If you have read our guide on judging a group before scraping it, this is the version you actually use. Same thinking, laid out as a sheet you fill in while reading a community.
The reason to write it down rather than judge in your head is comparison. Any group looks reasonable when it is the only one you are looking at. Put three side by side and one is usually clearly better, which is a decision you cannot make from memory a week later.
How to use this resource
Fill in one of these for each group you are considering, then compare them side by side. The comparison is the whole value: on its own every group looks plausible, and it is only when you put three next to each other that the differences become obvious. It takes about ten minutes per group and saves you from building a campaign on your first search result.
What to prepare
01A worksheet to copy per group.
02The two answers that end the evaluation immediately.
03How to estimate what you will actually get.
04What to add after the campaign runs.
The sheet
Copy this once per group. Square brackets are yours to fill in.
- Group: [name and link]
- Members: [the number, noted and then largely ignored]
- What they actually talk about: [in your words, after reading, not the group's description]
- Target roles seen: [how many of the ~25 profiles you checked looked like your buyer]
- Active people, not messages: [roughly how many different people posted this week]
- Rules on DMs and scraping: [what the pinned message or description says]
- Would my message make sense here: [yes / no / only for a narrower offer]
- Expected usable contacts: [members × the fraction that looked right in your sample]
- Decision: [scrape now / read another week / skip]
The fourth line does most of the work. Open twenty-five profiles of people who actually post, count how many look like your buyer, and you have a far better estimate than the member count gives you.
Two answers that end it early
If either of these comes up, stop filling in the sheet.
Everything else is a matter of degree, and a group can be mediocre on several lines and still be worth scraping. These two are not degrees, they are exits.
- The rules forbid it. Plenty of groups state that members must not be DMed or scraped. Doing it anyway means being reported by people who told you not to.
- You cannot tell who anyone is. No bios, no posts, nothing to read. Without something to go on, you have no basis for a relevant message and volume will not compensate.
Estimate the yield before you commit
The line most people leave blank, and the one that makes the sheet worth filling in.
If four of your twenty-five sampled profiles looked like genuine prospects, that is roughly 16%. Applied to an 8,000-member group, you might expect somewhere over a thousand usable contacts, and in practice fewer, because sampled posters are more readable than the silent majority.
The number is rough and it does not need to be precise. Its job is to make you compare: a 900-member group at 40% relevance is a better source than a 20,000-member group at 2%, and without writing it down you will pick the big one every time.
Halve your first estimate. Sampling the people who post overstates quality, because posters are more visible and more readable than the rest of the list.
Come back and finish the sheet
Two lines added after the campaign turn a worksheet into something that compounds.
After three or four groups, your predictions start being right, which is a genuinely useful skill. You will also spot the pattern in what makes a good source for your particular offer, and it is rarely the thing you expected before you started measuring.
- What you actually got: usable contacts after filtering, versus what you predicted.
- What the conversations were like: interested replies, and how many people were confused about why you contacted them.
Keep the filled-in sheets. A folder of them is a private map of your market that no competitor has, and it costs ten minutes per group to build.
How we checked this guide
The estimates here are deliberately rough, and the sample of profiles you can read is biased towards people who post. Treat the yield figure as a way of ranking sources against each other rather than a forecast, and remember that a group's rules and Telegram's own access limits govern what you can legitimately collect.
Put it to work
Fill this in for the next three groups you are considering, before scraping any of them. The exercise usually reveals that your instinctive first choice was the largest rather than the best.
Ten minutes per group, against a month of outreach into a community that was never going to work.
Keep the evaluation attached to the results
TeleBoost records which group each contact came from, so the yield line on your sheet can be checked against the conversations that source actually produced.
Keep learning