How to Spot Vote-Farm Shills on Crypto Twitter

You search a coin on X and the timeline is full. Dozens of accounts, all excited, all saying it is the next big one. It feels like a crowd has found something early.

Often it is not a crowd. It is a farm: a set of accounts posting the same words about whatever coin they are paid or told to post. This lesson shows how to tell the two apart by eye, and how PumpPill counts them for you.

Search by the contract, never the ticker

The first rule is the one that clears most of the noise.

A ticker is not a coin. Anyone can launch a coin with any ticker, and popular tickers get copied within hours. A post that says only "$TICKER is sending" could be about any of five coins, including the copy someone launched to catch the people searching.

A contract address points to one coin and only one. So PumpPill reads X by the contract address, never by the ticker. A post that does not contain the contract is not counted as someone pushing that coin.

That rule sounds strict. In practice it is what separates the people who actually want you to buy a specific coin from the people filling a timeline. A real caller posts the contract, because they want you to find the right coin. A farm account often never does.

What we found when we looked

PumpPill has stored 231,396 X posts that carry a contract address, from 26,837 accounts. Reading them by contract, and comparing their wording, turned up a pattern that is hard to see from inside a single timeline.

• One popular Robinhood Chain coin's "pushers" were all copy-paste accounts. Searched by ticker, it looked like it had a large following on X. Searched by contract, the accounts behind that buzz had never once posted the contract. Every one of them was posting the same kind of template.
• One farm template appeared on 720 different tokens. The same sentence shape, with the ticker swapped in, posted about coin after coin. That is not a community finding a coin. It is a script.

When a coin's X buzz is mostly this kind of copy, the buzz tells you nothing about who is buying.

How to spot a vote farm by eye

You can catch most farms in a minute of scrolling. Look for these.

Identical wording. Read five posts in a row. If they share the same phrases, the same emoji order, the same line breaks, with only the ticker changed, they came from one template. Real people describe a coin in their own words, and they disagree with each other.

Ticker but no contract. Count how many of the posts contain the actual contract address. If almost none do, those accounts are not trying to get you to a specific coin. They are trying to make a ticker trend.

New or empty accounts. Open a few profiles. Farm accounts are often recent, have few real followers, and have a timeline that is nothing but coin posts. Look at the replies too: a ring of the same accounts liking and replying to each other is part of the same setup.

Posting dozens of coins a day. Scroll an account's timeline. A real caller posts a handful of coins a week. A farm account posts dozens a day, each one "the next big one". PumpPill flags this kind of account with a plain count, for example "posted 833 contracts in 7 days". Nobody researches that many coins.

Same minute, many accounts. When a wave of accounts posts about one coin inside the same few minutes, that is coordination, not discovery.

One of these on its own can be innocent. Three or four together is a farm.

What Lens shows you

You should not have to do all of that on every coin. PumpPill Lens, the browser extension, puts the count on the token page you are already reading. On each token it shows a line like:

"N accounts posted the contract on X. M more posts were vote-farm copy."

The first number is the accounts that posted the actual contract, the ones that count. The second is the copy we set aside. When the second number dwarfs the first, the coin's X buzz is mostly manufactured. When it is small, the attention is closer to real.

Real attention still needs checking

Separating real posters from farm copy tells you who is talking. It does not tell you whether the people talking are any good.

Some accounts post the contract every time and still have a poor record, because they post coins after those coins have already run. To check that, look up the account's record with the KOL track record checker, or read our report on 1,585 callers. For groups rather than single accounts, see how to judge a crypto call channel.

And whatever X says, the coin itself still has to pass the basics: run it through the rug pull checker before you buy.

What this does not tell you

Our farm flags come from wording and posting patterns, so a careful farm that writes each post by hand can slip through, and an honest account that reuses a phrase can be caught. A farm-heavy coin can still run, at least for a while; the farm exists because it sometimes works. The point is not to avoid every coin with a farm on it. It is to know that the crowd you are looking at may not be a crowd.

Try it

Pick a coin your timeline is excited about. Search X for its contract address instead of its ticker and compare how many posts you get. Then open its page with Lens on and read the line: how many accounts posted the contract, and how many more posts were vote-farm copy.