What works and what doesn't inside Google's labs (photos)

After all the code is written, product developers put their visions to the test in front of real people inside the Google Usability Testing Labs.

James Martin

Oct. 14, 2010 8:43 p.m. PT

1 of 8

James Martin/CNET

User experience researcher at Google

MOUNTAIN VIEW, Calif.--For a familiar Internet property like Google, where billions of queries are made each day, tweaking a well-known system too dramatically can frustrate users, complicate once simple tasks, and have dramatic impacts on ad revenue and the bottom line.

This week LaDawn Jentzsch, user experience researcher at Google, took CNET inside Google's Usability Testing Labs at Google's headquarters here, one of the last stops for a new product before it is rolled out to the public.

In an industry known for backend programming, engineering, and algorithms, the difference between the coded intention of a product and the human experience can be vastly different.

2 of 8

James Martin/CNET

Inside Google's Usability Testing Labs

Inside Google's Usability Testing Labs, product development teams get a chance to see their work in action on real live human subjects, an important part of the development cycle.

Feedback in a controlled environment like this gives product teams the chance to try out new ideas and product concepts. Feedback is important. While the ultimate goal is an improved user experience, care must be taken to not upset the existing experience too abruptly.

3 of 8

James Martin/CNET

Eye-tracking lasers

Part of the user experience assessment for the Google Instant research required the labs to track the eye movement of the participants. Lasers mounted in the frame of the monitor emit light, and that reflected light is able to track the eye movements around the screen, identifying where the user's attention is being directed.

Here, the red line maps the path of the eyes, and the larger red ball shows the current point of focus.

4 of 8

James Martin/CNET

A team outlines its goals

Once a team outlines its goals and assigns tasks, test participants are brought in to carry out the assigned tasks. During the testing for Google Instant, the newest search product, which provides instantly updated search results, testers were asked to find the answer to questions.

LaDawn Jentzsch, user experience researcher at Google, and CNET's Tom Krazit sit down in the research room inside the Google Usability Testing Labs in Mountain View, Calif.

5 of 8

James Martin/CNET

Part of the user experience assessment

Two sample questions asked the testers to identify and name an image of a certain type of architectural column, and find the name of the books that make up the Girl Who Kicked the Hornet's Nest trilogy.

6 of 8

James Martin/CNET

Results

The eye tracker gives researchers the chance to see how important certain information is to each user, and whether key areas are getting the proper attention, or if users are being distracted by certain page elements.

Seated at the computer with a list of tasks

7 of 8

James Martin/CNET

Seated at the computer with a list of tasks

Seated at the computer with a list of tasks to accomplish, participants read through their processes in a research concept known as the "talk aloud protocol"--a testing technique in which the subject essentially thinks aloud, giving observers a better insight into the participant's intentions and what they are looking at, thinking, doing, and feeling.

8 of 8

James Martin/CNET

Video cameras monitor these experiments

Video cameras monitor these experiments, broadcasting the audio and video into the adjoining viewing room, where researchers, separated by a two-way mirror, can study the interaction with the product and watch what works and what doesn't in the design and implementation.

It requires a lot of work to take a product from start to finish, and often after hundreds of hours of development and back-end programming, seating a real person in front of a product can be the pinnacle moment, when real people encounter real problems. It's an element of crucial human feedback that Google has yet to replicate. We still need humans, for now.