Topic: Request: condense Upvote and Favorite functionality into one concept

Posted under Site Bug Reports & Feature Requests

This topic has been locked.

Problem: many very popular posts in terms of # of Favorites receive a small fraction of that same number in Upvotes, relegating them to not display on the Popular page. This presents a distorted view of what's popular on e621 to the userbase.

Solution: take the approach nearly every other social media website uses, by providing just one "upvote/favorite/heart/star" functionality. Why have separate functions? It's clearly not being used effectively.

Updated by bitWolfy

hhrff said:
Problem: many very popular posts in terms of # of Favorites receive a small fraction of that same number in Upvotes, relegating them to not display on the Popular page. This presents a distorted view of what's popular on e621 to the userbase.

Solution: take the approach nearly every other social media website uses, by providing just one "upvote/favorite/heart/star" functionality. Why have separate functions? It's clearly not being used effectively.

I like having separate upvote/favorite options. They can be searched separately, which is convenient for organization. Upvotes can be cancelled out by downvotes, while favorites don't have that. Favorites are easily visible to other users while upvotes aren't.

Using upvotes to measure popularity makes sense to me since it accounts for downvotes. If an image has many favorites but a much lower rating, then I'd assume it's because of downvotes that imply the image is less popular than the favorites might suggest. Maybe there's a way to combine both systems for measuring popularity?

sys-yok said:
I like having separate upvote/favorite options. They can be searched separately, which is convenient for organization. Upvotes can be cancelled out by downvotes, while favorites don't have that. Favorites are easily visible to other users while upvotes aren't.

Using upvotes to measure popularity makes sense to me since it accounts for downvotes. If an image has many favorites but a much lower rating, then I'd assume it's because of downvotes that imply the image is less popular than the favorites might suggest. Maybe there's a way to combine both systems for measuring popularity?

I doubt upvotes are meaningfully counteracted by downvotes in the vast majority of cases, but only the admins could shed light on the voting statistics there. I suspect, rather, that most users are not motivated enough to Upvote AND Fave, or don't realize how the two systems interact/are different from each other.

Reduce complexity.

I may explicitly upvote certain images but not favorite them, since I use my favorites in much the same way other sites use 'em: for categorization and to show fondness towards the post. FA only has a favorite category, so they get squished in there when I could absolutely like an image, but not like it enough to want to see it in my favorites section.

Vice versa, whenever I favorite a post here, I also upvote them. Now, I'm willing to bet that some people choose to keep their upvotes and favorites separate, so I don't believe we should implement anything that may hamper a person's ability to keep them separate. If a person wants to find the typical "opinion" of a post, they can use the upvote/downvote total; if they want to see the total amount of support or appreciation of a post, they can use the favorite total instead. Even though it is possible to upvote, downvote, or favorite for literally any reason, these are typically reliable values to use to determine the aforementioned.

So, I'm gonna say "no, this shouldn't occur". This is entirely a "you" problem, and though others might share your opinion a "solution" should not alter the site for everyone. Not everyone sees it as a problem...

It will not happen. Users treat voting and favorites the way they want to. There is also a favorites limit that users are already hitting, but no voting limit.

Also, the problem you mentioned doesn't seem to exist. Unless a post has a higher downvote ratio, it will be high on the page. date:today order:score and date:today order:favcount return similar results.

The real solution to the Popular page would be to implement the code I wrote in topic #24996. I might update that since I have another idea for it.

Again, I think it's all too easy to see user intent where the driving force is actually habitual usage and feature apathy. I understand some people use the two features differently, as they should, to accomplish different things. That's great.

However, I would bet a month's paycheck that most site users are not intentionally choosing to Favorite, or choosing to Upvote, making careful decisions on each post. I suspect most people swiftly become accustomed to taking one action to indicate they "Like" a post, and that's the one action they take each time out of habit -- treating the two features as the same thing in their mind. In any case, I think reworking the Popular page is a great idea. Thanks for linking to that thread; I left a comment there.

Also, I don't understand your comment that "the problem I mentioned does not seem to exist." It very much does exist: the Popular page appears to be a simple ranking of vote score, and does not factor Favorites into the ranking at all. A post can have an infinite number of Favs, but without sufficient upvotes, it won't appear in the Popular page.

That's the main thrust of this discussion.

Updated

tbh the only confusing part is when both upvotes and favorites are used to measure a post's popularity, in that case having two separate scores that mean the same thing feels very superfluous

right now upvotes function the same as they do in other social media, while favorites essentially add the post to your own "private" gallery

those are pretty distinct uses for both of them but it seems somewhere along the way a lot of people end up treating favorites as a secondary voting system, I'd think it might be because the number is displayed along the post's score? making it feel like it has the same level of "importance" in determining the popularity of the post

I really don't think upvotes and favorites should be fused into a single metric, nor they should be both used together as a criteria to put a post in a popular page or anything like that. that would just muddy things up even more trying to make both of the numbers mean something they shouldnt

if anything, what should be changed is how and where the number of favorites is being displayed to users in order to convey they're not "Upvotes 2.0", like not displaying it in the posts thumbnail for instance, only in the post page itself

An incredible number of posts on e621 have a higher total of Favorites than Upvotes. By the logic offered by others in this thread, that outcome would be rare or non-existent, yet it's very commonly seen.

If a user likes something enough to Favorite it, that should be treated as an Upvote for the purposes of ranking posts in the "Popular" listing. I understand that the reverse is not true: an Upvote does not imply a Favorite.

tl;dr

People are clicking +Favorite but are simply too lazy to click Upvote or don't realize the value of clicking Upvote. This is reflective of user ease, folks, and not an indication of deliberate user intent. Don't make the mistake of assuming the userbase possesses your own sophistication and knowledge. You're making a product for the general public -- design the interface with that in mind.

Updated

hhrff said:
An incredible number of posts on e621 have a higher total of Favorites than Upvotes. ...If a user likes something enough to Favorite it, that should be treated as an Upvote for the purposes of ranking posts in the "Popular" listing. I understand that the reverse is not true: an Upvote does not imply a Favorite.

People are clicking +Favorite but are simply too lazy to click Upvote or don't realize the value of clicking Upvote. This is reflective of user ease, folks, and not an indication of deliberate user intent. You're making a product for the general public -- design the interface with that in mind.

Alright, this clarifies the point a lot more now.
Honestly, dunno about others, but this seems like something that could be better solved by changing Favorites to, like, a "save to gallery" thing, mostly in name only. Plus, that's how most other art sites do it. There's a lot of value in having Upvotes for support/banter while also having the reaccessible, more statistically-unfaltering Favorites (as they are now).

Don't make the mistake of assuming the userbase possesses your own sophistication and knowledge.

This, though... Don't know how to feel about it. I doubt that a lot of features integrated into websites these days meet the exact same concept that the developers made them for. It'd be pretty bogus if there was some sort of crackdown against people using custom statuses (hint hint to a certain platform) as a sort of portable "about me" instead of writing down their own thoughts and feelings at the moment.

hhrff said:
If a user likes something enough to Favorite it, that should be treated as an Upvote for the purposes of ranking posts in the "Popular" listing

A favorite and an upvote mean completely separate things like I said above

honestly, I feel that a lot of confusion might be due to the fact that FA only has a favorite system right now, so the only way for users to "vote" on submissions is by adding them to their favorites, which in turn leads people to do the same here

funeralopolite said:
Honestly, dunno about others, but this seems like something that could be better solved by changing Favorites to, like, a "save to gallery" thing, mostly in name only

I don't know if this would be ideal solution, but it certainly is on a much better track to disambiguating upvotes/favorites. This line of thinking keeps the scores as they are and at the same time helps users not to be confused about how they're used in the site

funeralopolite said:
Honestly, dunno about others, but this seems like something that could be better solved by changing Favorites to, like, a "save to gallery" thing, mostly in name only. Plus, that's how most other art sites do it. There's a lot of value in having Upvotes for support/banter while also having the reaccessible, more statistically-unfaltering Favorites (as they are now).

Personally I much prefer seeing Favorites as a better gauge of the post's popularity. If I see a post with a ↑20, what does that actually tell me? Is that 20 upvotes? 30 upvotes and 10 downvotes? 100 upvotes and 80 downvotes? And since people upvote and downvote for different reasons, ranging from technical ("this is well made, even if I don't like the subject matter" / "this isn't that good, even though it tickles my fancy") to supportive ("this may not be great, but you got potential") to personal ("that's hot" / "no blacklist for me, I must downvote everything I find morally repugnant"), so when one upvote == one downvote, that number tells me nothing substantial.

Favorites suffer less from this problem. While yes, people can still favorite a post for arbitrary reasons, they're not counteracted by equally arbitrary downvotes. As a result, posts tend have more favorites than upvotes, and I think works a bit better as a gauge for how well liked a post is. But it does also still suffer from the problem of upvoting and favoriting being separate, as two people can have the same opinion on a particular post, one upvotes and the other favorites, causing both numbers to be lower by some unknown amount right off the bat.

The admins have made it clear that they don't intend to change anything about this, though. So unless someone manages to come up with more convincing arguments than those that have been made in previous threads where the topic came up, we just gotta deal with it.

The big argument against using score as a measure of popularity is that, when a post is replaced by a superior version, favourites carry over, but votes don't. In some ways this is a good thing, for example when a post got downvoted for being a thumbnail. But I have noticed that, when an already high-scoring post is replaced by a larger resolution version, it invariably never regains its former score. Usually it doesn't even come close. I have no idea what the sociological reasons behind this are. (Surely if it got upvoted once, it should get upvoted again? People's memories of every image they've previously seen aren't that accurate, right?) Nevertheless, it's an easily provable effect, and leads to some posts occupying much lower positions in the score rankings than their number of favourites would suggest.

funeralopolite said:

Don't make the mistake of assuming the userbase possesses your own sophistication and knowledge.

This, though... Don't know how to feel about it. I doubt that a lot of features integrated into websites these days meet the exact same concept that the developers made them for. It'd be pretty bogus if there was some sort of crackdown against people using custom statuses (hint hint to a certain platform) as a sort of portable "about me" instead of writing down their own thoughts and feelings at the moment.

In this specific example (Upvotes vs. Favorites), there's no evidence that the users on e621 have cleverly re-purposed concepts/features in ways that the developers never intended or imagined.

Rather, the users on e621 appear to be unaware of the value/utility of Upvotes and, as a result, even the most popular posts on the Popular page have much higher Favorite counts than Upvotes. I appreciate the nuance the developers were aiming to achieve, but it's not being used that way by most users.

mabit said:
A favorite and an upvote mean completely separate things like I said above

The fact that you know this does not mean that most users on e621 know this. The imbalance between Favorites and Upvotes clearly shows that the distinction is lost on most users. As a direct result of this, the Popular post ranking is currently meaningless. It doesn't have to be meaningless. Let's fix it.

watsit said:
Favorites suffer less from this problem. While yes, people can still favorite a post for arbitrary reasons, they're not counteracted by equally arbitrary downvotes. As a result, posts tend have more favorites than upvotes, and I think works a bit better as a gauge for how well liked a post is. But it does also still suffer from the problem of upvoting and favoriting being separate, as two people can have the same opinion on a particular post, one upvotes and the other favorites, causing both numbers to be lower by some unknown amount right off the bat.

This is a great argument for condensing the Favorites and Upvote functionality into a single feature, or at the very least, reworking how the Popular ranking is calculated. Thank you.

Updated

hhrff said:
The fact that you know this does not mean that most users on e621 know this. The imbalance between Favorites and Upvotes clearly shows that the distinction is lost on most users. As a direct result of this, the Popular post ranking is currently meaningless. It doesn't have to be meaningless. Let's fix it.

Rebuttal: The imbalance between upvotes and favorites shows that there's a clear distinction on how most users use them. If the users did see no difference between them, that would mean it would make little difference to the userbase in general between upvoting and favoriting a post, so their numbers would be a lot more uniform. If there's a noticeable difference on their numbers, that means users perceive them as different scores and use them for separate purposes

versperus said:
that's like combining the like/share buttons on twacker

Not quite, since liking someone's post increases its engagement ranking to the system, while sharing it "retweets" it for people following you to also see it (and even then, people are strongly encouraged to both like and share for the post to get proper visibility, not just one or the other). Last I knew, e6's favorite system isn't meant as a timeline blog that other people follow so they can see what you want to signal boost, it's more of a personal set of what you like.

mabit said:
Rebuttal: The imbalance between upvotes and favorites shows that there's a clear distinction on how most users use them. If the users did see no difference between them, that would mean it would make little difference to the userbase in general between upvoting and favoriting a post, so their numbers would be a lot more uniform. If there's a noticeable difference on their numbers, that means users perceive them as different scores and use them for separate purposes

Counter rebuttal: That doesn't indicate the distinction is the same for different users. One person can upvote things they find technically impressive (and downvote things they find technically boring or bad) and favorite things that make them laugh or get horny, while another person can upvote things that make them laugh or get horny (and downvote things that don't do it for them or are boner kills) and favorite things that they find technically impressive. Or worse, the same person can even use different criteria for different posts (e.g. upvote one post because it's technically impressive, then downvote the next post because it offends their morals). As these are completely opposite criteria, and you have no idea which criteria each person used to get the current score and favorite count, what do the separate numbers really tell you?

hhrff said:
Problem: many very popular posts in terms of # of Favorites receive a small fraction of that same number in Upvotes, relegating them to not display on the Popular page. This presents a distorted view of what's popular on e621 to the userbase.

Problem: This is not actually a problem. For a gauge of how much of a non-issue this is, I didn't even know we had a Popular tab, because I never looked, because I never cared to look. I don't come here to see who is popular. I never once in my life asked if a piece of media was popular before I liked it, and I ain't gonna start anytime soon. Popularity is academic, that's all.

Solution: Stop caring.

hhrff said:Solution: take the approach nearly every other social media website uses, by providing just one "upvote/favorite/heart/star" functionality. Why have separate functions? It's clearly not being used effectively.

1.) This is not a social media site, and 2.) You've clearly never been to other social media sites because this is absolutely not true. Popularity algorithms are toxic, period; sequential order was just fine but noooo, social media sites had to squeeze out that extra penny even it made the rest of us miserable.

Speaking as someone who is constantly validation-starved, this is really sad. Let people just dislike stuff. Stop invalidating other people's negative feelings to protect yours.

If you weren't aware of a Popular tab, and have no desire to interact with it now that you're aware of it, then this issue is not a concern for you.

Regarding your second point, you've clearly never been to other social media sites like, oh, I don't know, Twitter and Facebook: you can Like/Heart something, and that's it. One function. It works.

mabit said:
Rebuttal: The imbalance between upvotes and favorites shows that there's a clear distinction on how most users use them. If the users did see no difference between them, that would mean it would make little difference to the userbase in general between upvoting and favoriting a post, so their numbers would be a lot more uniform. If there's a noticeable difference on their numbers, that means users perceive them as different scores and use them for separate purposes

Maybe the Favorite button, being much larger than the Voting buttons, is more visible and a more familiar concept, thus gets used more by unsophisticated users (i.e. vast majority of userbase on any website). You really seem to want to believe your personal "nuanced usage" applies to the entire userbase. Do you interact with the public much in a professional capacity?

Updated

watsit said:
Counter rebuttal: That doesn't indicate the distinction is the same for different users. One person can upvote things they find technically impressive (and downvote things they find technically boring or bad) and favorite things that make them laugh or get horny, while another person can upvote things that make them laugh or get horny (and downvote things that don't do it for them or are boner kills) and favorite things that they find technically impressive. Or worse, the same person can even use different criteria for different posts (e.g. upvote one post because it's technically impressive, then downvote the next post because it offends their morals). As these are completely opposite criteria, and you have no idea which criteria each person used to get the current score and favorite count, what do the separate numbers really tell you?

That is exactly my point. All these examples you've posted show users making conscious decisions to use upvotes and favorites separately right? It doesn't matter if they are currently not agreeing with each other, but all of them still see a distinction between the two scores, so there's still merit to keep the scores separate

The problem here may be that it's not clearly signalized to them the "intended" usage is (in quotations because as I talked about it's not really a rigid thing, one changes the score up or down and the other just saves the post to your personal favorites gallery), but even then we still have the majority of the userbase using the two scores to mean different things.

My point here is that regardless of how users are currently treating the system, merging the two scores together would be a bad idea because users are objectively not using them to mean the same thing. So getting rid of one of them leads to a massive loss of information one way or the other

hhrff said:
Maybe the Favorite button, being much larger than the Voting buttons, is more visible and a more familiar concept, thus gets used more by unsophisticated users (i.e. vast majority of userbase on any website)

You understand that I can also make the exact same arguments in the opposite way? The up/downvote buttons come first in the post, are color coded, show the direct effect of your voting on the post's score (something that you have to search elsewhere for favorites) and are also a familiar concept in major sites like youtube and reddit. So by that logic the users would have even grater incentive to interact with the voting system than favorites right?

In the years I've been around here, I've seen both in forum discussions and in the scores of my own posts that e621 users treat upvotes and favorites differently, with the "problem" being that each user might have different notions on what each of them mean. My suggestion up there was to seek ways to make it a bit clearer to the user what the best way to utilize the scores. Having both scores is still completely valid regardless of any changes though

watsit said:
Not quite, since liking someone's post increases its engagement ranking to the system, while sharing it "retweets" it for people following you to also see it (and even then, people are strongly encouraged to both like and share for the post to get proper visibility, not just one or the other).

I remember when Twitter didn't have a Like button. It was better TBH.

mabit said:
My suggestion up there was to seek ways to make it a bit clearer to the user what the best way to utilize the scores. Having both scores is still completely valid regardless of any changes though

Can you explain to me why the most popular posts, even if the material is relatively uncontroversial (i.e. not likely to garner high number of downvotes), have much higher Favorite counts than Vote score? Are you really going to maintain your position that this is all working as intended and no change is needed?

Example from yesterday's Popular page: https://e621.net/posts/3089512 https://e621.net/posts/3089525

540 Vote score and 1229 Favorites in the first example, and a similar proportion in the second. Are you going to tell me that almost 700 people downvoted that milquetoast material? Or maybe a huge number of users failed to vote either up OR down because it's not clear to them what "Voting" really accomplishes? Which is more likely?

Just gonna pop in again real quick to say that we're making a lot of assumption, at this point, what many users' goals with upvotes and favourites are, relative to some other intentional use of those. As much as I'd like to debate that favourites may seem wicked more practical to the layuser of this site, who knows.

I wonder if there's some sort of poll/data for a similar upvote/downvote system, e.g. Reddit?

hhrff said:
Can you explain to me why the most popular posts, even if the material is relatively uncontroversial (i.e. not likely to garner high number of downvotes), have much higher Favorite counts than Vote score? Are you really going to maintain your position that this is all working as intended and no change is needed?

Because people are more likely to favorite than to upvote, it is really that simple
Yes, no change is needed. The system is working correctly and people are engaging with it the way they want to

hhrff said:
540 Vote score and 1229 Favorites in the first example, and a similar proportion in the second. Are you going to tell me that almost 700 people downvoted that milquetoast material?

You can see the number of downvotes in a post btw, just hover over the score
So no, people are not massively downvoting these pictures. People are seeing both the voting and the favorite buttons and making a choice to favoriting pictures more often

hhrff said:
Or maybe a huge number of users failed to vote either up OR down because it's not clear to them what "Voting" really accomplishes? Which is more likely?

Neither, they're just choosing not to upvote
It is incredibly clear how voting on a post works, it's the most prominent post statistic shown in both the listing and in the post pages themselves

mabit said:
Because people are more likely to favorite than to upvote, it is really that simple
Yes, no change is needed. The system is working correctly and people are engaging with it the way they want to

You can see the number of downvotes in a post btw, just hover over the score
So no, people are not massively downvoting these pictures. People are seeing both the voting and the favorite buttons and making a choice to favoriting pictures more often

Neither, they're just choosing not to upvote
It is incredibly clear how voting on a post works, it's the most prominent post statistic shown in both the listing and in the post pages themselves

Clear to you, yes, we've established that a dozen times. So what?

You're telling me half the people who Favorite a post make a conscious decision not to Upvote it, deliberately ignoring an opportunity to make that post (that they clearly enjoyed!) more visible to other users that might also enjoy it?

Why? Please, speculate. Give us a plausible explanation. Maybe they only had energy enough to click Favorite, and then their little fingers got tired, eh?

I think this thread has run its course. Let's put a stop to this before it devolves any further.
We are not getting rid of the voting. It has been proposed before, and rejected by the community every time.

  • 1