Honest question: is your opinion that the 29% is composed of automated bots who were blocked by the CAPTCHA before, or is it composed of people who just don't like CAPTCHAs?
...or people who can't do CAPTCHAs. I regularly have to retry CAPTCHAs 3 or 4 times because they are utterly illegible. Sometimes I just give up and move on. Maybe I'll build a bot to help me solve CAPTCHAs!
Same. It's ridiculous that someone with 20-20 vision and good enough vocabulary to ace the verbal SAT fails roughly the first 3 or 4 tries at many CAPTCHAs these days. Some are just damn near impossible to decipher, and some I could have sworn I got right and checked multiple times b/f hitting submit, but failed anyway.
I'm so sick of them now that I only do the ones on really consequential websites, like ones handling my money. Any that I can just skip, I do.
Once or twice I've gotten a truly bizarre word I suspected doesn't even exist, and looked it up on dictionary.com to find it's not even in there. I don't bother expanding the search to the Oxford English dictionary for those, since I'm already annoyed with the wasted time.
It seems CAPTCHAs are one are where the tech is getting worse, not better.
They probably use a hidden javascript captcha, which falls back to a real captcha if not completed. I'd be hugely surprised if they dropped any kind of bot prevention.
Having said that, maybe they figured "well, bots are getting in with the captcha, so let's just drop it". But I doubt that.
The article's own title is: Gradual Engagement Boosts Twitter Sign-Ups by 29%
You're referring to the HN title, which unfortunately has become the focus of discussion here. Gradual engagement is intellectually interesting. eg:
- make people successful along the way to a signup
- games should allow people to succeed within the first 5 seconds
- postpone account creation
- end registration with a set of relevant topics (not dump users off a cliff)
EDIT they don't mention this, but it would fit with postpone account creation to push back the captcha validation of an account until the user requested an action where it matters.
I was slightly confused by the graphic and though step 2 was the redesigned step 1. I thought it was clever to see what your friends are up to before you join.
Selecting hobbies and interests beforehand is good as well, since some people might not have social circles that use Twitter.
So I agree with the author in that if account creation was last, it would be the most interesting.
That suggestion page has an awful lot of similarities to the new Digg sign-up process. Who's came first? I would have to imagine Twitter but I wasn't paying particularly close attention.
Comparing the Before and After:
- Before uses CAPTCHA, After does not.
- Before uses green "arrowed" image buttons, After uses plain grey CSS buttons.
- Before does not show users a progress bar, After does.
- Before does not use floating field explanation screens, After does.
- Before is set against a green background, After is set against blue.
I buy in that the extra step is helping, but attributing 29% to that single change is inaccurate.
The handful of points above properly tested against ~350,000 new signups a day certainly could have driven the majority of that impact.