Total Blocking Time: meaning, examples and practical use
Total Blocking Time, or TBT, is a laboratory performance metric that sums the portions of long main-thread tasks exceeding 50 milliseconds during the measured interval. It indicates blocking work that could delay responsiveness rather than measuring an actual user interaction delay.
Also known as: Total Blocking Time
At a glance
- A 120 ms task contributes 70 ms of blocking time.
- TBT is a lab diagnostic; INP examines actual interaction responsiveness.
- Smaller downloads alone do not prove less main-thread work.
On this page

How is it calculated?
For main-thread tasks longer than 50 ms, the portion beyond that threshold counts as blocking time. An illustrative 120 ms task contributes 70 ms, while a 40 ms task contributes nothing. The TBT documentation defines the measurement interval and aggregation.
Why does the main thread matter?
Many JavaScript and rendering tasks compete for the same browser resource. An uninterrupted long task can make handling a new user action wait. INP examines actual interaction latency, so it cannot be replaced by TBT.
Example: an article with many scripts
Imagine a page downloading HTML quickly, followed by several large scripts executing together. The text is visible, but its menu reacts late. Increasing transfer speed alone does not eliminate the processing burden. The amount downloaded and the work executed are related but different concerns.
What should be improved?
Identify the long task and its source. Removing unnecessary code, deferring work or breaking it into smaller tasks can help. Compare changes under consistent laboratory conditions, then examine real-user responsiveness too. A better lab score is useful evidence without proving every visitor's interaction is now fast.