* Didn't realize I had the name and expiration field to enter as well. I thought it was just show casing the number entry. It wasn't until I came back here to HN to read comments that I realized that there was more to the demo.
* Tab order was wrong.
* 1Password couldn't fill in the data.
* No indicator telling me what credit cards you accept.
* CVV was hidden. There were several cases where I'd finished filling out the form, but couldn't change the CVV.
* Errors are not displayed. Try 10/10 expiration, and you are basically stuck. Without additional information, there is nothing to indicate anything is wrong.
* You can tab from the date to the hidden CVV and enter information there, but you don't see it. This is easily done because the automatic progression is inconsistent, so I've had to tab twice before getting to the year, and because it automatically moves me there, I end up on the CVV. This makes things really awkward.
* Overall, this was fairly confusing. I failed at entering the CC data the first time.
I spent 10 years handling CC processing, and basically living on the payment page. Keep the page clean, quick, and simple. You don't want to do anything confusing. Let them enter the CC data and move on. That fancy CC form will NOT sell a single thing. It will however stop people from paying. It does nothing to convert.
The more friction you place between a customer willing to pay and the actual transaction, the less chance you have of making a sale.
Exactly. This isn't a usability improvement, it's fluff.
If you want to see an actual usability improvement, look at what Stripe has done with jQuery.payment: https://github.com/stripe/jquery.payment A good example is in their Checkout javascript component: https://stripe.com/docs/checkout
I should also point out that this seems to be just one part of the checkout process, so there are cases where it would not work (though that's probably not the intended use-case for modal form).
Do you have any data that backs that up? (excluding call-to-authority and anec-data)
I know you probably won't use it -- and that's okay. I just wanted you to know I heard you :) All of the things you mentioned are things I'm actively working now to address.
And I applaud you for that!
> All of the things you mentioned are things I'm actively working now to address.
Awesome!
I should make it clear, I'm not posting this as a "OMG He's wasting his time!" As I said, I spent a decade working with this, and I can appreciate the difficulties of a frictionless payment process. I wanted to share my experience with you just in the hopes that you can do exactly what you are doing: improve it.
=)
Adding to Jason's issue list:
* Fill out everything including CVV on back, erase everything on front, option to switch to the blank back persists.
* Not sure where it's supposed to be, but backside CVV for Mastercard seems positioned partially off the top left of the card on Chrome/Mac.
* Card numbers highlights existing digits when tabbing/shift-tabbing, but expiration date does not.
* If you Tab to a card number field with less than 4 digits, they are selected and typed over. If it autoadvances, they are not selected and your cursor is placed before them. This then does not autoadvance.
* D/DD in expiration followed by Tab appears to be accepted but does not work.
* Entering complete invalid Mastercard number, then clicking and entering expiration date (without name), then Tab leaves you apparently nowhere. But it's actually on the CVV, so if you type something to see where it goes, odd things happen.
* Invalid date format, or expired date is not highlighted until you type the first character of name. If out of order, you have to erase previously entered name completely before highlighting recurs.
* Tab from last field takes you off the form with no resolution.
* Enter four digits in card number field, position cursor in middle, then tab to next field. Pressing delete erases the number where you last had the cursor, not the last digit.
* Enter 4 numbers in card field, autoadvancing to next field. Shift-tab back. Tab forward. Delete erases all 4 number in previous field, rather than last as expected.
I think it's a nifty idea, visually pretty when it works, potentially useful as a standard, but right now I feel like I hit bugs faster than I can type them up. Hard to say how useful it will be when actually working, but seems worth finishing to see.
Yes.
> Could some other arrangement have less? And isn't it great that someone else is experimenting to try to find it, so you only have to switch once it's established to actually be better?
Yes. I say as much in a reply to the OP here: https://news.ycombinator.com/item?id=6143898
Citation needed.
>It does nothing to convert.
Beside the point. Any other credit card form I've seen does nothing to convert either -- and it's not it's job: by the time I'm entering my credit card details, I'm already "converted".
Most of the objections are dead simple to fix (tab order, 1password), and address nothing inherently wrong with the core idea.
>by the time I'm entering my credit card details, I'm already "converted".
Our numbers say otherwise. Optimization of our checkout page is a huge deal any change, even text, comes under scrutiny, and for good (i.e. measurable) reason.
Would I put this demo on our page? If the bugs and confusion can be fixed, I might try it for a very small test group and see what happens. Not sure the confusion can be fixed though.
No, when you finish entering your credit card information, you're "converted."
I have been ready to purchase but abandoned orders because I was doing multiple things. By the time I remembered to go back, I didn't feel as strong a compulsion to purchase the product. In some of these cases, if checking out were easier, I would've been done and I wouldn't have felt strongly enough to cancel the order.
Shitty forms cost you money even when it's the last form you have your users fill out.
Unfortunately, I've never published anything to any official standard. I can only go by my experience. That experience involves transactions north of $100M.
> Any other credit card form I've seen does nothing to convert either -- and it's not it's job
Perhaps I worded it poorly. However, a CC form's job is to convert a potential customer into a paying customer. Just because you've entered your CC details does not make you a paying customer. You'd be surprised as the number of people who do enter information, and then never actually complete the transaction. It it's a recordable amount.
In the end, a credit card form that creates friction will cost sales. The more friction, the less sales. The amount of problems with this setup will stop sales from going through, even if just the bugs are fixed.
> Most of the objections are dead simple to fix
They aren't objections as much as they are a report of what is wrong. Consider them a list of problems that need to be addressed.
> and address nothing inherently wrong with the core idea.
Correct, the idea is fine, and should be explored. Keep in mind, this isn't the first time I've seen someone use a credit card layout as a form. The CVV on the back is new, but than I'm going to say that's a bad idea the way it's implemented.
Anyways, I already replied to the author of the form, and made it clear I'm not arguing against his attempt. Hell, I applaud him. I just wanted to share my experience because that is what's needed.
I'm getting pretty tired of this attitude. Don't get me wrong, I'm all for healthy skepticism, but if you can't grasp the basic understanding that the biggest hurdle to conversion is friction in the purchasing process then your in the wrong business mate.
This isn't rocket science.
That said, there are some problems: * Amex CVV doesn't allow the proper 4 digits * Tab-order is weird * Confusing because sometimes you tab to next field, sometimes you don't * Name usually not needed since already entered elsewhere (inexplicable why so many checkout forms do this) * Difficult to change card number
That said, my experience indicates that this is not going to convert well.
To mirror the author's own quote:
"Every question a user has to ask themselves during the checkout process is another reason for them not to complete it."
What's presented here is a drastically different experience from the norm.
It doesn't behave the way you'd expect if you've ever bought anything else online.
Beyond that it has to load images dynamically based on card type which, on a slow or interrupted connection, will create even more confusion.
I love the concept, I love the execution, I just don't think this is going to be a conversion driver.
But I'll test it anyways.
If it works, and people like it, great! I'm glad to have helped a little. Please let me know if you run into any trouble (my Twitter and email are on the site).
Unfortunately, for unrecognised card types it makes it impossible to proceed. I have just tried a maestro card number (beginning 6759) and am unable to enter any of the additional details (expiry, etc).
It may be worth adding a generic fallback for cards that are not recognised to account for unexpected edge cases.
Keep up the good work!
I hope that came across in my original comment - appreciate you giving us testers more fuel for the fire!
The "flip to the other side" to fill out the CVV is great. I might provide a second button/link, appearing once the CVV is completed, that says something to the effect of "valid input, next step"
Your message for disabled Javascript is really, really weird.
> Either you have Javascript disabled, or you're using an unsupported browser, amigo! That's why you're seeing this old-school credit card input form instead of a fancy new Skeuocard. On the other hand, at least you know it gracefully degrades... > Never tell me the odds! Enable it anyway.
Write like this instead:
> You can still use this form if you have Javascript disabled or an unsupported browser. For a better experience, enable Javascript or upgrade your browser.
- You can have invalid inputs in the month field. "20" or "00" are not months.
- You can put invalid inputs in the year field.
- Tab order is wrong. Go from top left to bottom right.
Also, on Chrome 28/Mac, the validation flow seems to have some issues. I can put invalid months like "22" or "00" on the card without it highlighting the month. But it looks like the month needs to be accurate before it will show me my Luhn checkdigit is wrong. After playing with it, I'm not sure what exactly the flow needs to be, but I'm able to get green "Back to the other side..." messages with invalid data fairly consistently.
Other than that though, very neat!
I didn't realize this until I read your comment.
Please do. It's true that "it doesn't behave the way you'd expect if you've ever bought anything else online", but the (somewhat rational in a business sense) fear of it is why we're stuck with things that suck, the "first designs" that got popular and that now we're too afraid to change.
It depends what you are trying to sell. If you have used all known tricks to convince user to buy "lose 50 pounds in 2-weeks" product, then yes unusually good looking CC form nothing circa 97 can discourage user from converting. On the other hands, many times you offer product that user truly wants, and when your checkout is broken, they will send you a long explaining email, showing you how far they would go to get what they want.
>Beyond that it has to load images dynamically based on card type which on a slow or interrupted connection
I noticed that too, but thats mostly because this post > 420 upvotes. All you have to do is load them all in div style = "display: none" on body onload and viola! all are preloaded and hidden, waiting for your browser to use one from your cache.
Did I enter my freaking CC number right?
For those hunt and peck typists, the larger size, the proper handling of spaces, and the ease of comparing to the actual card makes that so much easier, even for the average users.
I will note that the expiry and name input could be highlighter better - I missed the expiry until I accidentally clicked on it.
For those power shoppers who've memorized their card number, this will still work for them too.
This actually happened when I refreshed the page while playing with it and it was replaced by a normal form. Pretty handy. I think I will also be testing it out.
One additional nitpick - it says "YourBank" at the beginning, so while I was aware this was supposed to a credit-card form because of the title of the submission, it can also be mistaken for a bank account number submission.
Because it's completely unneeded.
Typical web designer wet dream. Fancy-schmancy visual design fluff, robbing screen real-estate and bandwidth for a freaking "credit card" background image, and will it still work even 2 or 5 years from now when all web dev standards have once again turned 180° and future browsers render stuff like this way-off or not at all? Oh it won't, but you web designer dude will happily fix this at a charge of only 1.5 man-days? Well, neat!
Don't get me wrong, it's fun to spend a week or three "optimizing" an approximate 0.000001% of your web app's UX with ideas like this. I get it. Been there, too.
I see another issue --- with the CC now looking more real, the user also gets closer to the "physical reality" of "I'm parting with cash". Once I see my card's logo I'm instantly more hesitant. It's been shown that the more "layers" away from real physical payment a transaction is, the "easier" a customer proceeds with it.
Isn't it a good thing, if people are more responsible with their money?
I mean, yes, most web merchants would be much happier if all their customers spent the diaper bill and the mortgage payment on product, but I don't think you're supposed to say that out loud on a public forum. It's considered gauche.
"Wait, where do I put the CVC, name and expiration date?"
...we have a usable, familiar, industry-tested form of data entry. =D
My general attitude: it's trying too hard to look exactly like a credit card, and that makes it not enough like a normal form. Dial it back a notch or two. Have the fillable areas look like normal form fields, don't be afraid to explanatory or error labels on the form, and show front & back at the same time rather than a "flip" interaction.
I gave up after the fourth extraneous tap.
So I guess "why aren't all credit card forms like this" is because people use iPads.
(In all seriousness, great idea but disappointing first impression.)
For those not on a ipad now, what happens is when trying to tap 5 digits in a row for 4 digit field: the keyboard disappears at the fourth tap, and the fith tap misses it's target and lands on the page.
If the 'somewhere' happens to be a link (i.e. the 'view on github' link on the left if right below the '4' when the keyboard is displayed) you get out of the page, and have to start over again. Trying to enter important information is stressful enough, getting out of the page every pack of digits and start over all again is a punishment in itself.
[1] Don't Make Me Think, Steve Krug: http://www.sensible.com/dmmt.html
EDIT: And fill out the entire front.
By the way on the Mastercard on Chrome 28 on Win 7 it shows the CVC on the top left corner of the card so that the text overlaps with the border. On my Mastercard atleast the CVC is in the middle of the card.
Would you be open to a patch like that?
* As others have mentioned, there's no advertising of accepted cards
* None of the empty inputs look like inputs until you tab into them. They also cease looking like inputs when you tab out. Correcting data that is not objectively invalid (e.g. expiration in the future, but wrong month) is kind of unintuitive. Someone who clicks into form fields will likely be entirely stumped by this form.
* It doesn't seem to work on mobile. On my iPhone the series of inputs were positively irritating to interact with, somehow each one only took three characters instead of four, and the lack of a tab meant I had little direction on where to go when the first four digit input disappeared.
* The CVV is clumsily poorly handled. You can actually tell in that you're breaking away from skeuomorphism to get the user to that input. It might be a better idea to show the "back" next to the card when you do the initial breakout.
This is a great step up from that. The one thing I felt was not perfect was the multiple entry boxes for the different card types. Erasing/backspacing is not so easy. I would suggest a single entry field, and do something about adding padding spaces automagically. This also plays in to the auto-complete/form fillers these days. A credit card number is a single number, not 4 groups of numbers you copy and paste independently.
Plus, I don't know anyone anywhere anymore that's confused by a few credit card fields.
Plus, tabbing from the number goes to the name field and not the expiration.
Then you'll still have the problem of probably having to enter a valid billing address on separate fields.
I don't know if this really improves anything other than the aesthetic is very pretty.
You should use data urls or sprites for the CSS, then you don't have the the loading image coming in with delay.
Changing stuff once you have filled out the number is hard. You have to navigate your way back through the 4 blocks, and the auto-tab and the way it reacts to arrow keys means that the cursor doesn't end up where I anticipate it.
One bug I noticed: if you enter one of the sample numbers such that the number splits into multiple fields, then backspace starting from the end, you'll end up moved to the beginning of the first field with the first few digits still filled in, rather than backspacing over them.
More generally, any time you construct your own custom input rather than using browser input fields as designed, you end up with this kind of brokenness because you have to reimplement browser logic.
I'd love to see an <input type="creditcard">, with an input-creditcard.js polyfill, and native support added to browsers. On platforms that already know your credit card number, this would allow for simple auto-fill of all relevant information. This would also make life easier for things like Google Wallet that construct virtual credit cards for you, and it would allow for browser extensions that integrate with credit-card vendors to produce one-time tokens.
As others have said, some minor instructional text would help people get started. Also, making the name and expiration fields more obvious from the start (more obvious that they're fields and more obvious that they will need to be filled out) would also be helpful.
Having to click to flip the card is probably the biggest problem (though fixable). Typically when filling out a form, I suspect most people use the tab key to go from field to field. Having to go back to the trackpad/mouse and click on a spot slows things down. Maybe have the card automatically flip if you hit tab after the expiration field?
#contrastrebellion http://www.contrastrebellion.com/
Maybe you could get rid of the text entirely, MM/YY should be enough of a cue.
The major flaw (fail) for me is that LastPass doesn't know what to do with it.
Still, I passed to our dev team for consideration.
I do, however, like the idea of offering a "live preview" of data filled in via a normal credit card form (using the stripe checkout library for example which does all the nice card type detection and input formatting stuff for you).
This type of "preview" could also be useful in a confirmation step, ie. show the front and back of the card so the user can confirm it looks okay (I don't think there would be much advantage in "flipping" back and forth between front and back, though).
I would be concerned about reducing conversions though. Unfamiliarity can be scary.
One annoyance for me - I can't hit Command-A to select everything and start over. I'm also annoyed generally by auto-advancing text fields, though this implementation seems to work around some of the issues.
It's horribly wrong to even suggest that. The instructions should say "type 4111.... I don't want your credit card number", and should pop up an error if a valid checksummed card number is entered. Don't miss an opportunity to help people lean safe habits.
- ALL of the inputs should look like inputs and only after a user has put in their information should the input ui disappear and the text be shown as it is on a real credit card. Otherwise the ability to enter your name and expiration is ambiguous.
- I don't like how AFTER you enter your full credit card number, the input fields become segregated, trying to change the credit card number becomes difficult as you now have to click on each of the 4 new text fields to enter a new credit card number. (If a card fails to go through a user might try to use a new one). Yes you can just hold down BackSpace and it will travel from field to field deleting the number but that's not obvious for a user to figure out. Don't segregate the numbers into separate fields just leave them in one.
That said, I totally agree with the premise. I'm totally baffled why so many forms require you to unnecessarily select a card type or insist that you enter a number with spaces or only without spaces.
But I liked the concept. The "flip over for CVC" worked for me on Firefox, and that was an elegant solution. I think a lot of the criticism could be avoided by just making it static: don't load the image and redraw the card based on the input numbers. It then boils down to, essentially, the image of a generic card with some generic fields on it. Add some red text for fields that require input, and I think it's an elegant way to guide people when entering CC info. I'd be interested if anyone did A/B testing with this variant (and others) against regular forms.
What problem is it a solution for?
I actually find the lack of any label on the CVV field a bit confusing, but maybe that's just me.
My only gripe would be that it doesn't innately explain which cards are acceptable... E.g., you might not accept Diners Club and Discover, but do want to accept Visa, MC, and Amex. Having little icons there make that more explainable. We did it like that on our contribution form: https://contribute.publicradio.org/contribute.php
NB: I comment on amazing HN submissions (Show HNs usually) primarily so that I can find these tools again by browsing my comments from my profile when I feel the need to use them. This is primarily a comment to that end. Thank you so much for this Ken!
http://stackoverflow.com/questions/72768/how-do-you-detect-c...
However, there is a few things that makes me not want to use it.
Firstly, no security code.
Secondly, it's hard to tell which fields aren't being filled out. Especially the YOUR NAME in particular. Maybe you should make these fields red, or when you complete a field it automatically goes to the next field. For example, when you finish typing in your credit card number the cursor automatically goes to the expiration date, and when you finish typing the expiration date it automatically goes to the name. Just a suggestion.
Usually those granular types are on a select list, if one chooses badly, she gots "transaction refused" without any more info (clearly not nice).
card number -> expiration date -> name
instead of the current order of number, name, date.
It's a really slick implementation, I love the flipping animation, I used an extremely similar animation setup (face front/face back, and a flip property) on a site I built.
And what happens if javascript is disabled?
Not something I'd use right now. It's too clunky compared to just having some textboxs on the page. I'm not a fan of skeuomorphism in general though, so I'm probably a hard sell.
Doesn't mean you shouldn't keep trying. Thanks for putting this out there. I love the article, btw.
Regardless, fantastic piece of code! Will definitely think of incorporating this.
As for the input, the standard forms that we grew used to are probably here to stay for a little longer.
Making the whole thing look like a credit card doesn't seem necessary at all.
Without great experiments like these, there would be no progression. Kudos to Ken.
Small bug: Amex cards have 4 digit CCV. Although you do show "XXXX" (4 digits) before I start typing, when I get to that field I can only type "123" (3 digits).
I can think of all sorts of ways for it to drop conversions if it fails, but it is so perfectly done that I kind of don't care.
I'm going to keep an eye on this because it has the potential to really increase the comfort people have when entering a credit card. And because it is beautiful.
I'm glad you think it's beautiful. That's all I can really ever hope to hear as a designer and engineer.
Other than the nice visuals it doesn't help anything.
As a designer, you should think of how to make the process of entering the CC more efficient and less distracting, so your conversions don't go down the drain.
A simple clean design with standard input fields will outdo your pretty design any day of the week.
p.s. Does your design even work if JS is disabled?
There is a very good reason to model the physical card. On the assumption that most folks enter their CC information by taking the card out of their wallet & copying the information from the card into the form. Then for those folks following the physical object makes it easier for the user to locate the expiry date & CCV code on the back.
I assume that folks like us that use CC auto-fill tools are a minority. And yes for us this is totally unnecessary.
PS. Do any non-tech savvy people disable JS? And if so, why?
You learn where the CVV is the first time you successfully buy something online. You learn how this specific form design works the first time you enter a website that uses it. Which do you think is going to be happening more frequently?