How to Audit Your Twitter Following List in 30 Minutes
A following list that nobody has reviewed for years is full of silent, dormant and one-way accounts. A 30-minute method for finding out what is in yours, deciding who stays, and setting the clean-up running.
Tweeksocial Team
You can learn what is wrong with your following list in half an hour, without reading every profile. Count the list, sample a small part of it, sort the sample into a few groups, and scale the result up. Then protect the accounts that matter and start a clean-up at a gentle pace.
Most following lists have never been reviewed. They grew by habit over years, a follow here and a follow there, with an occasional burst from a campaign. The result is a list in which a large share of accounts are silent, gone or one-way, and nobody knows how large a share, because looking is tedious.
This method makes it quick. It takes five steps, with a time allowance for each, and ends with a decision table you can apply to any account.
Why audit before you clean up
A clean-up without an audit is a guess. You pick a grace period, switch on a rule, and discover afterwards what it removed.
An audit takes half an hour and tells you three things in advance. It tells you how big the problem is, so you can choose a daily pace that finishes in a sensible time. It tells you which kind of account dominates, so you know which rule to switch on first. And it shows you the people you must not lose, which is the input the whitelist needs.
It also tends to change how you see your account. Many people expect a few dozen accounts to go and find that it is closer to a third of the list.
Step one: count and compare (5 minutes)
Open your profile and note two numbers: how many people you follow, and how many follow you.
Divide followers by following. That is your ratio. An account following 3,200 people with 1,100 followers has a ratio of about 0.34, which reads as somebody following far more than follow them back. Above 1.0 reads as an audience, and the guide to the follower-to-following ratio explains what each range signals.
Write the numbers down with today's date. In three months you will want to compare.
Next, work out the gap: following minus followers. In the example above it is 2,100. That is the largest number of accounts that could be one-way, and it gives you an upper bound for the clean-up.
Step two: sample 50 accounts (10 minutes)
You cannot review thousands of profiles, but you can review fifty.
Scroll your following list and pick accounts from different places: some from the top, some from the middle, some from the bottom. The oldest follows are at the bottom and the newest at the top, so a spread of positions gives you a spread of ages.
Open each of the fifty and answer three quick questions:
- Does the profile say "Follows you"?
- When did the account last post? Anything within the last three months counts as active.
- Do you know who this is, and do you still want to follow them?
Tally the answers. You are looking for how many of the fifty fall into each group below.
Step three: sort into five groups (5 minutes)
Almost every account on a long list falls into one of five groups.
| Group | How to spot it | What to do |
|---|---|---|
| Keep | Follows you, posts, and you get value from it | Nothing |
| Protect | Does not follow you, but matters: a customer, mentor, journalist or peer | Add to the whitelist |
| Past the grace period | Does not follow you back, and you followed them more than a week or two ago | Queue for removal |
| Dormant | No posts for six months or more | Remove with the inactivity rule |
| Doubtful | Looks automated, has no profile picture or a strange following ratio | Review, and remove if it fits the pattern in removing bot followers |
Count how many of the fifty land in each group. These counts are an estimate, not a measurement, but they are a good one. If twenty of your fifty are past the grace period, then roughly two fifths of the whole list may be too.
As an example, suppose your tally gives 14 keep, 3 protect, 21 past the grace period, 9 dormant and 3 doubtful. Scaled to a list of 3,200, that points to about 1,340 accounts past the grace period and 580 dormant, with some overlap between them. The numbers will be rough, but they show where the weight is.
Step four: build the protected list (5 minutes)
Everyone in your sample who landed in the "Protect" group goes on the whitelist, and so does anybody you thought of while you were scrolling.
Add the usual groups from memory: customers, partners, people who reply to you, journalists, mentors, and any competitor you watch. The guide to the unfollow whitelist lists seven groups and a quick way to build the list from your messages and replies.
This is the step that makes the rest safe, and five minutes is enough to start. You can add to it any time.
Step five: choose the rules and the pace (5 minutes)
Your tally tells you which rule to start with.
- If past the grace period is the largest group, start with the non-follower rule. The method is in the guide to unfollowing non-followers.
- If dormant is the largest, or you are cautious, start with the inactivity rule, described in unfollowing inactive accounts. It is the gentlest of the lot.
- If doubtful is large, review those accounts first.
Choose a grace period of seven days, or fourteen for a young account, and a daily pace that finishes in a time you can accept. For an estimated queue of 1,900 accounts and a three-week window, that is about 90 a day. The reasoning for each setting is in Twitter auto unfollow settings.
Use the preview that shows how many accounts each rule would affect. If it roughly matches your estimate, you can proceed. If it is very different, check the grace period and the sample before starting.
What the audit usually shows
People are often surprised by three things.
How many accounts are simply gone. Accounts are suspended, deleted or abandoned, and they sit in the list for years. They cost nothing socially to remove, because there is nobody there.
How many one-way follows are deliberate. The list of people you want to keep, though they never follow back, is longer than most people expect. That is why the whitelist comes before the rules.
How few accounts you actually read. The timeline improves noticeably after a clean-up, because the accounts you care about are no longer buried.
When to audit again
Run the audit every three to six months, and after any burst of following. It does not need to be repeated in full. A repeat of steps one and two, the count and a sample of twenty, takes ten minutes and shows whether the rules are keeping up.
If the gap between following and followers is shrinking and the sample shows fewer past-the-grace-period accounts, the rules are doing their job. If the gap is growing, the follow volume is outrunning the clean-up, and you should raise the daily pace slightly or tighten the grace period.
Frequently asked questions
How long does it take to audit a Twitter following list?
About thirty minutes for a first pass. Counting takes five, sampling fifty accounts ten, sorting five, building a protected list five and choosing rules five. A repeat audit takes around ten.
Do I have to review every account?
No. A sample of fifty from different parts of the list gives a useful estimate of the whole. The preview inside an unfollow tool then confirms the exact numbers.
How do I tell whether an account is inactive?
Open the profile and look at the date of the most recent post. Six months with nothing is the usual threshold, and a year suits niches with many quiet readers.
What should I do first if the list is huge?
Build the protected list, then start with the safest rule at the lowest daily pace. The inactivity rule is the gentlest, and the non-follower rule is the most effective. The question of pace is covered in the guide to the 5,000 follow cap.
Will the audit change my follower count?
The audit itself changes nothing. The clean-up that follows improves your ratio, and it can change who follows you slightly. Whether the other person is told is covered in what X tells someone when you unfollow them.
Can a tool do the audit for me?
A tool can list every account that does not follow back and when each last posted, which saves the sampling. You still need to decide who to protect, because that is a matter of knowing the people. For the list of who left you, see the Twitter unfollower tool.
The summary
Count the list, sample fifty accounts, sort them into keep, protect, past the grace period, dormant and doubtful, build the whitelist, and start the gentlest rule at a low pace. That is a half-hour job and it replaces guessing with numbers.
Twitter/X auto unfollow previews how many accounts each rule affects before it starts, protects your whitelist, and logs every removal. It is included in every plan from $27 a month, with a 3-day free trial.
