Scenario: Fastest page on your site loads in 300ms. Slowest is 1,500ms.

Have there been any studies on artificially slowing down the fast pages to match your slowest page? So that every single click is a consistent 1,500ms in the above case.

  • What are you trying to achieve with this? What benefit is there to users?
    – dnbrv
    Feb 13, 2015 at 18:39
  • 1
    Aside from the fact that I agree it is a bad idea, there is fundamentally no way for you to achieve a guaranteed constant load time. All the building blocks of the internet have no guarantee on timing. Your user could be on a slow connection. The DNS cache could have flushed. The OS could have decided to page the browser tab they are looking at. Or they could just be on a damn slow computer.
    – Bowen
    Feb 13, 2015 at 21:21
  • This is the same concept as Facebooks 60fps vs 30fps vs somewhere in between research. I am just wondering if anyone has published research on it. Feb 13, 2015 at 21:39

2 Answers 2


Don't slow down your page intentionally. Seriously... Don't.

The average expected loading time of a page needs to be no more than 1s and that in itself is too long. If a page doesn't appear to be loading instantaneously, people will leave because nothing is showing up.

Instead of attempting to get your fast pages to match your slow page, how about you do it the other way around... Where you try to get your slow page to match with your fast one.

It's very important to have websites (and the pages within) load quickly because then people will leave out of frustration.

  • 1
    That isn't my question though. I know the obvious, have a fast page. I am looking for research on my stated question. Feb 13, 2015 at 17:54
  • Why would someone do research on something that's known to be a UX flaw. It's like asking to see if anyone did research on whether having the navigation jumping around on different pages is a good idea because "they saw it once and they were intrigued by the idea." Just my take on this.
    – UXerUIer
    Feb 13, 2015 at 18:03
  • Hey maybe I'm wrong @ElijahLynn - maybe someone can put up an answer with the research described?
    – UXerUIer
    Feb 13, 2015 at 18:23
  • 1
    +1 for good UX advice when the actual answer to this question is There is no data on this because it isn't possible to make every page have a consistent load time
    – DaveAlger
    Feb 13, 2015 at 21:39

It is certainly possible to slow a site down but it isn't possible to make every page consistently take the same amount of time to load.

That said, here is the closest data I could find on how page loading speed affects users...

Neuromarketing, Relaxation and Web Design ...

Remember that people are happier to wait for a specific service that they have requested on an internal page – but slow-loading home pages are an absolute no-no.


There is nothing we can do about a slow connection speed – but always remember that applications on the web should be optimised to be as swift an easy as possible, even for Grandma on dialup.

Also, here is a pretty good info graphic that tries to put a bottom line price on slower loading pages.

How Loading Time Affects Your Bottom Line ...

website speed survey

  • Let's say that the delay was due to some database call. Using Go, I can think of at least one way that I could enforce a minimum response time of 1500ms. So I would have to say that it is possible to do. Whether it's a good idea though... That's a different issue.
    – JohnGB
    Feb 13, 2015 at 21:42
  • Well sure there are infinite ways to make your code less efficient and your application run slower. What's not possible is controlling the pipes of the internet in a way to ... "artificially slow down the fast pages to match your slowest page"
    – DaveAlger
    Feb 13, 2015 at 21:48
  • It depends on a number of factors. You can control how long it takes your server to respond, and in most cases that's the majority of the time. For example, I can ping almost anywhere in the world and get a sub 50ms response time (usually more like 25ms). If a server is giving me a 1500ms response, then it's reasonable to consider all the delay due to the server for all practical purposes.
    – JohnGB
    Feb 13, 2015 at 21:51
  • Okay well I found some data that is probably about as close to answering the initial question as possible.
    – DaveAlger
    Feb 13, 2015 at 22:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.