Image Links vs. Text Links

Last week, on the Interaction Designers discussion list, Johan wrote:

Hey, I read an article/study a while back on the efficiency of html text links compared to image links.
Can’t find it now..

I know we’ve published it somewhere, but damned if I can find it now. 🙂

Years back, we compared successful clickstreams (clickstreams that resulted in users accomplishing their goals, as observed in tons of usability tests) with unsuccessful clickstreams (clickstreams where users abandoned their goals before completing), looking for any clues that would help us predict behaviors in one that we didn’t see in the other.

One factor we looked for was whether the clickstreams contained image links versus text links — does one type of link show up more often in successful clickstreams than the other.

Our finding was when users clicked in image links they were just as likely to succeed or fail as when the clicked on text links. There was no statistically-meaningful difference.

Our inference from this was a well-designed image link will work as well as a well-designed text link. A poorly-designed image link will fail as often as a poorly-designed text link.

Since image links are significantly harder to design “well”, our recommendation to clients has been to favor text links. They are more efficient to create and manage and produce the same results.

We also concluded, from this same research, that there are three different types of images that can appear on a page:

1. Content images — images containing information within them that assist or accomplish the user’s goal.
2. Navigation images — images containing scent to inform the user what clicking will produce for them. (These may or may not be links — they just have to inform the user about links.)
3. Ornamental images — images creating mood, displaying professionalism, organizing the page (such as rules and fancy frames), and otherwise enhancing the experience.

The quick summary was that we saw well-designed content images helped shorten clickstreams, we saw well-design navigation images helped with scent, and we couldn’t see any advantage to the presence of ornamental images, beyond pure layout assistance. (Inotherwords, no matter how we measured, users beliefs about the site as professional, as fun, as well-designed, was independent of the presence of these images. At the time, it was this last finding that branded us as graphic-design-haters, since we were dipping into the rice bowl of the visual design community.)

Of course, there’s a lot more depth in this research — I’m only summarizing before I run to a meeting. But that’s essentially what we found. Hope this helps.

[Editor’s note: We’ll be discussing this topic in-depth at the UIE Roadshow: Web Design Foundations sessions coming up this spring.]