IBM Power Systems are faster than the competition across all workloads and have a real

advantage when it comes to complex workloads: they process them in milliseconds. The

human eye blink is, on average, 350 milliseconds. We realized we could cut through the clutter of “speed claims” by comparing Power Systems data processing to the speed of a human eye blink.

Using a combination of facial recognition algorithms, pupil tracking, and motion

detection technology we tracked the eyes of participants and were able to capture the

speed of their blink. To make this as social and scalable as possible, we built this

experience in javascript using web technologies. It was a challenge, but using the latest

APIs, we were able to track blinks down to the millisecond. No one had ever measured

and photographed the blink of a human eye to create a dataset before. No one had ever used an amalgam of human eye blinks as a dataset for a business case before. Matching unique blink speeds to the amount of data processed by a Power Systems server was a perfect pairing of human input with a computer output.

The participant was an integral part of the Blink Test design. To measure the blink, we needed to devise an interface that guided them through the alignment process, regardless of their height. We used a camera to help users to align their face—the same camera that measured their blink speed. At any moment their eyes closed, we captured a still image to prove the abilities of the “blink capturing” system and used this image as the basis for social tiles to promote the product and experience.

Visually, the IBM Blink Test uses overlapping radial shapes to evoke speed and simultaneous processing. We then transformed these shapes into charts to tell our data story. The motif is derived from the concentric circles of the Power Systems logo and the graphic feel of the IBM brand campaign at that time.

At a conference that held 5,000 of our exact target audience, we set up our station in

the highest traffic area. 390 people took the blink test and many were there to watch

the process. Those with Twitter handles shared their results right then and there,

resulting in over 10,000 impressions and 2,151 visits to our Twitter handle.

More importantly we sent everyone who took the blink test directly to the Power8

Demo at the IBM booth. The booth had queues of people waiting to talk to product specialists. We started a conversation with an audience that could actually buy our

product but previously hadn’t been able to quantify the competitive advantage.