To check Instagram fake followers before you buy a post, you are looking for a pattern, not a single smoking gun. Purchased followers, follow/unfollow bots, engagement pods and recycled comment spam all leave slightly different traces. No public tool is “100% accurate”, and a clean-looking grid can still sit on a hollow audience.
This is a buyer’s checklist: red flags you can see without logging into their account, the Insights screenshots to request, and the category of tools that estimate audience quality. Pubshr does not include a fake-follower scanner — do not expect one in the product. Use this pass together with how to choose an Instagram influencer, not instead of it.
Key takeaways
- →Fake followers show up as a bundle of flags — ratio, reach, comments, spikes, geo — not as one magic percentage.
- →Ask for Insights: reach, profile visits, saves, audience countries on recent posts, not a follower screenshot.
- →Comments that could be copied onto any photo are a stronger tell than a slightly low like count.
- →A sudden follower spike with no matching viral post is a pause, not a bonus.
- →Third-party audits are directional. They are not a courtroom exhibit, and Pubshr does not provide one.
Why
Why fake followers wreck the media buy
You pay for attention. Fake or inert followers inflate the denominator (follower count) and starve the numerator (real reach, saves, questions). A $400 Reel on a 50k account that only reaches 3k real people in the wrong country is not a bargain — it is an expensive CPM once you do the maths in the Instagram ads cost comparison.
Not every quiet account is fake. Small, niche, or recently shadow-limited profiles can have honest low reach. The job is to separate “small and real” from “large and hollow” before the invoice.
Flags
Red flags you can check in 15 minutes
Work down this table on the public profile plus any Insights they send. Two or three flags together are a no. One mild flag (for example a high following count on a new creator) is a question, not a verdict.
| Signal | What it often means | What to do |
|---|---|---|
| Followers far above recent reach | A large inert or purchased audience; the algorithm has no one to show | Ask for reach on the last 8–12 posts; if reach stays ~5–10% of followers with no story, walk away |
| Following almost as many as followers (on a “big” account) | Follow/unfollow growth or a non-creator behaviour | Normal for small personal accounts; odd for a 40k “influencer” |
| Engagement vs reach mismatch | Likes from pods, or reach bought without comments | Compute ER by reach — see engagement rate — and read the comment text |
| Comment quality | Bots, pods, giveaway leftovers | Sample 30 comments; see the next section |
| Sudden follower spikes | Bought batches or a contest that attracted prize-hunters | Ask what content caused each spike; no matching Reel is a problem |
| Audience geo mismatch | Bought cheap followers from unrelated countries | Compare Insights top countries with the market they claim and the market you serve |
No single score: tools that output “87% authentic” are estimates from models you cannot see. Use them to decide whether to ask harder questions, not as a substitute for Insights screenshots.
Comments
Comment quality, pods and empty praise
Open three recent organic posts (not the sponsored one they want to sell you) and read 10 comments on each. You are looking for language a stranger could paste under any photo.
Comment patterns that fail the check
- →Generic praise. “Nice pic”, “Great post”, “🔥🔥🔥” from accounts with no profile photo, or the same phrase stacked.
- →Wrong language. A French food creator whose comments are mostly another script and never mention the dish.
- →Name-dropping spam. Comments that are just a mention of random users — a classic pod leftover.
- →Burst timing. Fifteen comments in two minutes, then silence — more typical of a swap group than of a slow real audience.
- →No questions. Real buyers argue, ask sizes, ask prices, tag a friend. A product account with zero questions in a month is odd.
Giveaways confuse this test. If the last three posts are “follow + tag a friend to win”, ignore that engagement and ask for a normal week of content. Pods can look slightly more human than bot comments; they still will not survive a geo and reach check.
Spikes and geo
Follower spikes and audience geography
Public third-party charts (the category of historical follower-graph tools) are useful for one job: seeing whether growth is a stair of bought chunks versus a slope that matches content. A vertical jump of several thousand followers in a day, with no Reel that could explain it, is the classic purchased batch.
Geo is the flag brands skip and then regret. If you sell in Germany and 45% of their audience is in countries you do not serve, almost half of the quoted reach is not in your market. That belongs in the effective CPM, not in a footnote.
Match language to countries
A Spanish-language creator with top country United States can be real (diaspora). A local-services creator with top countries that never appear in comments is not.
Check Stories vs feed
Stories replies and sticker taps are harder to fake at scale than feed likes. If they will not share Stories Insights, that is information.
Non-follower reach
Healthy Reels often reach a large share of non-followers. An account that only ever reaches its own follower list, and barely, may be stuck — or padded.
Age and gender vs the product
Not a fake-follower test, but a wasted-budget test. A 70% male audience for a product bought almost only by women is a fit fail even if the followers are real.
Insights
What to ask the creator for (Insights screenshots)
Native Instagram Insights are the primary source. Third-party estimates are a backup. Ask for dated screenshots (so they cannot recycle a good month) covering the last 8–12 organic posts plus the audience tab.
Per-post reach and views
Feed and Reels. You want the median, not the best outlier. Plays without reach is incomplete.
Interactions
Likes, comments, saves, shares. Saves and shares are harder to buy in bulk than likes.
Profile activity
Profile visits, follows from the post, external link taps if they used a sticker. This is what a buyer actually wants from a post.
Audience
Top countries, cities if available, age brackets, gender. Compare to the market in the media kit.
Accounts reached
Followers vs non-followers on recent Reels. A paid post that never leaves the follower set will not do discovery work.
A growth explanation
If a chart shows a spike, they should be able to name the content or event. “It just grew” is not an answer.
If they stall, offer to sign a one-page NDA. Creators who still refuse Insights after that are asking you to buy a number they will not evidence. That is enough to stop.
Tools
Tool categories — without the “100% accurate” claim
You do not need a branded shopping list. You need to know what each category can see. None of them replace Insights, and none of them are offered inside Pubshr.
| Category | What it is good for | What it cannot do |
|---|---|---|
| Native Instagram Insights | Real reach, saves, audience geo from the account itself | You only see what they send; they can omit bad posts |
| Historical follower-graph tools | Spotting batch spikes and dead-flat paid growth | They do not prove why a spike happened |
| Audience-quality auditors | A second opinion on suspicious geo/gender mixes and mass-follower patterns | Model scores are estimates; vendors disagree with each other |
| Comment and caption sampling (manual) | Catching pods, spam and language mismatch cheaply | Slow; does not scale to 40 posts without a process |
| Your own UTM on a test link | Whether their Stories sticker or bio mention actually sends people | Only works if they agree to a tagged URL |
Honest product note: Pubshr schedules Instagram feed posts, carousels and Reels, writes captions and keeps a calendar. It does not detect fake followers, does not crawl Instagram audiences, and does not score influencers.
Use an auditor when the fee is large enough that a false positive is cheaper than a bad post. For a $150 nano Reel, the 15-minute manual pass plus Insights is usually enough. For a mid-tier package, two categories (Insights + a graph or auditor) is a reasonable bar. Then go back to the selection maths in how to choose an Instagram influencer.
Frequently asked questions
How can I check Instagram fake followers for free?
What follower-to-following ratio is suspicious?
Are fake-follower checker tools accurate?
What Insights should I ask an influencer for?
Does Pubshr have a fake-follower scanner?
Wrap-up: evidence, then the invoice
Run the red-flag table, read real comments, demand dated Insights, and only then convert the fee into CPM. Two or three flags together are cheaper to walk away from than to explain to whoever approved the budget.
Keep Pubshr for your own publishing workflow. The fraud check is a buying skill, not a button in the scheduler.
Publish on an audience you actually own
Pubshr helps you schedule Instagram posts, Reels and captions to your own account. It will not scan someone else’s followers for fakes.
Start for freeFree plan, no credit card required