Raygun Feature Request

Feature Request

Define proper load times per page

Current Status:

New

Votes:

8


Avatar

mcartwright

Many of our application's pages aren't designed to load in less than 1 second. The end result is that the performance always looks bad on those pages (which is invalid). You should be able to customize the allowed time to load bands. I.e. Excellent is under 3000ms, Good is between 3000ms and 5000ms, needs work is between 5000ms and 8000ms and poor is over 8000ms. The current/static health scores don't account for real work applications.


Avatar

Raygun

John-Daniel Trask

Posted on
Dec 21 2018

Thanks for this request.

We've debated this for a while, and we think we will add something here in future as it is a semi-common request.

However...

The reason for our slowness on this, is that it doesn't feel like the right move for your users. No user is going to care if a developer thinks 10s is an acceptable load time generally.

Our experience score isn't based on what Raygun thinks is a great score either, but rather off the research done by companies like Google and others that measure user satisfaction with load times. This measures when users will be frustrated - the 'why' isn't a relevant factor in satisfaction - if it's slow, it's slow, and that's not good.

Celebrating a "Good" score where you moved the goal posts to say a long time is "good" isn't really the intention of a product that measures user experience. Fast software is good software.

John-Daniel Trask
Co-founder
Raygun

Avatar

ncrown

Posted on
Dec 21 2018

I think the key to this is that B2C is much different than B2B.

Avatar

Raygun

John-Daniel Trask

Posted on
Dec 21 2018

Can you provide more insight into that?

Business users are the same people as consumers - their satisfaction doesn't change when they're at work does it? I still get grumpy when something is slow, and I totally understand when our own users get annoyed when our software gets slow - it directly impacts our conversion rates, satisfaction scores etc.

There are a few things that are outliers that I can think of, such as file uploads. No amount of optimization is going to make that request go super fast if the file is huge and the pipe is small. But outside of things like that, when would long load times be acceptable?

Avatar

ncrown

Posted on
Dec 21 2018

If you are going to make the point that a business user working in a web-based ERP system should expect the same experience they have when viewing the news feed in a social media platform, then you have not spent enough time in IT to understand the argument.

Avatar

Raygun

John-Daniel Trask

Posted on
Dec 21 2018

I have used such systems - and no, they're not "good".

As mentioned, we'll probably add this feature sometime next year for various reasons, but I do think it's lying to ourselves that it's a good experience. That user isn't having a good experience.

Avatar

ncrown

Posted on
Dec 21 2018

Good point.

It's hard to make a crappy job "good," but providing a solution for a user that might take a few seconds today that in the past took them several minutes, which included looking at paper-based forms, making phone calls, etc. is still a much better solution.

As you know, it is relative, and I get the argument that B2B providers should work to provide experiences that are similar to those that are achievable in the B2C world. That said, there are certain functions (e.g. load optimization in the transportation world) that can easily span multiple seconds in page load time. Again, doesn't mean that we as B2B providers should settle, as there are all sorts of cute tricks you can play to be more responsive and hide the real, behind-the-scenes, data crunching.

Appreciate the consideration...

Avatar

mcartwright

Posted on
Dec 21 2018

Interesting perspective that some would necessarily (and without context) equate speed with good (I don't think that is a good approach)? By way of example, if i return a web page quickly but it doesn't contain ALL or much of the information the user wants, I wouldn't think that would be a good experience?

Also, i'd arguing that we aren't trying to "move" the goal posts, we are trying to properly set the goal posts to the correct starting point and then detect deviation and quickly/accurately report on said deviation.

www.raygun.com takes approx 3 seconds to get initial load (87 requests 2.87 MB / 1.02 MB transferred Finish: 25.38 s DOMContentLoaded: 2.41 s load: 9.88 s).

You could certainly return only your logo much faster.. But i much prefer to wait the 3 seconds to get the rest of the information this is super interesting. :)