Center Store
MegaStudy: Grocery
Shopper Technology Institute
Science of Shopping

PRESS RELEASE

Magic Numbers: Five Ways Network Operators Can Measure Their Audience

aka TV - August 24, 2005

ByKevin Massy

Visit www.aka.tv

Original article can be found here

SAN FRANCISCO – Much discussion of measurement in digital signage leads to more questions than answers. What is being measured? How important is a standardized metrics system? What constitutes an impression? How do you compare metrics across different networks?

In this second part of an ongoing series, aka.tv takes a look at five innovative solutions that are attempting to tackle the measurement elephant head-on, and assesses the strengths and weaknesses of each for different applications.

Seeing is believing: VideoMining The use of cameras to track digital-signage viewership is developed even further by VideoMining, a division of Pennsylvania-based Advanced Interfaces.

VideoMining’s suite of products range from generic behavior-monitoring services such as its Customer Traffic and Queue Management solutions, to its sophisticated Customer Segmentation, an application that wouldn’t be out of place in the movie Minority Report.

According to VP Jeff Hershey, all VideoMining’s products use advanced software and video-processing techniques
to derive behaviors and characteristics captured from ordinary video footage. For simple impression-counting, VideoMining’s cameras can register faces from a distance of several feet even when viewers are in motion, giving an exact count of impressions. Hershey explains that the screen-mounted cameras can be calibrated to correspond to the size – and therefore visibility – of the display, and that network operators can set threshold limits, so that only views that last over a certain amount of seconds are counted as impressions.

With Customer Segmentation, analysis or ‘mining’ of the video footage is taken several steps further to create a demographic profile of the viewers. Guided by certain data points on a viewer’s face, the application processes captured images through a series of algorithmic calculations to produce detailed data on a viewer’s gender, age range, and ethnicity (for example, Caucasian, African-American, Asian).

This information can then be used in real time to tailor messaging on the in-store displays to the specific demographics. Hershey explains that the longer viewers stand in front of a screen, the more frames the computer can process and therefore the more accurate the demographic profiling can be.

VideoMining’s products are currently being used by a number of retail clients to understand the link between certain demographic characteristics and purchasing patterns. Hershey says that one global quick-serve restaurant chain has used the application for “several years” to better target its customers by collating data on eating habits, geographical region and day-part.

To date, most deployments of VideoMining have been standalone cameras, but Hershey says that the company has begun to offer an integrated display solution with the application built into digital-signage network screens.

Good for: Providing exact viewership counts, as well as demographic profile of audience.

What it doesn’t do: Make coffee. With the availability of more dependable tracking technology and the emergence of third-party measurement agencies such as TDSA, the issue of impressions for digital-signage networks is being addressed from numerous angles. The next part of this aka.tv series will cover another challenge for retail-network operators: measurement of sales uplift.