5 ms·

How I increased conversion on my checkout form

71 points by wasd 12y ago

andrewljohnson 12y ago
I'd like to see performance vs. Stripe Checkout, which we use. Like this: https://stripe.com/img/blog/posts/checkout-new-features/blog-post-hero.png https://stripe.com/img/blog/posts/checkout-new-features/blog... I would also like to see any analysis of which of these many changes made the difference. Did only particular thing account for most of the difference?
- eps 12y ago
  I'd like to see performance vs. PayPal. I launched a project a month ago (a desktop Windows app) and only had the Stripe checkout. Guess what? No one outside SV and HN has any idea what the heck Stripe is and half of them "don't feel like giving you my credit card info" even though it says right there that it goes directly to Stripe. So after 2 weeks I grudgingly added PayPal and added itsy-bitsy hard to find link to it on the checkout page... Drumroll... nearly 50% increase in sales!
  - DustinCalim 12y ago
    This post makes me sad, but reminds me just how big of a technical gap there is between the valley, the early adopters across the country, the rest of the population in the middle, and then the older/lagging adopter generations at the tail. (I'm picturing a normal bell distribution here). A lot of companies are creating solutions to problems the majority of people don't have yet, and I think the gap is widening. Bear in mind it wasn't long ago that nobody had any idea who Paypal was...
mmastrac 12y ago
Now that these items are all in, it might be interesting to see him run a multivariate test to see which of the factors actually contributed to the 60%, and if all of them are net positive.
pbreit 12y ago
Great info. It's beyond the call of duty but wonder if Pud or anyone else could quantify any of the individual changes?
b0b0b0b 12y ago
Did the strength of the ssl decryption change? Do some browsers not support 256 bit ssl encryption?
- scott_karana 12y ago
  No kidding; I thought AES256 was present even in SSLv3.
kuyan 12y ago
Why does the before form say "Secured with 256-bit SSL encryption" when the after form says "128-bit SSL"?
- nhebb 12y ago
  He mentions that he was inspired by this article: http://baymard.com/blog/visually-reinforce-sensitive-fields http://baymard.com/blog/visually-reinforce-sensitive-fields From which it looks like he copied the "128-bit SSL" text verbatim.
larrys 12y ago
Noting that on the before the "you're one step away" immediately jumps out at you. Personally I don't like that it seems like overselling to me. I'd like an a/b test just with that one phrase actually. The ideas presented sound good in general and make sense. I'd guess you'd have to tease out which of them actually did the trick. Anyway, another possibility is that there was something negative on the sparse form (like the "you're one step away") and by adding a few things to the page it detracted from that. In the same way that if you have a big nose it's less obvious if you do your hair (and makeup if a woman) in a certain way. So it's not the fact that you are adding makeup or hair as much as it is that the nose doesn't appear as big.
pdenya 12y ago
> Added card-type dropdown ... Their selection isn’t stored or used for anything. I wouldn't have thought of on my own, adding extra fields that do nothing except increase conversion.
- gk1 12y ago
  > adding extra fields that do nothing except increase conversion. Well, that's not certain. Any of the other changes could be responsible for the increase, and not this particular change.
brianbarker 12y ago
Adding a credit card type dropdown seems unnecessary. There are standard number patterns for each card provider, and making users select one is silly since cc gateways don't typically ask (they use the number patterns). I'd like to know if something about adding that step actually helped conversion or had no affect at all.
- fookyong 12y ago
  I think it helps the user feel confident in the form's accuracy. In some cases, abstracting or automating away information from the user can make them confused or increase distrust in the form. There's something to be said for making things 100% clear and upfront to the user. This is something you'd test, it's not something I'd say with confidence would increase or decrease conversion rates; but I definitely understand the hypothesis behind it.
- brianbarker 12y ago
  Alright, I'm fucking done with HN comments. So damn tired of being down-voted over nothing. It's like a hive of 5 year olds.
  - RobAley 12y ago
    I'm going to guess that you got downvoted as people felt your comment didn't add anything to the discussion, which is the prescribed use for downvotes. The article already explained the rational for adding it and that it wasn't actually used. It's likely clear to you why your response was also downvoted.
    - brianbarker 12y ago
      I got downvoted -12 a few weeks ago for saying that an article mentioning John Regehr was the professor from the University of Utah that I respected. What the fuck is that? This website is overrun with nerds with small egos.
      - DanBC 12y ago
Hoozt 12y ago
Compared to how it looked before - great job!
kosei 12y ago
It's a bummer that all of these were shipped at once. Tough to tell what the true takeaway is from this (security vs. reinforcing value proposition). I'm sure this was quickly put together to ship, but it would have been nice to have distinguished the valuable changes from the red herrings.

71 points by wasd 12y ago

I'd like to see performance vs. Stripe Checkout, which we use. Like this: https://stripe.com/img/blog/posts/checkout-new-features/blog-post-hero.png https://stripe.com/img/blog/posts/checkout-new-features/blog... I would also like to see any analysis of which of these many changes made the difference. Did only particular thing account for most of the difference?

eps 12y ago
I'd like to see performance vs. PayPal. I launched a project a month ago (a desktop Windows app) and only had the Stripe checkout. Guess what? No one outside SV and HN has any idea what the heck Stripe is and half of them "don't feel like giving you my credit card info" even though it says right there that it goes directly to Stripe. So after 2 weeks I grudgingly added PayPal and added itsy-bitsy hard to find link to it on the checkout page... Drumroll... nearly 50% increase in sales!
- DustinCalim 12y ago
  This post makes me sad, but reminds me just how big of a technical gap there is between the valley, the early adopters across the country, the rest of the population in the middle, and then the older/lagging adopter generations at the tail. (I'm picturing a normal bell distribution here). A lot of companies are creating solutions to problems the majority of people don't have yet, and I think the gap is widening. Bear in mind it wasn't long ago that nobody had any idea who Paypal was...

mmastrac 12y ago

Now that these items are all in, it might be interesting to see him run a multivariate test to see which of the factors actually contributed to the 60%, and if all of them are net positive.

pbreit 12y ago

Great info. It's beyond the call of duty but wonder if Pud or anyone else could quantify any of the individual changes?

b0b0b0b 12y ago

Did the strength of the ssl decryption change? Do some browsers not support 256 bit ssl encryption?

scott_karana 12y ago
No kidding; I thought AES256 was present even in SSLv3.

kuyan 12y ago

Why does the before form say "Secured with 256-bit SSL encryption" when the after form says "128-bit SSL"?

nhebb 12y ago
He mentions that he was inspired by this article: http://baymard.com/blog/visually-reinforce-sensitive-fields http://baymard.com/blog/visually-reinforce-sensitive-fields From which it looks like he copied the "128-bit SSL" text verbatim.

larrys 12y ago

Noting that on the before the "you're one step away" immediately jumps out at you. Personally I don't like that it seems like overselling to me. I'd like an a/b test just with that one phrase actually. The ideas presented sound good in general and make sense. I'd guess you'd have to tease out which of them actually did the trick. Anyway, another possibility is that there was something negative on the sparse form (like the "you're one step away") and by adding a few things to the page it detracted from that. In the same way that if you have a big nose it's less obvious if you do your hair (and makeup if a woman) in a certain way. So it's not the fact that you are adding makeup or hair as much as it is that the nose doesn't appear as big.

pdenya 12y ago

> Added card-type dropdown ... Their selection isn’t stored or used for anything. I wouldn't have thought of on my own, adding extra fields that do nothing except increase conversion.

gk1 12y ago
> adding extra fields that do nothing except increase conversion. Well, that's not certain. Any of the other changes could be responsible for the increase, and not this particular change.

brianbarker 12y ago

Adding a credit card type dropdown seems unnecessary. There are standard number patterns for each card provider, and making users select one is silly since cc gateways don't typically ask (they use the number patterns). I'd like to know if something about adding that step actually helped conversion or had no affect at all.

fookyong 12y ago
I think it helps the user feel confident in the form's accuracy. In some cases, abstracting or automating away information from the user can make them confused or increase distrust in the form. There's something to be said for making things 100% clear and upfront to the user. This is something you'd test, it's not something I'd say with confidence would increase or decrease conversion rates; but I definitely understand the hypothesis behind it.
brianbarker 12y ago
Alright, I'm fucking done with HN comments. So damn tired of being down-voted over nothing. It's like a hive of 5 year olds.
- RobAley 12y ago
  I'm going to guess that you got downvoted as people felt your comment didn't add anything to the discussion, which is the prescribed use for downvotes. The article already explained the rational for adding it and that it wasn't actually used. It's likely clear to you why your response was also downvoted.
  - brianbarker 12y ago
    I got downvoted -12 a few weeks ago for saying that an article mentioning John Regehr was the professor from the University of Utah that I respected. What the fuck is that? This website is overrun with nerds with small egos.
    - DanBC 12y ago

Hoozt 12y ago

Compared to how it looked before - great job!

kosei 12y ago

It's a bummer that all of these were shipped at once. Tough to tell what the true takeaway is from this (security vs. reinforcing value proposition). I'm sure this was quickly put together to ship, but it would have been nice to have distinguished the valuable changes from the red herrings.