Instagram Live analytics: viewers vs peak viewers

Instagram Live's Viewers metric counts unique accounts that saw your content during the live session. Peak concurrent viewers is the largest audience watching at the same time. Use both when reviewing a broadcast: they describe total attendance and the busiest moment, but neither tells you how long people stayed.

What the Live metrics measure

Instagram's Live insights guide distinguishes Viewers, Peak concurrent viewers and Accounts engaged. Accounts engaged counts unique accounts that interacted during the live video. Comments and Shares count actions, so several comments can come from the same person.

Keep the metric names from your source in your report. If an older export says Accounts reached, check its definition and scope before putting it in a column labeled Viewers. A change of label alone does not establish that two historical figures are comparable.

For a previous broadcast, Instagram directs you to your dashboard, then Content you shared, where you can choose the live video. Live insights are available in the mobile app for up to 30 days after the broadcast ends. Save the figures with the broadcast's date and the time you collected them.

A bigger peak can accompany a smaller total audience

Consider these invented results for two product demonstrations. Both ran for the same length of time, and all the counts describe the live session.

Metric Demonstration A Demonstration B
Broadcast length 30 minutes 30 minutes
Viewers 800 650
Peak concurrent viewers 120 160
Accounts engaged 50 60
Comments 90 72

Demonstration B reached fewer distinct accounts, yet its busiest moment had more people watching together. Its higher peak does not prove that it held a larger audience for most of the broadcast. A short burst could produce that result.

B also had more accounts interacting despite receiving fewer comments. The comment total alone would miss that difference. Read the questions before choosing what to change: questions about buying the product have a different implication from repeated requests to fix the sound.

Neither set of numbers proves that the host, topic or start time caused the result. Record those differences so you can choose one to test in a follow-up broadcast.

Why peak divided by viewers is not a retention rate

For Demonstration A, 120 / 800 × 100 = 15%. That calculation compares the maximum simultaneous audience with everyone who appeared during the session. It does not tell you that 15% stayed to the end, or that 85% left early.

The people watching at the peak might have arrived at different times. Later arrivals could replace earlier viewers. To establish how long people stayed, you need data on viewing duration. A chart of the audience over time can show changes in the total watching, but it cannot identify whether the same people stayed. The peak and total alone provide neither kind of detail.

Peak concurrency also cannot give you average watch time. Keep any available watch-time metric under its own name, and mark it as unavailable if your source does not provide it. The guide to Reels average watch time explains a separate video metric; applying its formula to a Live audience count would mix different units.

Compare broadcasts with the same purpose

Before your next live session, write down what you want to learn. A demonstration might need to answer common product questions. A launch might aim to get viewers to a booking page. Choose an outcome you can actually record, then keep it beside the audience figures.

Record the topic, host or guests, start time and broadcast length. Note how you promoted the event and whether the offer changed. A long interview and a short product announcement give people different reasons and opportunities to join.

For the next test, you could repeat a demonstration with the same host and offer but bring the product demonstration forward in the running order. Compare the questions it produces and the outcome you chose. Repeat a promising result before treating the format as a dependable improvement.

Keep the replay separate in your report

A replay or a clip published after the event gives people another opportunity to see the content. Record those results on separate rows, with their publication dates and measurement cutoffs. Adding unique audience counts across the live session and a later publication can count the same account twice.

If the event is meant to send people to your website, use a distinct tagged link for each placement you want to compare. The Instagram UTM guide covers how to label those visits. Attribute results to the tracked placement; the appearance of a booking after a broadcast does not by itself prove that the broadcast caused it.

In your monthly Instagram report, keep the saved Live insights beside a short note about the content and its outcome. Write down the next change you plan to test and the result you will measure before that broadcast starts.