Measuring your TikTok LIVE — analytics

Vanity metrics versus operational metrics.

Peak concurrent viewers, total likes, total shares — these are the numbers TikTok shows on the celebratory screen at the end of a LIVE. They are useless for forecasting. A broadcast can hit 4,000 peak viewers because of a single FYP spike at minute 12, then collapse to 200 for the remaining 90 minutes; the peak number lies about audience quality. Operational metrics — average concurrent viewers, gifts per minute, gifter conversion, return-rate — describe what your community actually does over the broadcast. They predict Diamonds. The vanity numbers do not.

The five numbers that matter.

One: Average Concurrent Viewers (ACV) — total viewer-minutes divided by session length. Two: session length — the duration you held the FYP signal, not just the wall-clock minutes you streamed. Three: gifts per minute — total Diamonds received divided by session length. Four: gifter conversion rate — unique gifters divided by unique viewers, expressed as a percentage. Five: return-rate — viewers from last week who returned this week. These five together explain 80% of variance in monthly Diamond totals across the agency's roster.

Cohort retention is the only real loyalty signal.

Forget vanity follower counts. The metric that separates a community from an audience is cohort retention: of viewers who joined Monday's stream, how many returned Monday next week? Above 30% is a healthy community. Below 15% means you are constantly re-acquiring strangers and your gifter pipeline is shallow. The agency tags every creator with their 4-week trailing cohort return-rate; below 20% triggers a content review. The cohort is built by repeatable broadcast features — a Monday opener segment, a Wednesday PK slot, a Friday Multi-Guest format — not by hoping people remember to come back.

The gifter conversion ladder.

Inside the live-analytics view, treat every viewer as occupying one of four states: lurker → engager → gifter → whale. Lurker watches without interaction. Engager comments, follows, returns. Gifter sends recurring small Gifts (1-50 Diamonds per session). Whale sends concentrated value (1,000+ Diamonds per session, often single creators driving 30-60% of monthly Diamond totals). A healthy creator moves one viewer one rung up the ladder per week. Track the ladder by counting unique users in each state across a 30-day window — TikTok exposes the raw event log; the agency aggregates it inside BlackOS.

ARPGifter beats total Diamonds as the agency metric.

Total Diamonds tells you what happened, not why. ARPGifter — average revenue per gifter, measured in Diamonds — tells you whether your community is deepening or just widening. A creator going from 50K Diamonds with 200 gifters (ARPGifter 250) to 60K Diamonds with 600 gifters (ARPGifter 100) looks like growth on the surface but is actually a deterioration: each gifter is now contributing 60% less. The agency tracks ARPGifter monthly, by server, and flags any 15% month-over-month decline regardless of total Diamond movement. It catches problems before they show up in the bank account.

Diamond conversion benchmarks per server.

Healthy gifts-per-minute ranges, observed across the agency's roster in 2026: MENA 8-25 Diamonds/min for managed creators, FR+ 15-40, IT+ 12-30, DE+ 20-50, US 25-80. These are blended ranges including PK nights and standard broadcasts. A MENA creator running consistently below 5 Diamonds/min is operationally underwater; a US creator under 15 is the same problem. Cashback is always denominated in Diamonds and paid back into the creator's BlackOS wallet — the absolute USD value depends on TikTok's settlement which we do not control, so we manage the Diamond rate, not the dollar rate.

How to read your LIVE Analytics dashboard, step by step.

One: open the TikTok app, tap your profile, tap the three-line menu, tap LIVE Center, tap Analytics. Two: select the last 7 days as the date range — weekly is the right cadence, daily is noise, monthly is too late. Three: find Average Viewer Count (TikTok's label for ACV) on the overview card; record it. Four: find Total Diamonds and Total LIVE Time on the same card; divide one by the other to get gifts-per-minute. Five: scroll to the Audience tab to find New vs Returning Viewers — the returning percentage is your return-rate proxy. Six: scroll to the Engagement tab for unique gifter count. Seven: enter all five numbers into a tracking sheet (or BlackOS if managed). Eight: repeat every Monday.

The agency's weekly creator review agenda.

Every Monday at 10:00 local time, the manager runs a 20-minute review per assigned creator. Step one: compare ACV this week vs the trailing 4-week average. Step two: compare gifts-per-minute against the server benchmark. Step three: pull the gifter conversion ladder counts — how many lurkers moved to engager, how many engagers to gifter. Step four: name the top 3 gifters by Diamonds contributed this week and the bottom 3 gifters who lapsed. Step five: identify the single broadcast feature that performed best. Step six: assign one experiment for the week — new opener, new PK partner, new schedule slot — with a numeric success criterion. The review ends with one written next-step in BlackOS.

Diagnostic: why your ACV dropped this week.

If ACV fell more than 20% week-over-week, run this in order. One: did session length shrink? An ACV drop from a shorter broadcast can be a planning issue, not an audience issue. Two: did the schedule shift? Moving a Monday stream to Tuesday breaks cohort return mechanically — give it 3 weeks before re-baselining. Three: did the opener change? Audit minute 0-5 retention — if first-minute retention dropped, your FYP signal is weaker and ACV follows. Four: check for platform-wide events (Ramadan opening, major sports finals, election nights) that pulled audience away. Five: check competitor activity — a same-server creator running a multi-hour PK tournament steals viewers from adjacent broadcasts. Six: if none apply, the content itself has gone stale — schedule a content reset week with the manager.

When to declare a LIVE bad and pull out.

Some broadcasts are not recoverable in-session, and the agency teaches creators to recognize the signal early. Pull-out triggers: ACV is below 30% of trailing average at minute 20; gifts-per-minute is zero at minute 30 with chat active (it means you have an audience but the gifter conversion is broken — wrong tone, wrong segment); FYP entry rate flatlines after a normal opener (algorithm is not pushing — close out clean, do not extend hoping for recovery). A pulled-out LIVE is recorded as such in BlackOS and excluded from the trailing-average baseline so it does not poison next week's diagnostic.

Inside Black Ads Agency, operational LIVE metrics are tracked per creator inside BlackOS (iOS app id6767493180) across the five-server stack — MENA (faction 108135), FR (115414), IT (117633), DE (120935) and US (128508). Average concurrent viewers, gifts per minute, gifter conversion and return-rate are baselined in the first 30 days of manager pairing and benchmarked against server-level cohorts so a Tunis-resident creator on MENA is compared to peers on the same 20:00–01:00 evening window, not to a Paris-resident creator on FR prime time. The agency reached Senior TikTok LIVE Partner status in May 2025; managed creators keep 100% of their Diamond payouts at 0% commission, with 1 Diamond ≈ USD 0.005 net.