Twitter Follower Analytics: The Six Numbers Worth Tracking (and the Ones to Ignore)

Open X analytics and you get impressions. Big number, goes up, feels good, tells you almost nothing about whether your account is heading anywhere.

Impressions measure how many times something appeared on a screen. Not read, not understood, not remembered. A post can do 90,000 impressions and produce two profile visits, and it has achieved essentially nothing.

Here are the numbers that actually predict something, in rough order of usefulness.

1. Net follower growth per week

Weekly, not daily. Daily numbers are noise dominated by whether you happened to post something that travelled.

Track gross gains and gross losses separately, not just the net. Two accounts both showing +100 a week can be doing completely different things: one gaining 110 and losing 10, the other gaining 400 and losing 300. The second has a serious retention problem hidden behind an acceptable-looking headline.

What good looks like: consistent positive net growth with gross losses under 20% of gross gains.

2. Churn rate

Gross unfollows as a percentage of total followers, weekly.

This is the number nobody tracks and it is the most diagnostic one available. Rising churn means you are attracting people who are not a fit, or posting things your existing audience does not want.

What good looks like: under 0.5% weekly. Above 1% weekly, something is actively wrong and worth investigating properly.

Match churn spikes against your posting history and the cause is usually obvious within ten minutes.

3. Follow-back rate by source

If you are following people to grow, this is your efficiency metric. Measured per targeting source, not in aggregate.

The aggregate number hides everything useful. Five sources averaging 18% might be one source at 40% and four at 9%. That is a completely different situation from five sources all at 18%, and it demands a completely different response.

What good looks like: above 20% for your best source. Below 10% means the audience is wrong or your profile is not converting.

4. Profile visit to follow conversion

Your profile visits divided into new followers. This isolates one thing: is your profile persuasive?

Everything upstream — posts, replies, follows, mentions — drives visits. This number tells you what happens when they arrive.

What good looks like: 5–15% depending on niche. Below 3% and the problem is your bio, your pinned post or your recent posts, in that order of likelihood.

This is the highest-leverage number on the list, because it multiplies everything else. Doubling it doubles your growth without changing anything about your traffic.

5. DM reply rate

If you send welcome messages or run outreach, this is the quality metric for your writing.

What good looks like: above 10% for welcome DMs, 8–20% for well-researched cold outreach. Under 5% means the message is wrong — almost always because it pitches, includes a link, or asks a question that takes effort to answer.

6. Reachable audience

Total followers minus dormant accounts minus obvious bots.

Most accounts discover their reachable audience is 60–80% of their follower count. That is the number worth planning against, and it prevents a lot of confusion about why engagement seems low relative to followers.

What good looks like: above 75% reachable. Below 60% suggests either bought followers somewhere in your history or targeting that has been pulling in low-quality accounts.

What to ignore

Impressions. Measures screen appearances. Correlates weakly with anything that matters.

Total follower count. The least informative number available. Ten thousand irrelevant followers is worth less than one thousand relevant ones on every metric that pays.

Likes. The cheapest possible signal. A like costs nothing and predicts nothing.

Engagement rate as a single headline. Too coarse to act on. Break it into replies, profile visits and follows, each of which tells you something different.

Follower growth of accounts you admire. You have no idea what they are doing behind the scenes, how long they have been at it, or what they started with.

A weekly review that takes ten minutes

Once a week, look at six things:

1. Net growth, with gross gains and losses shown separately
2. Churn rate, and whether it spiked on any particular day
3. Follow-back rate by source, with the worst source flagged for replacement
4. Profile visit conversion, compared with the previous month
5. DM reply rate by template
6. Which posts produced profile visits rather than impressions

Then make one change. Not five. One. Change five things at once and you will not know which one worked.

Building a feedback loop that improves things

Metrics only matter if they change what you do next.

Churn spike on a specific day? Look at what you posted. If a format keeps producing spikes, retire it.

Follow-back rate below 10% on a source? Replace the source, or check whether your profile is the actual problem by comparing across sources.

Profile conversion below 3%? Rewrite your bio today. It is twenty minutes and it multiplies everything.

DM reply rate below 5%? Rewrite the message. Shorter, no link, easier question.

Reachable audience below 60%? Tighten your targeting filters and stop bringing in low-quality accounts.

What most people get wrong

Checking daily. Daily numbers are noise. You will react to randomness and learn nothing.

Tracking everything. Six numbers is plenty. Twenty means you look at none of them properly.

Optimising for the comfortable metric. Impressions are comfortable because they are usually large. Churn is uncomfortable because it is a list of people who left. Guess which one is more useful.

Never writing anything down. Without a record you cannot see trends, and trends are where all the insight lives. A spreadsheet updated weekly beats any dashboard you check impulsively.

Frequently asked questions

What is a good engagement rate on X?

1–3% is normal, above 5% is strong. It varies enormously by niche and account size, so your own trend over time is far more useful than any benchmark.

How often should I check analytics?

Weekly for the review, monthly for trends. Daily checking is a habit that produces anxiety and no insight.

Does X's own analytics cover this?

Partially. Native analytics gives impressions, engagement and profile visits. It does not give churn, per-source follow-back rate, reachable audience or unfollower identity — which is most of what actually matters.

What is the single most important metric?

Profile visit to follow conversion, because it multiplies everything upstream. Fix that and every other channel you run gets more effective.

Should I track competitors' metrics?

Their public numbers, occasionally, for context. Do not obsess — you cannot see what they are doing behind the scenes and the comparison is usually demoralising rather than instructive.

How long before trends mean anything?

Four weeks minimum. Anything shorter is noise, especially for accounts under 10,000 followers where a single post can distort a whole week.

The summary

Six numbers: net growth with gains and losses split, churn rate, follow-back rate by source, profile visit conversion, DM reply rate, and reachable audience.

Review weekly, change one thing at a time, and write it down so you can see trends.

Ignore impressions. They feel like progress and they measure a screen refresh.

Get proper follower analytics in Tweeksocial — churn tracking, per-source follow-back rates, ghost detection and CSV export, from $27 a month.