Testing Some Myths about Testing and Targeting
Following are a few of the questions about testing that I get most often, and my best answers. Let me know if your experiences match mine.
Is it true that “ugly” pages always win?
Hardly. This is a misconception caused when an ugly page beats a high-concept design. Remember, ugly beats confusing any day.
Is simple always better?
My experience has shown that all things being equal, simple usually wins. Simple doesn’t mean Spartan; it just means a page that’s focused on the task at hand.
Anything on the page that contributes to that primary goal can fit into a framework of simplicity.
Should I put my registration process all on one page?
Surprisingly, testing shows that it’s often more effective to split a registration flow over more than one page. This makes each page short and sweet, and once a user starts the process they’re more likely to finish.
What’s the most common mistake you see on landing pages?
No clear call-to action, or competing calls-to-action. I find this usually happens because companies stop looking at their pages with a fresh eye. Make sure you show your users what to do.
What are the most important user segments to keep in mind when testing?
Good segments are big enough to matter, and also contain key information about the users. Here are the two that I find most actionable:
Referral source: With each of these sources, you know something about the user, such as a search keyword or the message they saw before arriving on your site. Use this info to target a reinforcing message.
· Paid search
· Organic search
· Display ad
Visit History: First time visitors may need an introduction to your services. For repeat visitors, you can target content based on their previous behavior.
· First time visitor
· Repeat visitor
What’s the best color for buttons?
Believe it or not, this is the question I get most often! Based on all our tests, red or orange buttons usually perform the best. On B2B sites, text links often perform better than buttons.
Where’s the best place to start testing?
The optimal place to start meets three criteria:
1) The page gets plenty of traffic
2) The page has a measurable success metric
3) The page is underperforming on that metric
It’s tempting to test on a page that meets only two of these three criteria. Resist the temptation – otherwise you’ll be testing for testing’s sake.
With all of these answers, there’s one common thread: At the end of the day, you won’t know until you test. Whenever I get complacent or feel like I know it all, I run a test that turns everything upside down.
I wouldn’t have it any other way… that’s what keeps my job so much fun.
Can you think of any I’ve missed?