How to Check Twitter Analytics
Where analytics live now, which numbers are worth reading, and which ones the dashboard counts differently from how you would expect.
Web and mobile access
Web Open X and navigate to the left panel Click More, then Analytics. You can access the same screen via Creator tools in that same menu on some accounts. The page which loads is the account-level dashboard. Seeing these numbers you do not need a third-party tool. With free accounts, it is just a small pack. More breakdowns appear on the same screens for paying subscribers. If you find yourself looking at the Analytics page frequently, ADD IT TO YOUR BOOKMARKS! The sidebar path is still okay when the post layout changes by update.
On mobile, skip the menu. Go to any of your own posts, and press on the view counter underneath it. That launches analytics for just that post and is the quickest route when you only care about one. From there, you can typically view the overarching dashboard. You may notice some period selectors and charts are collapsed in the mobile view as the space is tighter. Use web dashboard when you require a date range or follower change. Maintain both routes - sidebar for the account, view count of one post
Account dashboard metrics
The dashboard at account level, impressions & profile visits & engagement rate (how many people checked and clicked their posts) and change in followers At the top you select a period using date control. Impressions are the number of times your posts appeared in that period. Profile visits are those who opened your profile from those posts or from search. Follower change is that the net gain or loss for same dates These four are correlated side by side so you can tell if reach and follows move in unison or apart. For a tighter read, change the period to make it last seven days.
In addition to the above free features, paying subscribers see: extra charts, longer date ranges and splits by post type—all on the same dashboard. The core four are still available for free accounts to build from. A spike in impressions is not evidence that the writing improved. A change in exposure can raise impressions without any new interest appearing. Before you look to repeat, use engagement rate and follower change in addition to impressions. If profile visits grow while follows remain stable, people have looked and walked away. Check the bio next.
Individual post reports
Every posts give you the number of impressions, engagement, detail expired and profile visits, link click, new follow any likes your feedback is with reposts replies and bookmarks. View count on mobile or the post menu on web opens post analytics. Impressions for this are only on that post. The highlighted engagement metric is equal to the sum of… Detail extends count contacts went on to tap see more or open thread. Link clicks will only show up when you made a post that included a URL. New follows refers to the new followers you gained after they saw your post.
Few other scrolls on that list are as underrated as bookmarks. It suggests that if a bunch of people bookmark it, the intention is to come back which can be a stronger predictor of long term value than likes. Likes are cheap and public. Bookmarks are not public and require another tap. If bookmarks so far per like on a how-to or a thread are outpacing this, continue in that shape. Replies and reposts reflect the interaction with the post, if it sparked a conversation or was simply circulated. Profile visits from a post indicate that someone wants to know more info about you. View the row as an entire unit, not separate trophies.
Engagement rate reading
Engagement rate: This is the total of engagement divided by impressions. Because it adjusts for how far a post was promoted, its more informative than any raw total. For example, a post with 80 engagements on 2,000 impressions won over one with 200 engagements on 20,000 impressions. The first one held attention. The second was aired a ton and largely dismissed. This is the rate that X displays on your account dashboard and numerous post screens. When you choose what to write about next, use it. Raw likes reward distribution luck. Trust the rate.
Avoid going for a one-size-fits-all good rate. Accounts vary by size, so a solid number on an account with 500 followers can appear weak on a large one. The only equitable metric using is the median of your very own recent. Go back through your last 20 posts, take note of the engagement rate on each, and strike a mid line. Any of those well above that median attracted more interest than your average work. Anything obviously below that squandered its impressions. Stop guessing — write down the median once a month.
Format and timing patterns
Identify the patterns of success: format, theme, dayparting, thread vs single post. In a notes app, tag each of the last 20 posts with those four labels. Next to the label place the engagement rate. You will generally see a clump—not mystery. So you could have threads about one topic sitting above your median, whilst short jokes are beneath. You learn less from one viral outlier than you do from ten consistent posts. The outlier had assistance you will never have either. The cluster is repeatable.
After that, only time of day matters with format and topic held constant. For two weeks, run the exact same kind of thread at two hours and compare rates—NOT impressions. Impressions shift from hours to when your followers are engaged. Rate score tells you how much those people cared. If morning/evening threads have the same rate, post when you can write well. If 1 hour beats them on rate — hold that position. Resist the urge to copy posting times from accounts in other time zones or niches. Their audience is not yours. Ignore generic schedule lists.
Untracked profile history
Analytics only cover the posts that are still online. They say nothing about what impression your back catalogue leaves on someone who happens on to your profile and scrolls. A visitor lands on whatever is pinned and then a stack of dingy posts in reverse chronological order. But if those posts are on an off-topic basis or belong to another period, they educate the visitor that it is time to go. The dashboard will not trigger a warning for this. When a profile converts poorly, high engagement in the recent past can rest atop it. A week without metrics is not the same as a silent profile.
That audit is manual, and frequently better than another week of metrics. TweetSweep: TweetSweep provides you with all the history as a calendar so that you can see what is there and clean what no longer helps. Access the calendar, go through month by month and flag posts that would baffle a new visitor. Delete those first. Go back to the dashboard and see if profile visits are converting into follows. The figures following a cleanup represent a profile one is actually able to interpret. Stop blending all the old versions of you