Telegram Outreach Conversion Rate: How to Measure It and What Numbers Are Real

How to measure Telegram outreach conversion by stage: replies, MQL, SQL, deals. Formulas, a worked example, and five questions to ask about anyone’s "20% reply rate".

An outreach tool’s website says "20% reply rate". After your first week you’re at 3%. One of you is counting wrong, and often it’s both: the tool divides replies by messages sent in its best niche, while you divide by your whole list, including deleted accounts and people the message never reached.

Below: how to break outreach into stages, which formula to use for each, and why other people’s percentages almost never carry over to your funnel. At the end there’s a worked example you can repeat with your own numbers. If you source outreach leads yourself, Akula Marketing finds them by parsing groups and monitoring chats, and collects conversations from all your accounts in one inbox, so every funnel stage is visible without spreadsheets.

What is a normal conversion rate for Telegram outreach?

There isn’t one. We couldn’t find independent statistics on Telegram outreach, and the numbers on vendors’ websites differ several times over: one promises a 20% reply rate, another "15% replies, 8% warm leads and 40% conversion", a third shows 80% open rates. None of them says what the percentage is calculated from.

It’s more useful to compare a campaign with your own previous campaign. All that takes is counting the same way every time, using the formulas below.

Six funnel stages and a formula for each

Track absolute counts at every stage and measure conversion step by step. That shows exactly where people drop off.

StageWhat you measureFormula
DeliveredUnique people who received the messageThe base for every percentage
RepliedReply ratepeople who replied ÷ delivered
Positive replyShare of interest among repliespositive ÷ replied
MQLInterested people from your target audienceMQL ÷ delivered
SQLReady to discuss a purchaseSQL ÷ MQL
DealClose ratedeals ÷ SQL

End-to-end conversion is deals ÷ delivered. Its inverse, messages per deal, is the handiest number for planning: it tells you straight away how many contacts you need next month.

MQL vs SQL: where interest ends and a sale begins

An MQL (marketing qualified lead) is someone from your target audience who has shown interest: replied "tell me more", asked how it works, clicked the link in your second message. They aren’t planning to buy yet.

An SQL (sales qualified lead) is someone you can already talk to about a deal. They have a problem your product solves, they make or influence the decision, and they’re ready to discuss timing. Usually it shows in an action: they asked for the price, agreed to a call, or requested an invoice or a proposal.

Write the line between the two down before the campaign starts. Otherwise, on a good day a sales rep will call anyone who replied "interesting" an SQL, and MQL → SQL conversion will grow only on paper.

Five rules that keep your numbers honest

  1. Divide by delivered, not sent. Deleted accounts, closed DMs and spam blocks are a list or account problem, not a copy problem. Mix them in with silence from real people and you’ll start rewriting the message when the list is what needs fixing.
  2. Count people, not messages. A first message and two follow-ups to one person is one contact. A reply to any of the three touches is one reply.
  3. Attribute results to the campaign the lead came from. A November deal from an October campaign belongs to October. Otherwise a long sales cycle smears conversion across months, and your best campaign looks like a failure.
  4. Split replies by type. Positive, neutral, rejection, "stop messaging me". A 15% reply rate where 10 percentage points are unsubscribe requests is a reason to investigate, not celebrate.
  5. Exclude test and internal accounts. A colleague replying "test" inflates the reply rate of a small campaign more than you’d expect.

A word on "open rate". In Telegram you can see a message was read by the double check mark, but reading isn’t replying. Opens help you check delivery and say nothing about your copy or offer.

Worked example: how many messages it takes to close one deal

A hypothetical campaign; the numbers are for illustration, not a benchmark:

  • 1,000 people received the message;
  • 120 replied: a 12% reply rate;
  • 50 replies were positive: 42% of those who replied;
  • 40 MQLs: 4% of delivered;
  • 12 SQLs: 30% of MQLs;
  • 3 deals: 25% of SQLs.

End-to-end conversion is 0.3%, or about 333 delivered messages per deal. If you need 10 deals a month, plan for roughly 3,300 contacts. With the common advice to warm accounts up and start at 5–10 cold messages per account per day, that figure immediately tells you how many accounts and weeks you’ll need.

Add your costs (accounts, proxies, tools, a rep’s time) and you get cost per MQL, per SQL and per deal: campaign cost ÷ number of leads at that stage. That number is more honest than percentages, because you can compare it with other channels. How to turn it into CAC, LTV and payback per customer is covered in our unit economics glossary.

How to check someone else’s "20% reply rate"

Before you put someone else’s number into your plan, ask five questions:

  1. What was the percentage calculated from: messages sent, delivered or read?
  2. Are replies unique people or all incoming messages?
  3. Which niche? A number from Web3, where Telegram is the default work channel, says little about B2B services elsewhere.
  4. Do "replies" include rejections and unsubscribe requests?
  5. How many messages per day were sent from each account? A promise of "1,000 messages a day" sits badly with the warm-up and limits the same vendors recommend.

If three of the five have no answer, leave that number on the vendor’s website and keep it out of your plan.

Where to look when conversion is low

Find the step where the funnel drops the most:

  • Few delivered: a dead list or accounts in a spam block.
  • Low reply rate: targeting or the first message. We cover this in detail in 24 reasons prospects ignore cold outreach.
  • Many replies, few MQLs: you’re writing to the wrong audience. Narrow the list: the Telegram parser collects people who recently posted in niche chats, and chat monitoring catches "looking for", "need" and "any recommendations" messages 24/7.
  • Low MQL → SQL: qualification or the offer. People are interested but don’t see why they should buy now.
  • Low close rate: price, the sales process, or a slow follow-up after "let’s talk".

FAQ

Does a reaction count as a reply?

No. Count only text messages as replies, and if reactions matter to you, track them on a separate line.

How much volume do I need before the numbers mean anything?

With 50 contacts, one extra reply moves the reply rate by 2 percentage points. Draw conclusions about your copy after at least a few hundred delivered messages, and about close rate after your first few dozen SQLs.

Should I measure conversion separately for each account?

Yes, at least delivery and reply rate. If one account lags behind the others on the same copy, the problem is the account: it has been restricted or looks suspicious.

Find outreach leads and manage replies in one inbox →