Writing for the web

I love Steve Krug’s “Don’t make me think.” If you want to write better sites with the user in mind, then this is the book to read. Most of us have read it… or, at least, we have it on our shelves and we find ways to drop the title in conversation. Either way, the concept of not making a user think on the Web defines how we understand usability and, in turn, shapes how we approach online content strategies - from creative design and copy to information architecture and user interface design.

Question: Has “don’t make me think” become a mantra for *everything* webby that we do? Have we, as Web specialists, stopped thinking ourselves?

Consider the many ways we’ve stopped thinking:

A/B and A/B/C tests for landing pages, home pages, callout boxes.
Testing box shots versus human shots.
Testing calls-to-action on the left vs. on the right of an email.
Testing whether a simplified navigation bar will improve click-thru in emails.
Testing if the addition of a “Hacker Safe” logo will increase trust and, in turn, sales.
Forcing critical information into home page ‘hot zones.’
Choosing email creative that follows the F-pattern rather than making a bold statement.
Testing whether a testimonial (social proof) increases lift vs. showing a product shot.

When did we stop trusting our instincts? When did we start testing the copy in buttons?—and who has the audacity to say that ”Choose your product” increases conversion by 2.2% over “See the lineup” or that it sells even more if you place it on the right over the left?

Question: Do these sorts of tests actually impact conversion? For REAL? I’m not kidding. How much of what we test is useful?

We need to test the right things. So we get the right results - results the business can learn from. To get those results, be sure your testing involves at least these 3 things (from Avinash at Occam’s Razor, pure & simply-put genius [that’s for you, Lance ]):

1.      Have an hypothesis – Don’t start an A/B test saying, “Change this block of benefits messaging to a statistic.” Start with an hypothesis: “We hypothesize that people are more persuaded by social proof points than by marketing messages, which will lead to increased clicks for further information.”  

2.      Have at least one goal – Unfortunately, most tests I’ve been involved in have had general goals, like ‘see which one increases conversion.’ Not great. Why? Because you need to say from the get-go, “We’re looking for a difference in lift of at least 1.5% from this A/B test; if we don’t have at least that difference, we will not use the results.” When you put goals in place, you save the business from making potentially important decisions based on nearly negligible differences. 3.      Test and validate for multiple purposes – Go to Avinash’s blog now to get the skinny on this point

Something to consider the next time you want to test ‘that’ over ‘which’….

“Damn your a priori principles. Look!” 
- Galileo, to skeptics regarding the solar system revolving around the sun (heliocentrism)

Read a chapter of ‘Don’t Make Me Think’ on Steve Krug’s site

Order ‘Don’t Make Me Think’ from Amazon

Wondering what you should test?
Read Debasis Pradhan’s blog about what to test