What to look for
The five things worth checking before you scrape.
Three signs you should walk away immediately.
Why one busy day tells you nothing.
How to know afterwards whether the group was any good.
The temptation with a big group is obvious. Ninety thousand members, one click, and your list is full. It feels like progress, and it is the single most reliable way to waste a month of outreach.
Big general groups are mostly people who joined once and never came back, plus a healthy population of bots and resellers. The 400-member group where people argue about payment processors every day is a completely different proposition: everyone there has the problem, and they are talking about it in public.
Ten minutes of reading before you scrape will tell you which one you are looking at. Here is what to read for.
The simple rule
Spend ten minutes reading the group before you scrape it. You are looking for whether the conversations match a problem you solve, whether real people talk there or just bots posting links, and whether the members are the kind of people who decide things. Member count is the least useful number on the screen: a 400-person group of operators beats a 90,000-person group of lurkers almost every time.
The five things worth checking
None of these require tooling. Open the group, scroll, and read.
- Do the conversations match what you sell? Not the group's title, its actual messages. Plenty of groups named after a topic have long since drifted into something else.
- Are real people talking? Look for back-and-forth: someone asks, someone else answers, a third person disagrees. A wall of forwarded posts and links is not a community, it is a billboard.
- Are your people there? Read twenty or thirty profiles of people who actually post. If you cannot find your target role among them, scraping the other 89,970 will not conjure it up.
- Do they talk about problems, or just news? A group sharing price charts tells you nothing about who needs help. A group where people ask "how do you handle X" is showing you demand in real time.
- Would a message from you make sense here? If someone from this group received your opener, would the connection be obvious to them, or would they wonder how you found them?
The shortcut: the best sources are usually groups where people ask each other for help. Where there are questions, there are problems, and problems are what you sell into.
Three reasons to walk away, whatever else you found
Some groups fail on a single point, and no amount of other quality makes up for it.
- The rules say no. Plenty of groups explicitly forbid scraping members or DMing them. Ignoring that gets your account reported by people who were warned about you in advance.
- It is mostly bots and promo. If the visible activity is automated posts, copied content or obvious scams, the member list is full of the same. You will be messaging other marketers.
- You cannot tell who anyone is. If nobody has a bio, nobody posts, and there is no way to work out what people do, you have no basis for a relevant message. Volume will not fix that.
These are stop signs, not deductions. A group that trips one of them is out, even if it looks perfect on everything else.
Do not judge a group by one day
Groups have good days. A product launch, a market crash, or a viral post makes any community look alive for forty-eight hours.
Scroll back through a few different periods instead: this week, an ordinary week last month, and something from six months ago. You are checking whether the activity is a steady conversation or a spike followed by silence.
Watch for one more thing while you scroll: how many different people are actually talking. A group can post two hundred messages a day and still be four enthusiasts and a bot. That is not a lead source, it is a group chat.
A note on what you can actually see
Not every group will let you read its member list, and that is by design.
Telegram lets group admins restrict who can see participants, and large public groups often limit how much of the list is reachable. This is a normal platform behaviour, not something a tool can promise to work around.
In practice this means your best sources are public groups with visible membership and open discussion. When a group hides its members, the useful move is to participate in it and get known there, rather than to look for a way around the setting.
Find out afterwards whether you were right
The ten-minute read is a prediction. Campaign results tell you whether it was a good one.
Keep track of which group each contact came from, then look at where your good conversations actually originated. Almost every operator finds the same thing after a few campaigns: one or two sources produce most of the real opportunities, and several groups that looked promising produce nothing but polite confusion.
That list is worth more than any scorecard. Scrape more groups like the ones that worked, and stop revisiting the ones that did not, no matter how many members they have.
Judge sources by conversation quality, not reply count. A group producing ten replies where people understand why you contacted them is beating a group producing fifty "who is this?".
How we checked this guide
Reading a group tells you whether it is a plausible source. It does not settle whether contacting its members is appropriate in your market, and it does not override the group's own rules or local marketing law. Access to a member list is controlled by Telegram and by the group's admins, and technical access is not the same thing as permission.
Try it on your next group
Before your next scrape, run this on three candidate groups instead of one. The comparison usually exposes something uncomfortable: that you picked the first group by size, and the third one was better.
Source quality compounds. Everything downstream, your list, your message, your reply rate, your reputation, is capped by the group you chose in the first ten minutes.
Check the group, then keep the trail: TeleBoost's Telegram group scraper records which group each contact came from, so after a few campaigns you can see which sources actually produced conversations.
Keep learning