by Tom Starnes
Digital signage can be so much more than a PC behind a large monitor mindlessly looping through a PowerPoint presentation. Adding a few buttons along the side may give the viewer a chance to re-direct the program to another series of slides, but this interaction is fairly limited and dependent upon the viewer for explicit inputs. More exotic software is being developed that distinguishes digital signage as a truly new, exciting product category. This "more aware" digital signage senses enough about a passer-by to better attract and engage the person to stop and pay attention to the signage without their having to do anything. The system can keep better track of how engaged viewers of the signage are, allowing the operators to fine-tune the messaging to specific categories of people.
Such software enhances the experience for the passer-by, who is a potential customer. Sensing and analytics can better match products and messages to viewers, greatly improving the effectiveness of messaging. Remote management lets real-time updates be made conveniently to a widely distributed network of signs. The Intel® Audience Impression Metrics Suite (Intel® AIM Suite) establishes a base platform blending viewer sensing with analytics, which can increase effectiveness for the display and make a very positive impact on profitability for a store.
Know Your Audience
The AIM Suite adds intelligence to digital signage by detecting characteristics of the viewers of the signage on a moment-by-moment basis so the content being shown can be adjusted to better fit the viewer's likely interests. Just the presence of people looking at the sign can trigger software to jump into action. With AIM, that action can be a carefully planned sequence based on characteristics gathered about the viewer to better target their potential interests. Showing women's clothing to male viewers is not likely to generate a sale. Younger people often look for different things in the same products than older folks do.
Some fairly basic information about people in front of the digital signage can be quite useful. Is the person male or female? Are there children, teenagers, adults, or seniors? Are they passing by or pausing to look, and for how long?
Part of the AIM Suite software runs on the local digital signage hardware, classifying people viewing the sign by using Anonymous Viewer Analytics (AVA). Optical sensors on the display feed images to the processor which are analyzed to find faces in the image that are turned toward the display. Size, shape, and relationship of key features of faces and bodies in the frame are evaluated to determine how many people are there, how far away the subjects are, and whether they are facing the display. This is similar to what is done to determine focus and exposure in some digital cameras. The height and proportions of people help classify their age group. Bright and dark spots are evaluated to assess gender and age. Patterns that look like facial hair or a bald head are an easy indicator of men. Tracking the movement of subjects in the frame over time gives important information about whether people are paying attention to the signage or just passing by.
People lingering at the signage, facing the display, are probably interested in the subject matter (or at least curious enough to see what will come up next). If a person moves away after a few seconds, the follow-on image may have soured the person to the subject. A pattern of such abandonment may indicate that more work needs to be put into the follow-on image or the flow of the messaging. Was it too much information or just repetitive? Maybe this was the point where available colors or styles should have been shown. It might be time to start spelling out key features.
Intel is quick to highlight two important points about how the AIM software evaluates images. One is that it does not store any personally-identifiable information, and the other is that it uses algorithms developed to broadly classify people, not techniques that are designed to identify individuals as one might do looking at drivers licenses. Words like face detection, facial representations, audience characteristics, audience detection, demographics, and indications are used rather than facial recognition, matching, or identification. Privacy should not be a concern here.
These systems are not trying to pick out return shoppers, but are hoping to avoid wasting time showing peek-toe pumps (these are shoes) to men or socket wrenches to senior women. Guys can get the creeps walking through the ladies underwear department, which we're inevitably forced to do. Broad generalizations may imply certain biases that can be wrong, but their use is common in general merchandising. TV advertisers don't need to know that John Smith at 123 State Street watched the latest reality show, but they do want to know how many 18- to 49-year old females tuned in.
Knowledge of the makeup of the audience is key to effective advertising. The subject matter as well as the presentation are important for getting through to the audience. Action and music may grab the attention of younger viewers, but a more businesslike attitude may work better on some men. Fun and fashion may rule over facts and figures. Some groups are captivated by a lot of quick pictures, while others prefer slower-moving traditional imagery. It seems that pink versus blue is over-the-top for gender affinity, but such matters are beyond the author's forte.
The data gathered may not be perfect, but it will still provide insights by its volume. Hats, glasses, and bulky clothing may give false classifications. Three girls facing the sign may be talking to a fourth whose back is to the sign. They may be paying no attention to the sign at all and soon wander off. But if the products, music, and style of the images on the sign changes to something closely targeting teenage girls, the signage may soon catch their attention – exactly the intention of the digital signage – and a sale or three may be imminent.
Computer-based kiosks and digital signage can bring shoppers back to brick-and-mortar stores to browse, learn about products, and make their purchases right then and there. In recent years shoppers may have tried to do it all sitting at home on the Internet, stuck with only whatever information they could dig up there. Then after a few day wait for delivery of their order, they might find it wasn't what they had expected or it didn't fit. Interactive digital signage can bring the best of both worlds together at the local retail store.
While you can't go to the store in your pajamas, networked digital signage and kiosks can bring the wealth of information available on-line to a point in the brick-and-mortar store, near the merchandise and where a real person can help when the prepared information falls short. Shopping for clothing on-line, a person can see all the styles, colors, and sizes that are available. The in-store digital signage has the same capabilities when connected to a real-time database. Enhanced with the intelligence gained from the AIM Suite, the display and messaging should be more effective at offering suggestions than a user-driven browser. The added benefit in-store is that the shopper can immediately go see the actual clothing, feel the cloth, make sure the color is just right, check the workmanship, try it on to see how it looks, and narrow down the best size.
Shopping instead for new consumer electronics, after checking out detailed feature lists and comparisons and perhaps seeing a demo of a function or two from digital signage in a store, a patron can walk over to an actual system and try it out. How do those buttons feel? Is that screen bright enough? Can you feel that bass in your bones?
The key is that the resources of the digital signage can give the depth of information available on the Web that isn't on the package label or the placard on the shelf and might be difficult for a clerk to keep in mind for each product. Computer-driven signage can present information accurately, hitting the best selling points, fine tuned for the shopper according to characteristics determined by the AIM Suite.
The physical store gains two big advantages over pure Internet shopping. One is that a real person is available to answer questions, address negatives, demonstrate personal favorite features, and redirect to a better choice. A nurturing salesperson may still be needed to "bring home the sale," including assuaging concerns about the price, but on-line retailers don't get that human touch – the last ditch "are you sure?" banner when a person clicks off on their Web browser is pretty desperate. Even if the sale is lost, a live salesperson should at least come away knowing why.
The second advantage is that of instant gratification. Five minutes later, the store customer can be driving home with the new dress, sunglasses, or audio gear and can be showing it off later in the day. If the party is tonight or the water heater breaks, it can be a long two days waiting for the delivery truck to arrive. Again, a good salesperson will work this advantage.
For Your Information
Situations suggested here have mostly been set in a retail environment, but the concepts also work in many other applications. Increased sales may not be the direct goal of all applications, but digital signage taking advantage of the AIM Suite will improve effectiveness in more informational applications. Engaging the viewer is the key, and the better the system can classify the viewer, the better it can present the information in an easily digestible way.
Digital signage in museums, guidance, tourist information, and training all should be as useful as possible. More text. Lists. More audio. More pictures. More animals. More action. Music. Fred Astaire, John Travolta, or Psy? More interactivity. Simpler verbiage. More diagrams. Senior discounts, wheelchair accessible. Highlight neighborhoods, colleges, museums, restaurants, or bars? Matching content and presentation to the viewers can improve how readily information is absorbed in most applications. Effectiveness on some applications might be measured by how quickly viewers leave the signage rather than how long they linger there. After all, nobody wants to miss the train because they were still trying to figure out which train to be on.
Tools of the Trade
The Intel AIM Suite 2.2 is the current version of the software. Hardware based on 3rd Generation Intel® Core™ processors will provide the highest performance, quickest response, and most dynamic graphics for the content. With a 60% improvement in graphics performance over Sandy Bridge, many systems were able to eliminate independent graphics processors from the works. As always, buzz of even greater CPU capabilities is on the horizon.
For content creation, Flypaper offers the Intel OPS KIT which lets users develop and then manage Flash, video, and other media-rich content for digital signage. The Intel AIM Suite is integrated into this software with an AVA Component which lets the producer define their own rules within the content.
Take Aim at Customers with Digital Signage to Improve Brick-and-Mortar Sales
Digital signage takes a step forward using optical sensors with Anonymous Viewer Analytics in the Intel® Audience Impression Metrics Suite to characterize potential customers to better target them with appropriate messaging. The AIM Suite is a starting point, but hardware and additional software needs to be assembled for a fully managed system, and significant creative content will need to be developed for individual applications.
Solutions in this blog:
- Kontron Digital Signage Evaluation Kit
- Lanner LUGE LEC-7000 Fanless Embedded System
- Venture eIPC300 Embedded Industrial PC
- Flypaper Digital Signage Version
Kontron is a Premier member of the Intel® Intelligent Systems Alliance.
Lanner is an Associate member of the Intel® Intelligent Systems Alliance.
Venture is an Associate member of the Intel® Intelligent Systems Alliance.
Flypaper is a General member of the Intel® Intelligent Systems Alliance.