Hacker News new | past | comments | ask | show | jobs | submit login
The $300 Million Button (uie.com)
142 points by cynusx on July 27, 2011 | hide | past | web | favorite | 30 comments

The original reasoning that went into the design of the first form is (I think) the classic perspective taken from someone that has never had to use their own product.

"Why wouldn't someone want to create an account? Ship it!"

Then as soon as you are in a rush, shopping for some tents online for a camping trip (something you rarely do) and REI is forcing you to create an account just to order the clearance tents you found, you want to yell at somebody.

I think this story is valuable if for no other reason to make sites realize how much of a road block registration is.

Agreed. Another roadblock I hate is requiring me to enter my full name and address before they give tax & shipping info. A simple zip code field & calculate price button should not be that hard.

Does anybody publish some decent design guidelines for online shopping sites? So many sites try to follow Amazon without realizing Amazon can get away with terrible design choices because they are so massive.

And speaking of "a simple zip code field"...

If you're going to require a zip code to ship, because you only ship to the US, wouldn't it be nice if you said so up-front, so I could avoid wasting half an hour comparing your totally effing useless-to-me products to retailers that actually want to sell to me?

Couldn't agree more; as soon as I get a whiff of a website trying to dig more information out of me than it needs to perform a specific function, my tolerance for everything else on the website drops considerably.

"To calculate shipping, please enter all your personal info, credit card and hours a day no one is at your house."

As for Amazon, you are right; I think they make up for it with good prices and an excellent shopping experience. In a way, I think they can get away with it because they are so good at what they do.

I wish to go AWS Payments would allow payments without an account though so it could be used in lieu of PayPal... but that is an aside.

The key here is that the button leading to 'register' was 'checkout', and not 'register'. Because web links are essentially blind corners, where users can't see around the corner, the process involves a certain amount of trust between the user and the guide. For example, imagine being at a fair ground, and without knowing where you were there's a sign that said 'Ferris Wheel' and following that sign took you the ticket counter instead. Sure you need a ticket to get on the ferris wheel, but you were looking for the ferris wheel and not the ticket counter. This is the same thing. Checkout should lead to checkout. Register should lead to register. Checkout shouldn't lead to register. Once you've checkedout, you can close the transaction by creating an account for future purchases, but that's only because you've already completed all the necessary work, and so creating an account becomes a value added bonus.

Simple directions, simple prompts, and mechanize the rest. (Zip Code? -- Fill in the City automatically.) as mentioned by jobu below

Yes, I too want to scream every time they want to force me to register and I have also abandoned purchases just because I do not want another account.

I hope that people take note of this advice and at least make registration optional.

This is really funny -- I actually went to purchase the book "Web Form Design" that is referenced at the top of this post. It linked to the publisher's site, and when I tried to buy it, they asked me to register with email, password, etc. So I went over to Amazon instead and bought it there!

(I realize that Amazon requires registration to purchase as well, but I already have an Amazon account of course -- but the publisher is not Amazon and if they had allowed anonymous purchase I would have gladly given them my money instead, in which case they probably would have made more profit on the sale).

And the funny thing is it's largely psychological. You're giving them almost[1] the same information whether you create and "account" or not.

[1] The only exception being a user-defined password.

[EDIT] I think I've solved this in a win-win manner on my project.

  * No register or login.
  * Detail page with link at top: "Been here before? Login to pre-fill the fields if you want to."
  * An *optional* password field with a note: "If you want an account, enter a password."
  * If the user doesn't enter a password and wants to login later, a link in the password create/reset email sets them up.

You are giving them (mostly) the same information, but the presumption when not registering is that they will only store the information as long as necessary for their internal purchase history purposes, whereas an account remains "exposed" to the web, for a seemingly indefinite period of time.

That is likely the presumption for many people. It is in general naïve I'd have to say as your information is kept for a certain amount of time for returns or disputes and longer for archival and reporting purposes.

I don't have data, but I'd veture to guess that most non-account data is stored nearly as long - if not as long - as account data. (I'm thinking primarily of ecommerce applications here. Services like pastebin are quite different.)

It's not so much a matter of giving out information or not. It's a matter of having another account to remember and look after.

I just don't want the hassle.

That's why I like the "if you want, we'll email you and then you can log in" approach to return visitors which works nicely for occasional use sites.

If the user creates a "new" account with the same email, it'll silently use the same account and not show them their history until they validate.

Yeah, it sounds like you have a pretty good solution. Especially compared to some I've seen.

I actually LOVE LOVE website that let you shop as a guest. No, I do not want to sign up.

Brilliant. I'm amazed that more sites don't let you do this.

Why don't ecommerce sites just create an account for users on the fly? Why do they require the extra step?

That doesn't help the customers who come back later for a second purchase and then have to fumble to find their account info.

it would help, because users who don't want to be bothered can just create a new account on the fly.

you're already collecting all the information necessary to create an account when somebody completes a purcase, so just make an account for them. include a temporary account password on the email receipt, with a message telling users they can use that pw to log in to the site to check the status of the order. once a week, run a cron to clean up all the accounts of people whose orders have completed and haven't changed the password.

That's better, but I still don't like it because you're forcing a relationship upon the user, regardless of whether or not they want one.

Any idea what site this article refers to?

Based on my personal experience as a user of the site, feels like could be staples.com (I used to hate that form, and now it is gone, so I buy more often)....also the revenues match up to Staples of 25B.

It's Amazon.com.

Well then Best Buy has lapsed. Canadian site requires you to register in order to check out.

I guess only Jared Spool knows the true answer.

For the lazy: The answer is Best Buy, circa 04-05

I know from a couple months ago hearing the founder of omniture Josh James talk about changing a button and color to a homepage. He said it caused his client to grow significantly and caused sales to go up 15-20% almost overnight creating tens if not hundreds of millions of dollars in revenue.

The website they helped fix was still in its beginning years. Their client was ebay. I wouldn't be suprised if this was the website he's talking about.

I love this article because it nicely allows C-level execs to understand the ROI of usability and design. So often at larger companies, these areas lack budgets purely for the fact that it is so hard to quantify the returns i.e. with just general "positive" user experience. Here we can see a pure a before/after A/B type situation that had far-reaching consequences.

"But my organization is different because..."

I have seen lots of companies add that button, and I have never seen a company remove it!

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact