YouTube FTA analytics
Audience intelligence for free-to-air TV, measured live.
LiveFTA turns the YouTube simulcasts of East African broadcast television into a continuous, comparable record of who is watching, when, and what they are saying about it.
- channels tracked
- 15
- markets
- 3
- collection
- 24/7
- native timebase
- EAT
Channels tracked, by market
Kenya
- Citizen TV Kenya
- NTV Kenya
- KTN News Kenya
- KBC Channel 1
- TV47 Kenya
Uganda
- NTV Uganda
- NBS TV
- UBC TV
- Bukedde TV1
- BBS Terefayina
Tanzania
- ITV Tanzania
- TBC1
- Clouds TV
- Star TV
- Wasafi TV
What is on air right now, and who is watching it.
Every tracked stream is read every 10 minutes in prime time and hourly off-peak. Figures stay within their market and run on East Africa Time.
Last poll 20:40 EAT, prime-time cadence
- Live viewers, Kenya
- 60,130 across streaming channels
- Peak channel tonight
- Citizen TV Kenya 27,480 viewers at 19:40 EAT
- Live chat, last hour
- 4,210 messages collected
- Channels streaming
- 5 of 5 every tracked Kenyan channel is live
-
Citizen TV Kenya
Live24,610
concurrent viewers
-
NTV Kenya
Live13,820
concurrent viewers
-
KTN News Kenya
Live11,240
concurrent viewers
-
TV47 Kenya
Live6,070
concurrent viewers
-
KBC Channel 1
Live4,390
concurrent viewers
Sample figures for illustration. The live dashboard updates on the same schedule.
Introducing LiveFTA
Four stages, running continuously.
-
1
Detect
Find out whether each channel is live right now, and which broadcast it is streaming.
-
2
Capture
Record concurrent viewers and pull new live chat messages since the last read.
-
3
Interpret
Score each chat message for sentiment, then write a short account of what drove it.
How sentiment is scored -
4
Deliver
Serve a live dashboard, and export the same data as reports and extracts.
See the report builder
Audience size says how many were watching. Live chat says how they felt about it.
Each live chat message is scored positive, neutral or negative as it is collected. Hour by hour, the balance shows when a broadcast held its audience's goodwill and when it lost it.
One number, always with its sample
Net sentiment is the share of positive messages minus the share of negative ones. It is always shown next to the number of messages behind it, and flagged as a low sample below 30.
Described, never quoted
The written account paraphrases what drove the mood. It never quotes or names an individual commenter. When an hour has fewer than 15 scored messages, LiveFTA says there is not enough chat to characterise rather than guessing.
Live chat sentiment
Kenya, one evening bulletin, 17:00 to 23:59 EAT. Sample data.
+31 net sentiment, 821 messages
Neutral
Chat warmed through the evening and was most positive in the 21:00 hour. Negative messages peaked at 20:00. There was not enough chat at 17:00 to characterise.
Positive: reaction to the sports highlights segment
Negative: disagreement during a studio panel discussion
Neutral: greetings and viewers naming where they were watching from
Two output formats for two audiences.
A written executive document for the people who make decisions, and raw extracts for the people who model them.
The preview is the document
Set the market, stations, period and daypart once. The preview updates as you change them and shows exactly what you are about to download.
Extracts for your own models
The same scope exports as CSV, one file per analysis, with every reading timestamped in East Africa Time.
Watch a report build itselfExport
Executive Audience Report
YouTube Live Audience & Ad-Placement Brief
Kenya, 1 to 14 September 2026
Executive Summary
Prime time carried the largest average audience on every weekday. Weekend audiences peaked later, closer to 21:00 EAT.
Exhibit 1 · Audience trend
Exhibit 2 · Daypart audience
Preview with sample data.
For the people who plan around broadcast audiences.
Agencies and buyers
Bring evidence to a placement decision. Compare dayparts and stations on average and peak audience, on share of voice, and on how dependable each station's numbers actually are.
The standard report is written as an Ad-Placement Brief.
Broadcasters
See your own bulletins against your competitors' in the same hour. Know which slots are strengthening and which are quietly losing ground, and read what the audience said while it happened.
Researchers
A continuous, timestamped record of regional broadcast audience and public reaction, exportable as clean tables for whatever question you brought to it.
How it is built to be trusted.
Arithmetic and interpretation stay separate
Every figure comes from a query. Language models only describe figures they were handed. No number in LiveFTA has ever been produced by a model.
Silence over invention
Too little data produces an explicit "not enough data" state, at every level, rather than a confident summary of almost nothing.
Markets are never averaged together
Kenya, Uganda and Tanzania are analysed separately by design.
Failure is visible, not silent
Every collection attempt is recorded with its outcome. A failed poll is distinguishable from a channel that was not broadcasting.
History is permanent, at the right resolution
Viewership readings are kept indefinitely. Narrative detail is condensed on a schedule rather than growing without limit.
Access is controlled
The platform sits behind authentication, with verified email and single-use sign-in links.
The broadcast happened. The audience was there. LiveFTA is the part that remembers.