to get a percentage.
For example, with made-up numbers: a post with 2,000 views and 50 interactions has an engagement rate of 2.5%. Another with 20,000 views and 200 interactions has 1%. The second post reached more people, but the first landed better with the people who saw it. To learn what your audience responds to, follow the rate. To learn what spreads, look at views and reposts.
Two cautions. Small posts give noisy rates: 3 likes on 40 views looks great but tells you little. And posts keep collecting views for a while, so judge a post after a day or two, not after an hour.
Find the topics and formats that work for you
Single posts are random. Groups of posts show patterns. Label your last few weeks of posts by topic (what each one was about) and by format, then compare the typical engagement rate of each group.
- Formats worth separating: questions, posts with images, plain text, threads and posts with links.
- Topics: your two to five recurring themes, for example tutorials, opinions, behind-the-scenes posts and reactions to news.
You will often find that one or two formats clearly beat the rest for your audience, and that a topic you post about constantly earns less than one you rarely touch. GrowX's What works view does this grouping for you by engagement, so you can see which topics and formats pull their weight without building a spreadsheet.
Which posts actually brought followers
Engagement and follower growth are related but not the same. A joke can collect likes from your existing followers and bring in nobody new. A plain, useful thread might get fewer likes and still attract followers for days. If growth is your goal, you need to know which posts did that.
The manual way: note your follower count before a post and again a day later, and compare the gain with a normal day. It is rough, because posts and replies overlap, but over a few weeks the standouts show. GrowX Premium does this for you: for each post it looks at new followers in the 24 hours afterwards and highlights the posts that brought followers above your usual pace.
Look at the profile side as well. When a post clearly drew attention but few people followed, the post may be fine and the profile may be the problem. A clear bio and a pinned post that shows what new followers will get turn more of those visits into follows.
Common Twitter analytics mistakes
Most analytics errors come from reading too much into too little data, or from measuring the wrong thing.
- Chasing views alone. Reach without interactions or new followers is mostly people scrolling past.
- Overreacting to one viral post. A single outlier tells you little. Patterns that repeat across several posts are what you can act on.
- Changing everything at once. If you change topic, format and timing in the same week, you cannot tell which change helped.
- Ignoring replies. Replies are conversations waiting for an answer, and answering them is part of how a post performs.
- Comparing yourself with other accounts. Different niches and audience sizes produce very different numbers.
A 20-minute weekly analytics review
- Compare this week with last week: new followers, unfollowers and total views.
- List your three best posts by engagement rate and your three best by new followers. Note what they have in common.
- Check the weakest posts too. Was it the topic, the format or the timing?
- Look at when your posts did best. GrowX's best-time-to-post heat map (a week by 24 hours) is built from your own posts' engagement and when your followers are active.
- Choose one change for next week, such as one more thread, a different posting time or a new topic, and write it down.
GrowX Premium sends a weekly report every Monday with this week against last week (new followers, unfollowers and best posts), which covers the first step. If you schedule your posts, the post scheduler lets you turn best-time findings straight into queue slots.
Keep a short log of these weekly decisions. After a couple of months you can look back and see which changes actually moved your numbers, which is far more useful than any single week's report.
How GrowX Insights collects your numbers
GrowX does not use X OAuth or ask for your password. It reads your numbers through an x.com tab that is already open in your signed-in session, one small step per minute at most, and refreshes your post analytics every 3 hours while x.com is open. Everything is stored in your browser.
- Free: views, likes, reposts, replies, bookmarks and engagement rate for your latest 30 posts, What works by topic and format, top fans (who likes and replies to you most), a growth chart, daily goals and streaks, and 7 days of follower history.
- Premium: full analytics, including which posts brought followers, plus follower quality, the weekly report and your full follower history.
See the analytics feature page for details, and the guide to seeing who unfollowed you for the follower tracker.
Read your X analytics with GrowX
GrowX is free to install from the Chrome Web Store, and you can create a free account to try every Premium feature free for 30 days, no credit card needed. Then use the engagement guide to act on what you find.
Frequently asked questions
How do I calculate engagement rate on Twitter?
Add a post's likes, reposts, replies and bookmarks, divide by its views and multiply by 100. GrowX shows this rate for each of your recent posts.
What is a good engagement rate on X?
There is no single good number, because it varies by niche, account size and format. Compare each post with your own typical rate and work on raising that.
Do views and impressions mean the same thing on X?
Broadly, yes: both describe how many times a post was shown. X displays a view count on posts, and many guides still use the older word impressions.
How often does GrowX update my X analytics?
Every 3 hours while an x.com tab is open. GrowX reads the numbers through that tab, one small step per minute at most.
Are GrowX analytics free?
The core is free: your latest 30 posts with views, likes, reposts, replies, bookmarks and engagement rate. Premium adds full analytics, including which posts brought followers, plus follower quality and a weekly report.