OK, the few hundred people in the room here with me are the ONLY people at SXSW not hearing Frank Abagnale talk right now.
Wow. The great benefit of coming to see Jared Spool talk is getting to see him dance to Beyoncé’s All the Single Ladies.
I gotta call him out…he says we’re going to talk about science and the work of a well-known “woman scientist.” It’s a reference to Lisa Simpson, but the gender ID is sticking in my craw. Is it ironic? Can’t say.
So, here we are to the main point — the secret lives of links.
We’re starting 4/8/2011–the day Congress tried tried to shut down the budget “one of the times.” Ha.
CNN writes a story about it, and they put up a bunch of links. They kept restructuring the links. Spool shows a bunch of versions of the CNN home page from that day, each time with new links/photos/stories. He starts wondering about the work behind that.
Spool — I’m not a designer, I’m a researcher. “It’s a life goal of mine never to have to design a website.”
But he decides to design a page where where CNN doesn’t have to redo the page with all the new links.
So there’s a blank page with huge CNN logo, and these links:
- The Most Important Story
- The Second Most Important Story
- The Third Most Important Story
- An Unimportant Yet Entertaining Story
- Yet Another Snooki Story
- Ad for Crap You Don’t Need
Spool — Of course that wouldn’t work. It wouldn’t make you click. But CNN has this on its page, in effect. We don’t have to add links for all those things.
Links secretly desire to get you to the content.
Shows CNN from 2007 — design is different, but those same things are there.
Shows NY Times and LA Times from 4/8/2011, and Huffington Post, same elements, same links…different words. The links are all communicating to get us to the story.
Body of science about how links do this.
Research on the scent of information, originally from Xerox PARC. The way people naivgate large information spaces, mathematically, is the same as a bee looking for nectar. We are ingrained “informavores.” We use the scent of information to get us what we want.
Note: Ah, he’s taking me back to my smallbusiness.com days 12 years ago.
Difference in MIT and Oberlin home pages — MIT doesn’t have a lot of links telling you how cool they are — Oberlin does — “because if you don’t know how cool MIT is, you shouldn’t be on their home page.”
Ohio State makes it easy for you to find useful info about the school.
Walgreens site: Very busy page. 20% of users are clicking on “photo,” 16% on search and 11% to refill prescriptions. Overall, the 5 most clicked-on links account for 59% of traffic, but they only account for 3.8% of area of the page.
Prime violation of Fitt’s Law: If it’s big and close, it’s easier to hit. On the Walgreens site, you see what’s important to the company — it’s not the same thing that the user cares about. It’s because there was a meeting at Walgreens and the marketing people won.
“We all know what happens: When marketing people win, puppies die!”
“Skip this ad” — The 3 most helpful words on the Internet!
Links secretly live to emit the right scent.
Trigger words: Words that matches the user’s goal and signals where to click.
Initial theory about web design and usage in 1995: We expected that people who’d never used the web wouldn’t be able to find things as easily as people who had a lot of experience at using the web. Instead, some sites were really good at getting people to what they wanted, and other sites were really bad. It was the first time we realized the design mattered — design trumped the user’s experience.
Signs of design failure:
- Back button
- Using search box
Back button indicates design failure: Of all user research Spool has done, 42% of site designs help people reach needed content. If clickstream includes 1 back button, drops to 18% success. If 2 back clicks, only 2% success rate in finding needed content.
When a user loses the scent, they use the back button…but the page is no more helpful than it was before.
The back button is “the button of doom.” Once they do that, you’ve lost them.
However, do NOT code out the back button!! Ha, Spool says someone did that and blamed it on him. It’s a predictor — you don’t fix anything by removing the functionality.
Pogosticking: When the user bounces around the information hierarchy. Clickstreams without pogosticking: Success 55% of time, but with pogosticking only 11% success.
Only 3 ways to get from home to target page:
- Featured content links
Users behave the same way on all sites Spool has tested except one. Normally, user comes to page, scans page for trigger words. If don’t find trigger word, they type keyword into search box. On Amazon page, they don’t scan for keywords. They go straight to search box. “After many years of business, Amazon has trained every user of the web that they never put anything useful on the home page.” Big laugh.
When users use search, they type in trigger words. We should call the search box “BYOL — bring your own link.” If you look in your search logs, you will find a list of trigger words. Code your search logs to find out what page the users are searching from — so you’ll know what page to put the trigger words on — and possibly on the page before.
If users don’t use search, they succeed 53% of time. If they do, they only succeed 30% of time.
Spool evaluated more than 10 apparel shopping sites. Huge correlation between how many pages users visited and what they bought — the faster you get them to what they want, the more they buy.
Do not design for pogosticking — give people the information they need up front — don’t make them click. You’ll sell more and your users will be more successful.
When users say “clutter” they mean, there’s way too much stuff I don’t care about on this page. They do not mean, it’s too typographically dense. You can add density as long as you’re adding information they want and need.
Now he’s shouting…love this. “Learn more” is the 2nd most useless pair of words in web design, right after “Click here.” At what point on a product page do you click and NOT want to learn more???
Good design is invisible. You don’t notice when it’s working well.
The Bureau of Labor Statistics site has some great design…you might not guess by looking. Showing designs that work for expert and inexpert audiences: links that are trigger words and then short explanations for inexpert.
Links need to look really good, but they need to look like links. Links don’t have to be blue and underlined. And thank goodness — blue and underlined are two awful choices. Underlining changes the shape of the words, and blue is hard for people to discern. The only problem is when we don’t make it clear in a design what is a link. Links need to be distinct.
Spool pauses, then: I believe if you look up clusterfuck in the dictionary, you will see United. The man flies a lot, he would know.
Links don’t have to be blue and underlined, but you do have to establish a visual language for them.
We are making it hard for users to get to content. Now he’s calling out links in the middle of magazine articles, like on the Time magazine site.
All this computer generated crap is ruining the experience. Now he’s slamming those links to words in articles — article in Chicago Tribune about law firm from Alabama, and the word “Alabama” links to everything the paper has ever written about Alabama. Such a pet peeve of mine.
People decide what they’re going to click on before they move the mouse — rollover menus kill the experience. We throw other options in their face while they’re moving toward their intended target.
Other issue with rollovers: Users want to move their mouse in a straight line, but they leave the target area by accident and the rollover disappears.
Let’s check out here what links should:
- Deliver user to desired experience
- Emit the right scent
- Look good, while still looking like a link
- Do what the user expects