Time to Connect metric in network performance measurements

Question

Time to Interact appears to be a new metric used to measure the perceived speed of a web page. I am interested to know a little more about what it really is.

This term, apparently, was coined by Radware and is pushed to the most significant measure of performance (compared to things like Time to First / Last Byte, Time to Render, etc.).

It is described as:

the point on which the page displays its main interactive (I think clickable) content, and not the full page load.

This seems very subjective to me; What is the "main interactive content" of a web page?

There are messages with a link to the measurement results, so some of them are measured, and, in addition, it should be automated, since the result sets are quite large (~ 500 sites were checked).

Apart from the quote above, I cannot find any more information on how to measure it.

Since Google pays more attention to the content of the warehouse (or the visible content), I wonder if this metric really looks more like "Time to the first meaningful render", that is, it is contextual for the current purpose of the page. So, for example, on the product page of an eCommerce site, this can be the main image and a link to the basket.

I really want to understand this metric, since it seems to me that it is the most useful. So my question is, does someone measure this, and if so, how do they do it?

Answer 1

You answered your question, it is subjective and contextual for you in the current project.

What if I test the site only with HTML without any complex resources? There is no TTI measurement point. On the other hand, check out this demo site .

The big picture is here

The blue line marks the event "COMContentLoaded" (the main document is loaded and the markup is parsed), the red line indicates the load event, where all page resources are loaded. The TTI line will be between two lines, which is determined differently for each project, based on some resources that are important for the interaction of loaded events.

For example, let's say that the images on the demo site are not essential for the basic functions of the site. While the main site loaded in 0.8 seconds, 3 large snapshots took 36 additional seconds to load, so in this case using the total response time as KPI will give a response time of 36 seconds, whereas if you define TTI, excluding those large , resources, you end with <1 s.

I really want to understand this metric, because it seems to me that it is the most useful.

Definitely useful, but as you said in your question, this is project specific. You will not measure TTI on a simple, relatively static web application, you probably would measure the overall response time. I always define KPI “adapted” for the current project, instead of trying to use common indicators and “force them” in the project.

So my question is, does anyone measure this, and if so, how do they do it?

Having definitely used it before, you should define the main resources for your site, and when the last of these resources loads, this is your TTI. It could be a javascript, css file, etc.

Answer 2

Web sites are becoming more complex. While they may not always contain more content, they still have more resources to download as the user interaction / user interface is technically more complex. Ajax helps us load different parts separately. Therefore, instead of loading a single page, we load a few small things. And for each of these parts, we can measure download performance. But the site may have some details that may be more important than others. “Main interactive content” is that part of your presentation that allows the user to do what he intends to do, for example, buy a train ticket. If some advertising or special animation on the left side of the screen is not loaded, this does not prevent the buyer from starting to buy a ticket. But, of course, “primary interactive content” as a term is rather vague, and you should define it for your specific application. This is the moment when the average user can and will start interacting with the website, while some parts are loading the threshold.

This is how I understand the concept, and here I see the difference with "Time to the first meaningful render": you may have a basket displayed on your e-commerce page, but the graphical interface does not respond yet. So, you see something meaningful, but interactivity is not there yet. Therefore, TTI> = TtFMR.

TTI measurement requires you to determine which elements are necessary for interactivity, which depends not only on what the site is doing, but also on how it is done. It depends a lot on your implementation / technology.