Sorry Doug, Marissa was Right. Very Right.

Stop Sign - Douglas Bowman 2010

Would drivers be more or less likely to stop at this ‘redesigned’ stop sign?

There are calls to action in this world. In many cases these rely on visual cues, signs and symbols. As a society, and even as a species, we communicate through a ‘universal’ visual language. This language comes from experience, from government, from cultural trends and taste, and many other sources.

It’s a big f*cking deal, and you aren’t going to shake it out of someone’s head.

The time that we, as messengers, have to communicate important information, especially when calling another person to action, is often measured in milliseconds. The amount of time the human eye gives the top of a magazine cover in a newsstand comes to mind, where designers know they have a fraction of a second to try to convince a passerby to purchase a magazine. Perhaps these same designers take some comfort in the fact that their magazine has spent decades and millions of dollars on marketing and branding. People recognize the type treatment of the magazine, and can connect a value proposition to the item immediately, within this tiny window.

Google makes money when people click advertisement links.

Period.

Blue + underline = link

Guess what happens when you start to take the blue out of the link color? Click through rates go down. Way down.

What happens when CTR goes down for Google? Revenue goes down, way down.

Marissa Mayer had a question - the right question - and she answered it by testing.

“If you make links less blue, do we lose money as a business?”

Yes, yes you do.

Check out Josh Porter’s great Metrics Driven Design presentation, slides 7 and 8 show the test and the results. From that presentation:

From Josh Porter's 'Metrics Driven Design' presentation

This is why you test, because you have questions and you have assumptions, and design is ultimately intimately tied to business, as it creates the entire user experience - the value.

Google did not incrementally test for the purpose of optimization

They tested to find out if a design violated fundamental rules of signage, practically a priori rules of experience, tendency, and interaction of their users. I should note that the same might not be true for all web users, perhaps the blue-link-click-rate is the result of Google’s own years of user experience.

I’ve grown tired of debating such minuscule design decisions.

source

How is this a minuscule design decision? Hypothesis: people will click fewer links if the links aren’t blue enough, and we’ll leave tons of money on the table as a result.

Is that hypothesis true or false?

As a business, you should care. As a designer, you should care. Designers are not simply creating ‘pretty things’ on the web; designers are more than artists, that is why they are so important!

Constraints and creativity

It’s a designer’s challenge to identify the constraints that he or she faces in a given project space, and to deliver successful designs within these constraints. Testing is therefore a valuable tool for the designer, to better understand 1) the nature of the constraints and 2) whether the design is in fact successful from a business perspective.

It is my fervent belief that the ultimate reconciliation of data and design is to allow designers to have nearly total control, while demanding (and ensuring) that designers have the knowledge and the tools to recognize constraints, and to gauge the performance of their work.