Getting A Sense Of Consumers: How AIM Suite Works

Date: 
04/24/2013

Intel refers to the devices it uses for Anonymous Viewer Analytics as sensors for a simple, common-sense reason: sensors detect visual patterns, while cameras can record images.

AIM Suite technology just wants those patterns. That's at the core of how consumer privacy is protected when Anonymous Viewer Analytics (AVA) technology is deployed by retailers, brands and other companies that are looking to understand and optimize visual messaging and operations.

AIM Suite technology uses simple sensors - working with very sophisticated software in the background - to detect, count and analyze what it understands to be human faces. This pattern detection and analysis technology does nothing to recognize the faces of individuals.

What AIM Suite does is build a rich, granular sense of how consumers move through environments, their numbers and profile, and how they respond to visual cues.

How AIM Suite Works

The typical kit of parts installed at a location is a sensor - even a consumer-grade webcam - connected by USB cable to a PC. The PC has Intel's Audience Impression Metrics (AIM) software installed on it, and that software sends log files to Intel's cloud-based servers. Users subscribe to the AIM Suite service, and just need a web browser and login rights to view real-time analytics, customize reports, and manage their devices.

Customers - such as retailers with digital marketing displays in their stores - install AIM Suite technology in their environments to scan a targeted area and start to understand general consumer behaviour patterns and numbers. The sensor sends a steady pixel stream to the PC, which processes it frame by frame to detect whether the arrangements of pixels coming through that stream resembles the general pattern of human faces, using such factors as pixel density and the alignment around eyes.

A real-time view of what the sensor is doing would show the faces of people coming within range of the sensor, and looking in its general direction, being circled by the software and, therefore, measured and analyzed. The detection algorithm doing that work has statistically “learned” face patterns by being trained on an audience database of thousands of face images; it has not learned to recognize these individual faces, but rather the shapes and geometry of faces.

Every human face has distinguishable features and points that can be measured, like the distance between eyes, the length of the jaw bone and shape of a nose. These measures and points can be learned to not only understand the base pattern of faces, but also gender, and age range.

Process, Then Destroy

Every frame that gets processed by AIM Suite is then destroyed in real time. The algorithm does not have the capability to recognize individual faces, and there's no database of consumer faces kept to match against.

The anonymous data that gets aggregated by AIM Suite software allows subscribers to generate reports that provide a clear, accurate understanding of the characteristics and patterns of consumers passing through a defined area. That has a variety of implications and applications:

Real-time intelligence - Conventional audience-counting research involves humans with mechanical counters and clipboards, or mechanical counting devices like laser beams. Staffed research is costly and only tends to capture short time windows of activity, while mechanical meters can't do much more than provide gross numbers that offer few real insights. AIM Suite is a cost-efficient, ongoing tool to steadily report rich details on numbers and characteristics.

Messaging insights - AIM Suite not only counts and parses audiences, but it logs how long people look at messaging cues like posters and digital displays. With the latter, AIM Suite log files can be matched against content management system log files to determine whether different media files (like advertisements) make people watch longer. Understanding average engagement times allows brands and their creative teams to optimize the creative to the audience dynamics.

Tailored messaging - AIM Suite users are tying real-time data for gender and age range to content management systems to dynamically serve ads based on audience characteristics. For example, health and beauty brands can serve different hair care spots based on whether the viewers are male or female.

Optimized operations - Analytics generated by AIM Suite give retail and public venue operators a powerful sense of traffic and customer flow patterns. They get accurate data about the numbers of people by time of day and week, and how traffic flow in an environment can vary. That directly informs store operators on staffing - how many people to have on hand and where to place them.

The bottom line is that AIM Suite sensors are in place to sense patterns - not individual people - and deliver steady, actionable insights on what's going on.

Questions? AIM Suite Knowledge Base