How to read Instagram and TikTok analytics: a plain-English metric glossary with worked examples

Open Instagram Insights or TikTok Studio for the first time and you meet a wall of labels: views, viewers, accounts engaged, profile activity, average watch time, watched full video, traffic source types. Some sound alike but measure different things, some were renamed recently, and a few disappear entirely below a certain account size. If you run a small brand, say a candle shop in Manchester or a ceramics studio in Lyon, you do not need every chart. You need to know which few numbers tell you whether your content reaches the people who might buy from you.

This guide translates each metric into plain English using the platforms' own definitions where they publish them, shows how to compute engagement rate with illustrative numbers, explains how to check that your audience lives where your customers do, and ends with a weekly review template. The businesses in the examples are invented illustrations, not customer stories.

Where to find the dashboards and which timeframe to pick

On Instagram, any public account can open free insights. The app's overview groups Views, Interactions, New followers and Content you shared, and each tile opens a detailed screen with an (i) icon that shows the official definition ^[1]. You can pick a preset or custom timeframe, but only within the past 90 days ^[1], so export or note anything you want to keep longer.

On TikTok, personal accounts read analytics in TikTok Studio, while business accounts use Business Suite; both let you change the date range, and native data goes back up to a year ^[5]. Every video also has its own analytics screen, which is where most of the useful detail lives.

Instagram metrics in plain English

The first five definitions come from Instagram's help page ^[1]; profile activity is defined in Meta's developer reference ^[3]. Instagram labels views and viewers as estimated and still in development ^[1], so treat small week-to-week moves as noise.

Views, viewers and the end of impressions

If you learned Instagram analytics a few years ago, you remember impressions and accounts reached. Both have been reshaped. Meta's developer documentation states that impressions were deprecated as of 21 April 2025 and should be replaced with views ^[2], and at the post level impressions stopped being reported for media created after 2 July 2024 ^[3]. In the app, views now headline the overview ^[1].

The practical difference: views count every time content appears or plays, so one person watching a reel three times adds three. Viewers count that person once ^[1]. For a reel, a viewer is counted whether or not the reel actually played ^[1]. When someone asks what reach means on Instagram today, the honest answer is that the closest app label is viewers.

Worked example: a Brooklyn bakery reel

Suppose a bakery's weekend reel shows a views figure of 5100, a viewers figure of 3400 and 190 interactions. Views divided by viewers is 1.5, so on average each person saw it one and a half times, a sign people rewatched or met it in more than one place. That ratio is more informative than either total on its own.

TikTok metrics in plain English

TikTok's own tutorial describes total views as the number of times your videos were viewed and profile views as the number of times your profile was viewed, and it lists watch time, average watch time and audience traffic for each video ^[4]. More recent labels are described in third-party guides that follow the current app ^[5]:

Worked example: a Berlin fitness coach

A coach posts a 30-second tutorial. Average watch time is 12 seconds and watched full video is 18 percent. Twelve seconds out of thirty is 40 percent, so most people leave before the key move at second 20. The fix is editorial, not statistical: show the move first and explain afterwards, then compare the next video's percentage.

How to calculate engagement rate (and what counts as good)

Neither Instagram nor TikTok shows a single official engagement rate on the account dashboard, and the inputs vary between platforms and analytics tools ^[7]. Three conventions are common; pick one and keep it:

• By followers: engagements divided by follower count, times 100 ^[7]. Easy to compare over time, but punishes accounts with many inactive followers.
• By viewers (reach): interactions divided by viewers, times 100. Tells you how people who actually saw the post responded.
• By views or impressions: engagements divided by views, times 100; the impressions version is typical for paid or boosted posts ^[7].

Back to the bakery: 190 interactions over a viewers figure of 3400 is about 5.6 percent by reach. If the account's follower count is 2600, the same post scores 7.3 percent by followers, because non-followers also saw it. Same post, two very different rates, which is why the formula must travel with the number.

What counts as good? Published benchmarks differ widely by industry, account size and formula, so the useful benchmark is your own trailing average. A post beating your four-week median by a clear margin is good; a post far below it is worth examining.

Does your audience live where your customers do?

Instagram shows top countries, top cities, age ranges and gender for both viewers and accounts engaged, but only with preset timeframes and once a post or period has reached or engaged more than 100 accounts; if the last 7 days are empty, try 30 or 90 days ^[1]. Meta also notes that demographic lists stop at the top 45 entries and that their totals can fall short of your follower count, because only people with demographic data are counted ^[2].

1. Write down where your paying customers are: shipping addresses, booking locations or the cities you deliver to.
2. Compare that list with top countries and top cities for accounts engaged, not just followers.
3. Flag any market that brings attention but cannot buy, for example a UK-only candle shop with a large share of engagement in the US.

Illustration: a Manchester candle shop that ships only within the UK finds 58 percent of engaged accounts in the UK and 21 percent in the US. Rather than chase more attention overall, it starts posting at UK evening hours and mentions UK delivery in captions. Our guide to posting times for audiences in the US, UK and Europe shows how to schedule for the market you actually serve.

What bought follows, likes and views do and do not change

It helps to be precise. A purchased package raises the counted number it targets: the follower count, the likes on a post or the views on a video. That is all it does. It does not add saves, shares, profile activity or link clicks from people who want your product, and it does not create customers.

• Engagement rate by followers drops when the follower count grows without matching interactions, so record your rate before and after if you use one.
• Location and age reports can shift, which makes market-fit checks harder; read accounts engaged rather than follower demographics.
• TikTok's guidelines prohibit trading services that artificially boost engagement and say the platform removes fake likes, followers or other inflated signals it detects ^[6].

If you are deciding which counted number, if any, fits your goal, see followers, likes or views by goal, and for the wider trust picture read the small-business social proof playbook.

A 15-minute weekly review template

Block a quarter of an hour on the same day each week, open the last seven days, and fill one row per question in a simple sheet:

One change per week keeps cause and effect readable. After a month you will have four comparable rows, which is enough to see a direction without over-reading any single post.