Lesson 09 · TikTok Live Academy
Tracking: from a viewer to FTD
This source has a step that cannot be measured directly: TikTok does not report how many people followed the link in your profile. This lesson counts that step anyway, with two things you control yourself: named Telegram invite links and a stream log.
01
Where the chain breaks
It is worth seeing what exactly is missing first — otherwise it is easy to spend a month hunting for a number that does not exist.
TikTok Live
- 01 A slot in the feed Live views
- 02 Viewers and watch time Viewers · watch time
Telegram
- 03 Out of the profile Link clicks → joins
- 04 Warm-up in the channel Retention D1 / D7
Product
- 05 Registration and first deposit Reg → FTD
- 06 Repeat deposits RD / LTV
TikTok reports the first two steps itself: views, viewers, watch time. The fourth, how many people stay in the channel, shows in Telegram's own statistics, and the partner dashboard counts the last two. The third — out of the profile and into the channel — is reported by nobody: there is one link in a profile and no report on it.
Which means you tell one stream from another by what Telegram counts and by your own stream log, since TikTok reports nothing on the one link in a profile. The rest of this lesson sets up both.
Watch outUntil joins are counted stream by stream, every conclusion about the source is a guess. “Streams do not work” and “the channel does not hold people” look identical when nothing counts the people who move from the stream to the channel.
The whole analytics chainthe streamviews and viewersthe profile and TGTG joinsDMsregistrationFTDDeposit $repeat depositsLTV
02
A Sub ID of your own for the source
Without one, TikTok blends into Threads, Instagram and organic Telegram, and the quality of the source becomes unreadable.
- A separate Sub ID for the sourceOne for all of TikTok: what you want is to tell it apart from your other sources.
- A second level per streamerAs soon as there are two of them: each streamer gets their own Sub ID, so the team's numbers do not blend. How to build the team is lesson 10.
- A campaign tagIf you run several channels or GEOs — so they do not collapse into one line of the report.
| What you are telling apart | By what | Where you read it |
|---|---|---|
| The source as a whole | By whatSub ID | Where you read itThe partner dashboard's report |
| A streamer | By whatThe second level of the Sub ID | Where you read itThe same report, broken down |
| One specific stream | By whatA named Telegram invite link | Where you read itThe link's own stats in Telegram |
| Time of day and delivery | By whatThe stream log | Where you read itYour own table |
The anatomy of a partner link and the postback setup are covered in the Affiliate Academy — all that is needed here is the outcome: TikTok arriving in the report as a line of its own.
For a single stream, a unique link or a day-and-slot marker: that way an evening launch does not get mixed up with a daytime one.
03
Named invite links
Named invite links close the gap in the chain: the step out of the profile and into the channel, which TikTok does not report. Telegram counts joins per link, and that is the only way to learn which stream brought people.
- One link per streamerOne per person, permanent. The joins Telegram counts on that link are that streamer's joins.
- One link per stream, when you need to compareNot always, but during a test week always: otherwise an evening stream cannot be told from an afternoon one.
- Name the link after what you are comparing“Tuesday evening” or “mask, EUR/USD”. In a month you will not remember which was which.
- The profile link points at the current oneWhen you move on to the next thing you are comparing, say from an evening stream to an afternoon one, put its link in the profile: from then on, joins are counted on that link.
It is a compromise, and its limit is worth knowing: an invite link counts joins, not clicks. You still will not see the difference between “tapped” and “joined” — but joining is the step all of this exists for.
04
The stream log
One row per stream. After thirty to fifty rows it starts showing what no report will: which time of day and which delivery bring you more joins.
| Column | What goes in it | What it is for |
|---|---|---|
| Date and start time | What goes in itAs it is | What it is forTime of day is the first thing to test: whether evening streams bring more joins than afternoon ones |
| Length | What goes in itIn minutes | What it is forHow length relates to reach (lesson 07) |
| Views | What goes in itFrom TikTok's analytics | What it is forThe top of the funnel |
| Peak viewers | What goes in itThe maximum for the stream | What it is forRetention rather than reach |
| Joins | What goes in itFrom the invite link's stats | What it is forThe one number joining the stream to the channel |
| What I did differently | What goes in itIn one phrase | What it is forWithout this column the log becomes a table with no conclusions in it |
The “What I did differently” column matters more than the four numeric ones: length, views, peak viewers and joins. Thirty rows of figures with no context answer “what happened” but not “what to do”.
05
Four ratios
Put in this week and last. The comparison is only ever against your own previous week — there is no industry figure here and there will not be one.
| Step | This week | Last week |
|---|---|---|
| Stream views | ||
| Viewers | ||
| Joins to the channel | ||
| Registrations | ||
| First deposits |
Ratios
- Join RateJoins / Viewers —Last week:
- Reg RateRegistrations / Joins —Last week:
- FTD RateFirst deposits / Registrations —Last week:
- Viewer → FTDFirst deposits / Viewers —Last week:
An empty field is not a zero: a row with no data stays a dash rather than becoming nought per cent.
Join Rate is yours and nobody else's: it measures how much reason the stream gives to move to the channel. Reg Rate and FTD Rate are no longer about the stream at all: the stream brings a viewer into the channel, and from there on the channel and the product do the work. That is the boundary lesson 01 drew.
The fourth is there so the first three do not fool you one at a time. A week where Join Rate rose and Viewer → FTD fell usually means the wrong people went to the channel.
Watch outA stream with a thousand views and three joins is worse than one with a hundred views and fifteen. Reading views apart from the ratios is the fastest way to optimise the wrong thing.
06
Stream and conversion KPIs
The four ratios above are the minimum. The full set falls into two groups: what happens on air, and how people move on from it.
The quality metrics — Deposit $, repeat deposits, LTV — are lesson 15, and FTD per Live Hour is lesson 11.
07
Mini-task: start counting
Half an hour once, and after that every stream adds one row.
Logging the first stream usually gets put off until the numbers grow. That is a mistake: the first rows of a log are the baseline every later week gets compared against.
0/5