Bot-flavoured likes have measurable fingerprints: zero-post profiles, default avatars, recent creation dates, suspicious follower-to-following ratios. We give you a 7-marker scoring rubric and our own pass-rate against it (94%+ across 12 sample weeks).
“Real users.” “Real-looking accounts.” “High-quality profiles.” Every provider says it. Almost none let you verify it. Here’s a method to verify it yourself, in less time than it takes to argue about the question on Reddit.
The method
Pick one post that received likes from a paid plan (or any batch you’re curious about). Open the like list. Open 10 random profiles in new tabs — pick from top, middle, and bottom of the list, not just the top. Score each profile against the 7 markers below. Anything 5/7 or higher is a real-looking account; anything 2/7 or lower is a bot.
The 7 markers
1. Profile photo
Real accounts have a non-default avatar. The default Instagram silhouette is the single highest-precision bot signal — about 71% of cleanly-detected bot accounts have it, vs 4% of real users. +1 if avatar is custom.
2. Post count
Real accounts publish at least occasionally. Anything below 3 posts on an account older than 90 days is suspicious; zero posts is a red flag. +1 if post count ≥ 3.
3. Follower / following ratio
Bots tend to have massively more following than followers (the standard “follow-trade” pattern). Healthy personal accounts sit between 0.4× and 8× follower-to-following. Anything where following > 5 × followers is suspicious. +1 if ratio is normal.
4. Bio
Bots either have an empty bio or a string of 10+ emoji with no language content. Real accounts usually have at least a sentence, a hashtag, or a link. +1 if bio has natural-language content.
5. Engagement on own posts
Open the account’s most recent post. If it has likes and at least one comment, the account exists in a network. If it has literally zero of either, the account exists only to like. +1 if own posts have engagement.
6. Tagged photos
Real users get tagged. The “Tagged” tab on a real account usually has at least one photo. Empty Tagged tab on an account with 100+ followers is a marker. +1 if Tagged tab is non-empty.
7. Username pattern
Bots get names auto-generated by farm scripts: words plus random digits (jenna_4719) or vowel-heavy strings (aoaela_qq). Real users tend toward names, nicknames, or interest-based handles. This is the noisiest marker — give yourself benefit of the doubt. +1 unless username looks generated.
The rubric
| Score | Read | Action |
|---|---|---|
| 7/7 | Real account, well-maintained. | No concern. |
| 5–6/7 | Real-looking; might be inactive but is a person. | No concern. |
| 3–4/7 | Borderline — likely a low-effort real account or a high-effort bot farm. | Note it; if multiple liked profiles score here, ask the provider. |
| 0–2/7 | Bot account, near-certain. | If most of the sample scores here, the provider is selling pure bot inventory. |
Our pass-rate
We run the same audit weekly across our reseller panel mix (we use a small whitelist, not whoever is cheapest that week). The published score across the last 12 weeks:
Two notes on the methodology so the number is fairly comparable:
- We sample 30 accounts per audit, not 10, to reduce variance. The 4-minute version above is for fast personal use.
- We exclude accounts that have been deactivated in the days between the like and the audit; that’s a real signal but it inflates the bot count unfairly.
If your provider scores below 80%
Email them. Ask: “What proportion of your reseller inventory scores 5+/7 on a 7-marker quality audit?” The reaction is informative. Honest providers know the number; sloppy ones improvise; bad ones argue. None of those answers require them to share trade secrets — it’s a quality stat about their own product.
Start a no-password Instagram auto-likes plan in under a minute. No login, no shared credentials, transparent pricing.
START A PLAN