Twitter Analytics Explained: What to Track in 2026

The short version: Most people staring at their X analytics dashboard are watching the wrong numbers, impressions and follower count tell you almost nothing about whether your content is working. Track engagement rate against your own baseline, profile visit to follow conversion, and bookmarks, then run a proper five step audit once a month instead of refreshing the dashboard daily out of anxiety.

Why bother with X analytics at all in 2026

I'll say something that will annoy the platform's biggest fans: X's own analytics tool is thin compared to what LinkedIn or even TikTok gives you. It shows you impressions, engagements, engagement rate, and a rough breakdown of likes, reposts, replies and clicks per tweet, and that's more or less it. No proper audience demographics screen for most accounts anymore unless you're paying for X Premium, no cohort analysis, no real content type comparison.

So why track it at all? Because it's still the only source of truth you have for what your audience on that specific platform responds to, and because the platform's algorithm has shifted hard toward rewarding watch time, reply depth and Grok-assisted content surfacing rather than raw follower reach. If you're not reading the data with that shift in mind, you're optimising for a platform that stopped existing about two years ago.

The core metrics and what they tell you

Impressions

Impressions count every time your tweet loaded on someone's screen, including bots, including people who scrolled past in half a second. I've had tweets hit 40,000 impressions with 11 likes. That's not reach, that's noise. Use impressions only to spot outliers worth investigating, never as a success metric on its own.

Engagement rate

This is the one worth watching, but only against your own history. X calculates it as total engagements divided by impressions. A healthy account posting consistently should sit somewhere between 1% and 3%. Anything under 0.5% for weeks in a row means your content isn't landing, full stop.

Profile visits and follow conversion

This is the metric almost nobody checks and it's the closest thing to a pipeline number X gives you. If 500 people visited your profile last month and only 4 followed, your bio and pinned tweet are broken, not your content.

Bookmarks

Added to the analytics view in late 2023, bookmarks are private, so people save things they intend to use or return to. A tweet with a high bookmark to like ratio is doing real work, it's the closest signal you'll get to "this was useful" rather than "this made me smile for two seconds."

A story worth sitting with

A client of mine, a SaaS founder with about 6,000 followers, was convinced his account was failing because his impressions had dropped from an average of 15,000 per tweet to under 4,000 over three months. He wanted to abandon the platform entirely. When we pulled the raw data into a spreadsheet and compared engagement rate rather than impressions, his account had improved, engagement rate went from 0.6% to 2.1% over the same period. His impressions dropped because X had quietly deprioritised low quality reach and started showing his tweets to fewer, more relevant people. Fewer eyeballs, better ones. He'd been about to make a decision based on the one number that mattered least.

A monthly audit process that works

I run this for my own account and for clients on the first Monday of every month. It takes about 40 minutes.

• Export the last 30 days of tweet data from the native analytics dashboard (Settings, Creator Studio, Analytics, then the export button top right).
• Sort by engagement rate, not impressions, and pull the top 10 and bottom 10 performing tweets into two columns.
• Look for patterns in the top 10: format (thread, single tweet, poll, image, video), posting time, topic, length.
• Check profile visits versus new followers for the month and calculate the conversion percentage.
• Set one thing to test differently next month, one variable at a time, never five.

That last step is the one people skip. They read the report, feel briefly informed, then post exactly the same way next month. Analytics without a specific behavioural change attached to it is just a hobby.

The uncomfortable bit nobody tells you

Here's the part that gets left out of most guides on this topic: X Premium subscribers get an algorithmic boost in replies and, to a lesser extent, in the main timeline. Your analytics can look like they're telling you a story about content quality when a chunk of the variance between your best and worst performing tweets is simply whether X's paid tier gave you a lift that day. I've watched two nearly identical tweets, same topic, same time, similar phrasing, perform completely differently because one account had Premium and the other didn't. Read your analytics with that in mind before you conclude your writing got worse.

Comparing yourself to competitors

Your own numbers only mean something in context. I use a mix of manual competitor checks and dedicated tools to see whether an account I'm advising is underperforming its niche or just underperforming its own past self. If you've never checked whether your competitors are outperforming you on social media, do that before you panic about a dip in your own numbers, because a 20% industry wide drop looks identical to a personal failure until you compare.

Tools beyond the native dashboard

X's built in analytics stop being useful once you want trend data beyond 28 days or want to compare hashtags and conversations you're not directly part of. This is where social listening tools earn their keep, they'll show you sentiment, competitor share of voice, and conversation volume around your industry that the native tool simply doesn't track. I also pull my best performing threads and repurpose them elsewhere rather than letting one good tweet die on one platform, and a decent look at content repurposing tools for video, audio and written content will save you hours if you're doing this by hand right now.

What good numbers look like in 2026

For context, accounts under 10,000 followers posting consistently (4 to 7 times a week) tend to see engagement rates between 1.5% and 4%. Accounts over 100,000 followers usually see that drop to 0.3% to 1%, because reach scales faster than genuine interest does. If you're a small account seeing 0.2% engagement, that's not a platform problem, that's a content problem. Cross checking content ideas against what's working on other platforms helps too, I regularly look at what's performing best in a niche on TikTok before I decide what angle to take on X that week, because audience curiosity tends to travel across platforms even when format doesn't.

Reply depth and quote tweets, the underrated pair

The dashboard buries reply and quote tweet counts inside the general engagement total, but I pull them out separately every month. Replies show whether you started a conversation, quote tweets show whether people wanted to add their own take publicly, which is a stronger signal of relevance than a like ever will be. A tweet with 200 likes and 3 replies got agreed with quietly. A tweet with 40 likes and 25 replies started something. If you're building an audience with intent to sell something eventually, the second tweet is worth more to you, and it should shape what you post next.

When to stop checking daily

I check my own numbers once a week now, not daily, because daily checking rewards you for chasing spikes rather than building a body of work. Five years of rebuilding this business in public taught me that the accounts that grow steadily are the ones whose owners aren't refreshing analytics every two hours waiting for a dopamine hit. Set a day, do the audit, close the tab.

I keep every related walkthrough in the Twitter (X) Help: 21 Guides to Features, Verification, Video Fixes and Tools.

Related reading: does twitter save watch history.

Related: writing for us on data analytics.

Frequently asked questions

Is X analytics still free in 2026?

The basic dashboard, impressions, engagements, engagement rate and top tweet per month, is still free for every account. Deeper demographic breakdowns and extended historical data are locked behind X Premium.

What's a good engagement rate on Twitter/X right now?

For most small to mid sized accounts, 1.5% to 4% is healthy. Above 100,000 followers, 0.3% to 1% is normal because reach outpaces genuine interest at scale.

Should I track follower count as a main metric?

No. Follower count tells you almost nothing about active interest, plenty of accounts with 50,000 followers get less real engagement than accounts with 5,000. Track profile visit to follow conversion and engagement rate instead.

How often should I check my Twitter/X analytics?

Weekly for a quick look, monthly for a full audit with a specific change to test. Checking daily tends to encourage chasing spikes rather than building consistent, useful content.

Useful references

• the ICO's guidance on cookies and similar technologies