Mostly his advice is good but his conclusion that knowing that I’m seeing records 1-100 of 9001 is not useful is wrong in many cases. In my Gmail inbox for example, if I let it go for a couple days it’s no longer easy to know how many I’m looking at on the page or how many messages total are floating around in that folder. Giving that kind of count right by the page navigation is a) expected behavior and b) useful information. There’s a big difference in how long it’s gonna take me to clean up if it’s 150 emails vs. 520.
It’s also quite useful as feedback. I’m on page 7. I click next. Am I on the next page or is my browser reconstructing the universe from first principles off in some shadow dom? (Does it now say page 8?)
That is not a “user experience problem”, it’s the author’s personal opinion. I personally prefer the “Page x out of y” or “1-100 out of XXXX” and dislike most pages that use “timelines”.
If he doesn’t like the numbers he can ignore them and hit “Next” and “Previous” a million times till he finds what he wants. I appreciate being able to skip large chunks of results and the feedback of knowing how many pages I may have to look through still.
There’s still a good use case for the pagination that the author dislikes when the items being listed are things you have to pay for. I’m thinking it makes sense for showing how usage data (like phone calls) are billed. I really wouldn’t want infinite scrolling for that.
The author claims nobody ever wants to go to the last page. That’s just plain wrong. I frequently find myself wanting to go to the last page for various reasons - maybe to circumvent design choices (I can sort by upvotes, but only in descending order? I’ll just go to the last page then), or to skip most of the discussion in a forum if I’m mostly concerned about the conclusion. Claiming that skipping to the last page is useless is just asinine.
In blogs you might to go to the last page which is the oldest content and then read back in order, this is especially good if there is a 3 part post for example as you can open them in 3 tabs in the correct order (And yes, I am a tab hoarder).
I wish people would stop going on about this. “Page x out of y” is not complicated UI. Removing it isn’t going to make things easier for users, and it most likely won’t affect performance, either. The key to better UI isn’t to dumb everything down and provide as little information as possible.
Can anybody point to real users complaining about pagination?
I agree with you that the article doesn’t properly substantiate its point, but I want to note that the way to get an answer to a question about whether users care is to A/B test.
What users say is often out of line with how they actually use software (in particular, change-aversion makes it very hard to get anything useful out of written feedback). Most users aren’t going to know how to explain something programmer-y like pagination techniques, or meaningfully compare the actual thing in front of them with the hypothetical one. And complaining about software tends to be shouting into the void anyway, so a lot of people are just going to express generalized sentiments about the whole product as a proxy for talking about the specific issues that bother them.
Of course, this author makes no claim to have tested it, and it’s unfair to hold them to a standard they aren’t trying for. But I would classify this article as one of the time someone actually does complain about minor UI things. Its ideas are interesting, but it needs data.
And some more experimental evidence that “infinite scroll” increases amount of content consumed (if that’s your goal), “load more” buttons decreased the amount of paging people did (which means they will make more use of other navigation means) and traditional pagination buttons were rarely used outside of next/prev:
https://www.smashingmagazine.com/2016/03/pagination-infinite-scrolling-load-more-buttons/
A little more googling shows lots of other experimental results of various trustworthiness.
I myself have actually watched a user in an ad-hoc usability test click “next page” 50+ times when they could have found what they wanted with one click on the filter bar.
As for the performance problem: That’s a totally separate issue from pagination. You can still do page numbers and total record counts with key-set pagination. That is, you don’t use OFFSET, you use an inequality in your WHERE clause. You can still use a LIMIT (or TOP). In fact, this is one of the reasons that Microsoft SQL Server didn’t offer OFFSET for many years, but they finally caved to customer requests. My guess is because so many awful tutorials on the internet suggest using LIMIT/OFFSET for pagination.
Mostly his advice is good but his conclusion that knowing that I’m seeing records 1-100 of 9001 is not useful is wrong in many cases. In my Gmail inbox for example, if I let it go for a couple days it’s no longer easy to know how many I’m looking at on the page or how many messages total are floating around in that folder. Giving that kind of count right by the page navigation is a) expected behavior and b) useful information. There’s a big difference in how long it’s gonna take me to clean up if it’s 150 emails vs. 520.
It’s also quite useful as feedback. I’m on page 7. I click next. Am I on the next page or is my browser reconstructing the universe from first principles off in some shadow dom? (Does it now say page 8?)
That is not a “user experience problem”, it’s the author’s personal opinion. I personally prefer the “Page x out of y” or “1-100 out of XXXX” and dislike most pages that use “timelines”.
If he doesn’t like the numbers he can ignore them and hit “Next” and “Previous” a million times till he finds what he wants. I appreciate being able to skip large chunks of results and the feedback of knowing how many pages I may have to look through still.
There’s still a good use case for the pagination that the author dislikes when the items being listed are things you have to pay for. I’m thinking it makes sense for showing how usage data (like phone calls) are billed. I really wouldn’t want infinite scrolling for that.
Infinite scrolling for more than a couple of pages of stuff is pretty much the worst (Hi Youtube, Twitter, and Facebook!).
The author claims nobody ever wants to go to the last page. That’s just plain wrong. I frequently find myself wanting to go to the last page for various reasons - maybe to circumvent design choices (I can sort by upvotes, but only in descending order? I’ll just go to the last page then), or to skip most of the discussion in a forum if I’m mostly concerned about the conclusion. Claiming that skipping to the last page is useless is just asinine.
In blogs you might to go to the last page which is the oldest content and then read back in order, this is especially good if there is a 3 part post for example as you can open them in 3 tabs in the correct order (And yes, I am a tab hoarder).
This article buries the lead a bit, but the conclusion is strong:
I wish people would stop going on about this. “Page x out of y” is not complicated UI. Removing it isn’t going to make things easier for users, and it most likely won’t affect performance, either. The key to better UI isn’t to dumb everything down and provide as little information as possible.
Can anybody point to real users complaining about pagination?
I agree with you that the article doesn’t properly substantiate its point, but I want to note that the way to get an answer to a question about whether users care is to A/B test.
What users say is often out of line with how they actually use software (in particular, change-aversion makes it very hard to get anything useful out of written feedback). Most users aren’t going to know how to explain something programmer-y like pagination techniques, or meaningfully compare the actual thing in front of them with the hypothetical one. And complaining about software tends to be shouting into the void anyway, so a lot of people are just going to express generalized sentiments about the whole product as a proxy for talking about the specific issues that bother them.
Of course, this author makes no claim to have tested it, and it’s unfair to hold them to a standard they aren’t trying for. But I would classify this article as one of the time someone actually does complain about minor UI things. Its ideas are interesting, but it needs data.
Obviously, it depends on your application and goals.
There’s some experimental evidence that pagination slows visual scanning: http://usabilitynews.org/the-impact-of-paging-vs-scrolling-on-reading-online-text-passages/
And some more experimental evidence that “infinite scroll” increases amount of content consumed (if that’s your goal), “load more” buttons decreased the amount of paging people did (which means they will make more use of other navigation means) and traditional pagination buttons were rarely used outside of next/prev: https://www.smashingmagazine.com/2016/03/pagination-infinite-scrolling-load-more-buttons/
A little more googling shows lots of other experimental results of various trustworthiness.
I myself have actually watched a user in an ad-hoc usability test click “next page” 50+ times when they could have found what they wanted with one click on the filter bar.
As for the performance problem: That’s a totally separate issue from pagination. You can still do page numbers and total record counts with key-set pagination. That is, you don’t use OFFSET, you use an inequality in your WHERE clause. You can still use a LIMIT (or TOP). In fact, this is one of the reasons that Microsoft SQL Server didn’t offer OFFSET for many years, but they finally caved to customer requests. My guess is because so many awful tutorials on the internet suggest using LIMIT/OFFSET for pagination.