really? creating a twitter account takes about 15 seconds. if you're not one of the 150 million people with a Twitter account, then why not just be patient? i'm sure kyle will get around to supporting other login methods eventually.
i believe you're genuinely trying to be helpful with your suggestions, but comments like this reminds me of the vocal 0.1% of users who exercise troublesome preferences like disabling javascript, or installing ad-blocking plugins that break half the web, (which is fine!), but then complain that their (pathological) preferences prohibit them from enjoying the experience (which is not so helpful).
feedback like, "i can't be bothered to log in to give you feedback because i don't like your login" isn't great feedback. it might even be viewed as a sign to you that you're not the intended user. yet.
I wasn't aware not having a Twitter account was considered pathological in our Web 2.0 world.
Why do I need to sign up for a microblogging platform to use a completely unrelated calendar app???
Anyway, my comments about reducing friction aren't based on my own preferences, they're based on dozens of multivariate split tests I have conducted across many thousand unique visitors.
Every single time, forcing the user to sign up, login, enter their email or anything of the sort before getting to use the site SIGNIFICANTLY decreases engagement, time on site, conversions, and pretty much every other metric you should care about.
The "signup friction" argument's a straw men; no sane startup seeks to add friction to the signup process. the bogus (implied) assertion in the comment is that requiring a Twitter account to use a new web app in beta is overly restrictive / cumbersome.
The only way that would make sense is if beta testers don't have Twitter accounts, since OAuthing with an existing account reduces friction. Twitter's been around for four years and seems fairly well established among early adopters like HN readers, so counting on a critical mass of beta testers to have Twitter accounts hardly seems a stretch, even if those users aren't very active tweeters.
i for one was very glad not to need to enter an email address, pick a login name, etc; the oauth process handled that for me. ymmv
I think the suggestion was to not require a form of authentication (email or twitter) first. Instead, let the potential user "test drive" the app and then when they are a little invested prompt for a signup of some kind.
One-click registrations are really an innovative way to get people to sign up and try something. Normally you have to fill out an arduous form and decipher captchas. Services like facebook connect, twitter and Open ID Connect are much better from an UX standpoint(imho).
I was also pleasantly surprised when my login information was handled automatically as well.
> "i can't be bothered to log in to give you feedback because i don't like your login"
Eh? I wanted to try this product out, but I don't have a Twitter account. In fact, I deleted it a few months ago. Now I have to create a new account just to use a (admittedly sorely needed) Seinfeld calendar? No thanks, I'll just use Vim and a text file.
I realize Streak.ly is a free application, so I have no right to complain. What bothered me about your comment was the assumption that everyone on the Internet should have a Twitter account.
> "i'm sure kyle will get around to supporting other login methods eventually."
I bet he will. But by the time he has a simpler login set up, people would have forgotten about Streak.ly altogether. The average Internet user (read: me) is a fickle being.
it's not just signing-in using twitter, from a general public's perspective, it's being unclear about how the service will interact with twitter. will you service start tweeting my streaks ("50 days without sex!"), will it start inviting my friends to join? will start inserting self-promotional tweets? maybe it can't but will the public at-large understand that? and do i want twitter to know when i'm signing into your service?
honestly, i'd prefer to have my own sign-in for your service and i think many other people would, too.
comments like yours remind me of the 0.1% of lazy programmers who would rather tell users that they're pathological than spend the day it would take to implement and test a simple user registration and sign-in system for people who would prefer that over using twitter sign-in. i, personally, will not be a customer of your service because of your customer service! :-)
Actually, I'd like to provide a counter view point because this is really how I view it: but I actually prefer it if you had the "login with FB or Twitter or Google" because to me, I absolute get annoyed by having to sign up for yet another service. Another username, email, password, to remember == annoying for me.
That isn't really so annoying though - use the same username you do everywhere else, use the same password (click supergenpass so it doesn't matter if the site gets hacked) and make your browser remember it.
Because there sure as shit isn't any way I will just give some random website my google account - they get a dummy email forward, just like everybody else.
Same username? That's assuming it's still available. If it isn't ? Then you have to remember a one-off username. Also, I do not use the same password for all web sites. That's the most basic security FAIL. If someone figures out your password, then all your accounts are compromised. This is the worst security advice ever. I'm sorry - but you know this is true.
Yes - I also do the dummy email forward. I agree about not giving random sites your google account. So create a dummy google account.
Yep, needing to use Twitter account for login is a major minus. Maybe I am too old-fashioned, but if I click on Sign-In and it automatically logs me in, this is more of "WTF has just happened and where did it get my login credentials?" rather than "Ohhh, very cooooool, dude. Just saved 5 seconds and not needed to press 20 (twenty!!) keys. Awesome.".
This is actually refreshing to hear; I'm usually used to seeing the exact opposite - "what do you mean another account?!?". I wonder if adding an interstitial page with something to the effect of "you're signing up with Twitter FYI" would be a decent intermediary step.
Good idea, Kyle. I don't mind Twitter or Facebook OAuth logins for non-sensitive consumer apps (sensitive = apps dealing with financial info, dating sites, health, etc.). Just tell me upfront that you're using the service, and I'll be happy that I don't need to create a new account.
Thats because those people haven't reduced the friction to create an account yet - I have it down to about 20 seconds and making the browser remember the password means it is actually faster from then on.
If you absolutely must, then use OpenId. It is was pretty much made for that (and it works with twitter, facebook, gmail, your mom, etc).
Comments
For a simple webapp like this, please don't make me get a twitter account/sign in to use.
Generate a unique URL for me to use, or hash my email address or something like that. Reduce friction, I want to be able to use this with 1 click.
really? creating a twitter account takes about 15 seconds. if you're not one of the 150 million people with a Twitter account, then why not just be patient? i'm sure kyle will get around to supporting other login methods eventually.
i believe you're genuinely trying to be helpful with your suggestions, but comments like this reminds me of the vocal 0.1% of users who exercise troublesome preferences like disabling javascript, or installing ad-blocking plugins that break half the web, (which is fine!), but then complain that their (pathological) preferences prohibit them from enjoying the experience (which is not so helpful).
feedback like, "i can't be bothered to log in to give you feedback because i don't like your login" isn't great feedback. it might even be viewed as a sign to you that you're not the intended user. yet.
I wasn't aware not having a Twitter account was considered pathological in our Web 2.0 world.
Why do I need to sign up for a microblogging platform to use a completely unrelated calendar app???
Anyway, my comments about reducing friction aren't based on my own preferences, they're based on dozens of multivariate split tests I have conducted across many thousand unique visitors.
Every single time, forcing the user to sign up, login, enter their email or anything of the sort before getting to use the site SIGNIFICANTLY decreases engagement, time on site, conversions, and pretty much every other metric you should care about.
The "signup friction" argument's a straw men; no sane startup seeks to add friction to the signup process. the bogus (implied) assertion in the comment is that requiring a Twitter account to use a new web app in beta is overly restrictive / cumbersome.
The only way that would make sense is if beta testers don't have Twitter accounts, since OAuthing with an existing account reduces friction. Twitter's been around for four years and seems fairly well established among early adopters like HN readers, so counting on a critical mass of beta testers to have Twitter accounts hardly seems a stretch, even if those users aren't very active tweeters.
i for one was very glad not to need to enter an email address, pick a login name, etc; the oauth process handled that for me. ymmv
I think the suggestion was to not require a form of authentication (email or twitter) first. Instead, let the potential user "test drive" the app and then when they are a little invested prompt for a signup of some kind.
One-click registrations are really an innovative way to get people to sign up and try something. Normally you have to fill out an arduous form and decipher captchas. Services like facebook connect, twitter and Open ID Connect are much better from an UX standpoint(imho).
I was also pleasantly surprised when my login information was handled automatically as well.
> "i can't be bothered to log in to give you feedback because i don't like your login"
Eh? I wanted to try this product out, but I don't have a Twitter account. In fact, I deleted it a few months ago. Now I have to create a new account just to use a (admittedly sorely needed) Seinfeld calendar? No thanks, I'll just use Vim and a text file.
I realize Streak.ly is a free application, so I have no right to complain. What bothered me about your comment was the assumption that everyone on the Internet should have a Twitter account.
> "i'm sure kyle will get around to supporting other login methods eventually."
I bet he will. But by the time he has a simpler login set up, people would have forgotten about Streak.ly altogether. The average Internet user (read: me) is a fickle being.
it's not just signing-in using twitter, from a general public's perspective, it's being unclear about how the service will interact with twitter. will you service start tweeting my streaks ("50 days without sex!"), will it start inviting my friends to join? will start inserting self-promotional tweets? maybe it can't but will the public at-large understand that? and do i want twitter to know when i'm signing into your service?
honestly, i'd prefer to have my own sign-in for your service and i think many other people would, too.
comments like yours remind me of the 0.1% of lazy programmers who would rather tell users that they're pathological than spend the day it would take to implement and test a simple user registration and sign-in system for people who would prefer that over using twitter sign-in. i, personally, will not be a customer of your service because of your customer service! :-)
Actually, I'd like to provide a counter view point because this is really how I view it: but I actually prefer it if you had the "login with FB or Twitter or Google" because to me, I absolute get annoyed by having to sign up for yet another service. Another username, email, password, to remember == annoying for me.
That isn't really so annoying though - use the same username you do everywhere else, use the same password (click supergenpass so it doesn't matter if the site gets hacked) and make your browser remember it.
Because there sure as shit isn't any way I will just give some random website my google account - they get a dummy email forward, just like everybody else.
Same username? That's assuming it's still available. If it isn't ? Then you have to remember a one-off username. Also, I do not use the same password for all web sites. That's the most basic security FAIL. If someone figures out your password, then all your accounts are compromised. This is the worst security advice ever. I'm sorry - but you know this is true.
Yes - I also do the dummy email forward. I agree about not giving random sites your google account. So create a dummy google account.
Very good point. I've got plans to add "use it now, sign up later" stuff in, it's just not made it into production yet.
Yep, needing to use Twitter account for login is a major minus. Maybe I am too old-fashioned, but if I click on Sign-In and it automatically logs me in, this is more of "WTF has just happened and where did it get my login credentials?" rather than "Ohhh, very cooooool, dude. Just saved 5 seconds and not needed to press 20 (twenty!!) keys. Awesome.".
This is actually refreshing to hear; I'm usually used to seeing the exact opposite - "what do you mean another account?!?". I wonder if adding an interstitial page with something to the effect of "you're signing up with Twitter FYI" would be a decent intermediary step.
Good idea, Kyle. I don't mind Twitter or Facebook OAuth logins for non-sensitive consumer apps (sensitive = apps dealing with financial info, dating sites, health, etc.). Just tell me upfront that you're using the service, and I'll be happy that I don't need to create a new account.
Great point. I'm going to make it a top priority.
Thats because those people haven't reduced the friction to create an account yet - I have it down to about 20 seconds and making the browser remember the password means it is actually faster from then on.
If you absolutely must, then use OpenId. It is was pretty much made for that (and it works with twitter, facebook, gmail, your mom, etc).