Bir X hesabının takipçi kalitesini ölçüyoruz. Botlar, sahte hesaplar ve yıllardır uğramayan profiller anında ayrılıyor: parayı vermeden önce tam olarak ne aldığını bil.
Every sampled follower is scored against 12 independent signals, each with its own weight and each printed in the table below. The heaviest is default profile image at 25 points; the lightest is worth 5. A follower crossing 50 points is counted as a bot and 30 as suspicious. Those weights and thresholds are our model’s parameters, not a verdict X has issued about anybody.
The reason a sceptical model is needed at all is what an ordinary X account looks like. Measured on 24 August 2026 over a 45,308-account sample of the 18,486,455 X accounts PlayerSells indexes, the median account has 550 followers and 61.9% of the index has fewer than 1,000. Small, quiet and nearly empty is the normal condition of a real X account, so no single signal can separate a bot from a lurker; only the combination can.
This is a sample, and the page says so wherever it prints a percentage. X does not hand out complete follower lists, so the audit reads a slice of the most recent followers and reports what share of that slice looks automated. For an account with a few thousand followers the slice is most of them; for one with 89,454 followers - the 99th percentile of our index - it is a fraction of a percent, and the score leans on engagement and account health instead.
Signal weights and category cutoffs are parameters of the PlayerSells audit model. Follower distribution figures are our own measurement over the PlayerSells X index, last computed on 24 August 2026.
A follower starts at zero and accumulates points. Nothing here is proof on its own - a real person can have no banner and a numeric username - which is exactly why the model adds signals up instead of acting on any one of them.
| Signal | Points |
|---|---|
| Default profile imageThe egg. Bulk-created accounts rarely bother replacing it. | 25 |
| Zero postsAn account that has never posted has never been used by a person. | 20 |
| No bioEmpty description. Common on real lurkers too, hence the modest weight. | 15 |
| Zero followersNobody follows it back. | 15 |
| Follows accounts, has none of its ownOne-way traffic: the shape of a follow farm. | 10 |
| Follows over 500, has under 10 followersMass-following with nothing coming back. | 10 |
| Created under 7 days agoFresh accounts arriving in a batch. | 10 |
| Has never liked anythingReading X without ever tapping a heart is rare in real use. | 10 |
| Six or more digits in the usernameThe signature of an auto-generated handle. | 10 |
| Over a year old with under 10 postsAged inventory, or an account that was abandoned immediately. | 10 |
| Created under 30 days agoWeaker version of the 7-day signal. | 5 |
| No header imageWeakest signal in the set, and never decisive alone. | 5 |
50 points or more is counted as a bot and 30 or more as suspicious. Separately, a follower who has never posted and is over 90 days old, or has under five posts and is over a year old, is counted as inactive rather than fake: an abandoned account is a real person who left, and calling it a bot would be a different and wrong claim.
It cannot tell you whether the account owner bought those followers. Any account of any size collects bot followers it never asked for, and a large public account collects more of them than a small one. A high bot share is a reason to look at engagement rather than a verdict about the person.
A missing measurement is reported as missing, never as zero. If X returns no follower list or no recent posts, the audit says the sample could not be read rather than scoring the account at nought, because a working account told it scored nothing would act on it.
Tek tek bakmak yerine örneklem alınır: takipçilerin yüzde kaçı gerçek ve aktif kullanıcı, yüzde kaçı bot, sahte hesap ya da uzun süredir uğramayan profil, ona bakılır. Şişirilmiş takipçi sayısı gerçek erişimi gizlediği için, bir hesabı satın almadan önce yapılması gereken ilk kontrol budur.
Araç takipçilerin davranışına bakıyor: ne sıklıkla paylaşıyorlar, profilleri ne kadar dolu, kaç kişiyi takip edip kaç kişi tarafından takip ediliyorlar, gönderilere gerçekten karışıyorlar mı. Her takipçi bu sinyallere göre gerçek, şüpheli, bot/sahte ya da pasif olarak ayrılıyor.
100 bin takipçisi olup bunun %60'ı bot olan bir hesap, 50 bin gerçek takipçisi olandan çok daha az eder. Etkileşimi, gösterimi ve dönüşümü gerçek takipçiler üretir. Sahte takipçi hiçbir şey katmaz, üstüne platform cezasını bile tetikleyebilir.
Düzgün işleyen hesapların çoğunda bile doğal olarak %5-15 arası pasif ya da sahte takipçi bulunur. Sahte ve bot oranının %20'yi aşması tehlike işaretidir, genelde satın alınmış takipçiye işaret eder. %40'ı geçen hesaplar alıcı için yüksek riskli sayılmalı.
X, profilinden takipçileri tek tek kaldırmana izin veriyor. Ama toplu temizlik etkileşim metriklerini bir süre aşağı çeker. Satışa hazırlanmıyorsan sahteleri ayıklamakla uğraşmak yerine gerçek takipçi kazanmaya yüklenmek daha çok işe yarar.
Evet, tamamen ücretsiz ve üyelik istemiyor. Gerçek takipçi verisi üzerinden çalıştığı için herkese açık her X hesabında elle tutulur bir sonuç veriyor.
whether an account is being suppressed in search and replies
who is talking about an account and how far it reaches
what an X account is worth on the open market
engagement rate against accounts of the same size
the percentile a follower count sits in
how much two audiences actually share
the whole tracked market, sorted and filtered
PlayerSells is not affiliated with X Corp. Follower data is read live from public X profiles at the moment you run the audit and is never stored against your account. Signal weights and cutoffs are parameters of our own model, described in full above. The follower distribution quoted on this page is our own measurement over the PlayerSells X index, last computed on 24 August 2026. Nothing here is a statement by X about any account.
