Spotting fake followers on Bangkok creators in 2026

A Bangkok creator with 82,000 followers and 340 likes per Reel is not a mid-tier buy. That ratio is a red flag you can catch in ninety seconds with a phone and a calculator. In 2026 the fake-follower problem in Thailand is not shrinking, it is just getting cheaper to fake. A pack of 10,000 Southeast Asian followers costs roughly 400 to 900 THB on Telegram resale channels, and comment pods run 200 to 500 THB per post for 20 to 40 seeded comments. That is inside the noise floor of any brand marketing budget, which means the incentive to inflate is always live.

This guide is how we vet accounts inside Mai Influence before a brief gets sent, and what to do when the numbers do not add up. It is not paranoia, it is arithmetic.

The ratios that matter, in order

The single most useful number is engagement rate on the last twelve posts, not the pinned ones. Take likes plus comments divided by follower count, averaged across twelve posts, and expect these ranges for organic Bangkok accounts:

• Nano (1,000 to 10,000 followers): 4 to 9 percent.
• Micro (10,000 to 50,000): 2 to 5 percent.
• Mid (50,000 to 250,000): 1 to 3 percent.
• Macro (250,000+): 0.7 to 2 percent.

An 80,000-follower fashion creator posting Reels at 0.4 percent engagement is either burnt out, shadowbanned, or padded. Any of those three outcomes is a bad buy for a launch campaign. Our micro vs macro creators guide has the tier maths in more detail if you want to price the delta.

Comments-to-likes ratio is the second cut. Healthy Bangkok accounts run 1 comment per 30 to 80 likes on Reels, 1 per 50 to 150 on carousels, and 1 per 20 to 60 on TikToks. If you see 1 comment per 8 likes and every comment is three emoji, that is a pod, not a community. If you see 1 comment per 400 likes, the likes are bought and the comments never followed.

The audit takes ninety seconds

Before you pay any tool a single THB, do this manually on the creator's grid:

1. Scroll to the twelfth post back. Note the date. If the account has 80,000 followers and the twelfth post is from four months ago, the follower base is not being fed. Real audiences drop off dormant accounts inside sixty days.
2. Sort the last twelve by like count. If ten of them cluster inside a plus-or-minus 8 percent band, the engagement is being smoothed by a service. Real Instagram engagement is spiky: a viral post at 4x the mean, three flops at 0.3x, everything else scattered.
3. Open the comments on the three most-liked posts. Count how many commenters have profile photos, bios, and at least one post of their own. Under 40 percent real profiles means the comment layer is inflated even if the follower count is real.
4. Check story views against follower count. Stories bypass the algorithm, so they measure roughly the true active audience. Real accounts see story views at 4 to 10 percent of follower count. Padded accounts see 0.5 to 2 percent because bought followers do not open the app.

Ninety seconds. No tools. This will catch 70 percent of the padded accounts in the Bangkok market, and that is before you spend on the deeper checks below.

Geography is where Bangkok creators leak

The Instagram Insights tab shows top cities and top countries. Ask any creator you are seriously considering to screenshot both, dated within the last thirty days. For a Bangkok-based lifestyle creator you should see Bangkok as the top city at 25 to 55 percent of the audience, Chiang Mai and Phuket in the top five, and Thailand as the top country at 55 to 80 percent.

Warning signs that keep appearing in the Mai Influence vetting queue:

• India, Brazil, or Indonesia in the top three countries for a creator selling to a Bangkok F&B brand. That is bot-farm residue from a growth campaign the creator paid for in 2023 and never cleaned up.
• Bangkok under 15 percent of city split for a creator whose whole grid is Sukhumvit cafes. The follower base does not match the content, which means the follower base did not choose the content.
• A round-numbered split (exactly 40 percent Bangkok, exactly 20 percent Chiang Mai). Insights never produces round percentages naturally. The screenshot is edited.

If a creator refuses to share Insights screenshots, that is your answer. Every serious Bangkok creator media kit in 2026 leads with this data because it is the only cheap proof of audience quality that a brand cannot manufacture.

Tools worth the THB, and tools that are not

A tool that estimates fake-follower percentage from public data can be wrong by 15 to 30 points in either direction. Use it as a second opinion, never the whole vote.

Modash, HypeAuditor, and Phlanx sit in the 900 to 4,500 THB per month range depending on plan and check volume. They pull public engagement, cross-reference commenter profiles, and score audience credibility. For a brand running one campaign a quarter, none of them is worth the subscription. For an agency or in-house team vetting ten creators a week, HypeAuditor's pay-per-report at roughly 340 THB per audit is the cleanest option in the local market.

What is not worth paying for: any tool that claims to detect fake followers without also asking for the creator's Insights export. Public data alone cannot see logged-in city and age splits, which is where the real signal lives. Tools that skip this step are selling confidence, not accuracy.

What to do when the numbers do not add up

You have three moves, in order of severity. First, request a fresh Insights export and rerun the ratios. Genuine confusion happens. A creator who ran a giveaway six months ago might have residue followers from Vietnam that they can point to specifically. If the story checks out, proceed with a reduced rate. A 20 percent haircut on the flat fee is the standard adjustment when 15 to 25 percent of an audience is off-market. Our Bangkok creator rates guide has the tier bands to anchor the renegotiation.

Second, restructure the deal from flat fee to performance-linked. If you cannot trust the reach number, pay for the outcome. Base fee at 40 to 60 percent of the original quote, with the balance tied to verified impressions from the creator's Meta or TikTok Ads Manager after the campaign runs. Verified numbers, not screenshots. This is the shape Spark Ads deals already use and it transfers cleanly to organic briefs.

Third, walk. If Insights are refused, engagement is smoothed inside an 8 percent band, and story views sit under 1 percent of follower count, the account is padded and no discount fixes that. A padded account does not convert, no matter how cheap you get it. Move on to the next name on the shortlist and note the account in your internal do-not-book file.

Building this into the intake, not the panic

The mistake most Bangkok brand teams make is treating fake-follower checks as an emergency response after a bad campaign. The cheaper version is a one-page intake questionnaire attached to every creator brief that asks for: last-thirty-days Insights screenshots, story-view averages on the last ten stories, and top-three commenter usernames on the last three posts. Creators who cannot supply this in under a day are the same creators whose numbers you would find wanting anyway. The intake self-selects. We built Mai Influence's vetting layer around this idea because the cost of catching a padded account before contract is a five-minute review, and the cost of catching it after is an entire campaign budget.

The Bangkok market has enough honest creators that no brand should ever pay for inflated ones. The audit is arithmetic and the arithmetic is cheap. Do it once per shortlist and the padded accounts sort themselves out.