6 ms·

WHY is it a poor choice for you? I've always been impressed by your metrics-driven choices. Do you think that if you A/B tested your favorite choice from this

WHY is it a poor choice for you? I've always been impressed by your metrics-driven choices. Do you think that if you A/B tested your favorite choice from this service that it'd meaningfully change your success? I don't hate your current logo, but as an (sorta ex) designer, I wouldn't nominate it for any awards. What is the business value of "embodying the unique values and messaging" of your business? Related: trying to ignore what you've come to believe about Google, do you think their original logo embodied their unique values and messaging? I sure as heck don't. My view is that the logo is a pretty damn tiny part of the BRAND-- and the brand is where the value is. Your business is awesome because of your brand. A great logo or crappy logo can nudge the quality of the brand but not as much as designers think.

tptacek 17y ago
Patrick, I will absolutely buy you the 99d logo of your choosing if you'll A/B test it. I would love to know the results of this test. If only we could convince Colin to try this experiment.
- webwright 17y ago
  Ooooh, I'll chip in.
patio11 17y ago
Do you think that if you A/B tested your favorite choice from this service that it'd meaningfully change your success? Probably not, honestly. I'm currently in the process of setting up a graphical A/B test for all my conversion buttons, which I think will majorly move the needle and give me ways to move it in the future. (Tweak button texts, tweak button colors, tweak button orders, etc etc. These are all hard for me in the status quo because I don't have the buttons ready to go, and I have the graphic design ability of a drunken lemur.) This might not be obvious, but all A/B tests are not created equal in terms of implementation difficulty. The cheapest A/B test is a change to page copy: it is literally one line and done. I started three of them last night, testing microcopy in my shopping cart such as "You do not need an account to pay with a credit card through Paypal." Buttons are a wee bit harder to A/B test because they're often seen in many pages throughout my site and, stupidly, they're not currently implemented as some sort of render :partial => '/static/easily_replaceable_buttons'. That means it is going to take me a couple hours to get the button tests ready sitewide, and I have to disable page caching to do it. Yanking my logo is harder still. In addition to being on every page in the site (which means I have to turn off page caching entirely, which is going to require some thought about performance consequences), there are a few places where it is used offsite, for example in the top bar of my Paypal payment pages. Changing logos right at the shopping cart is probably not a great idea. Changing that sucker will require an hour or two of work, for boring technical reasons.

tptacek 17y ago

Patrick, I will absolutely buy you the 99d logo of your choosing if you'll A/B test it. I would love to know the results of this test. If only we could convince Colin to try this experiment.

webwright 17y ago
Ooooh, I'll chip in.

patio11 17y ago

Do you think that if you A/B tested your favorite choice from this service that it'd meaningfully change your success? Probably not, honestly. I'm currently in the process of setting up a graphical A/B test for all my conversion buttons, which I think will majorly move the needle and give me ways to move it in the future. (Tweak button texts, tweak button colors, tweak button orders, etc etc. These are all hard for me in the status quo because I don't have the buttons ready to go, and I have the graphic design ability of a drunken lemur.) This might not be obvious, but all A/B tests are not created equal in terms of implementation difficulty. The cheapest A/B test is a change to page copy: it is literally one line and done. I started three of them last night, testing microcopy in my shopping cart such as "You do not need an account to pay with a credit card through Paypal." Buttons are a wee bit harder to A/B test because they're often seen in many pages throughout my site and, stupidly, they're not currently implemented as some sort of render :partial => '/static/easily_replaceable_buttons'. That means it is going to take me a couple hours to get the button tests ready sitewide, and I have to disable page caching to do it. Yanking my logo is harder still. In addition to being on every page in the site (which means I have to turn off page caching entirely, which is going to require some thought about performance consequences), there are a few places where it is used offsite, for example in the top bar of my Paypal payment pages. Changing logos right at the shopping cart is probably not a great idea. Changing that sucker will require an hour or two of work, for boring technical reasons.