Is digital signage effective? From campuses to boardrooms, it's a pretty common question. And the answers are typically put either in anecdotal terms ("why are people still using the mailing lists and paper posters") or in the language of management -- ROI, metrics, and life-cycles ("If you can't measure it, you can't manage it"). Both can have a role to play, depending on the environment and objectives of your signage system. But before you can consider how to measure the effectiveness of digital signage, it's important to understand what those measurements entail and whether doing them will help.
Objectives
It's first important to consider what you're trying to achieve with digital signage. For "facilitating" signage near meeting rooms, in a lobby for a building directory, or technical support queues at a help desk, the placement and usage is clear, pending the practicality of power and networking installation.
For more general purpose signage in the realm of customer experience, event information, brand building, or advertising, placement is less clear and requires some thought. Oftentimes, some basic in-person observation, a strategically-placed webcam, or even the intuition of an employee on the premises who knows where people congregate can be key in determining if a screen will be a good fit for a location.
Metal and wheels
Back in 2008, there were well over two dozen vendors on the market with devices trying to measure relevant digital signage metrics -- everything from viewer gender to time viewed. At the time, displays and computers were still on the pricier side and there were still plenty of obvious deployment locations, and a deploy first ask questions later attitude, so few of them caught on.
Today, some prominent vendors doing these analytics are TruMedia (http://www.tru-media.com), Quividi (http://www.quividi.com/), GoCount (http://gocount.net) ,and StudioIMC (http://www.studioimc.com/measurement). Several other vendors such as Intel also advertise proprietary analytics capabilities, but there's little publicly available information past the usual buzzwords. All involve a camera connected to a ruggedized computer, which does the image processing. Without getting into the products in-depth, each unit tries to determine when a user is in the vicinity of the display, if they have noticed it (the "impression" in ad-speak), and how long they have stayed there.
These systems are useful for many things beyond signage, and you can learn a great deal about the people watching your display and gain some intuition as to your system's effectiveness.
Alternative Metrics
The trouble with the camera-based analytics products in this category is that they're a bit cumbersome. They depend on quite a lot of hardware, their user interface is dated (with the notable exception of GoCount), and they don't appear to be easily integrated with other products. If something like the Density sensor (http://www.density.io/) (which counts people in a venue with a small sensor and uses a simple API) came along in the signage analytics realm, it would decimate the existing players and be easy to recommend for even a simple signage installation.
But until then, there are some alternatives that are worth thinking about. The most obvious are QR codes. They're easy to embed in digital signage content, and can be tagged to indicate that the user scanned the code from a screen. But users need to be reasonably close and at not too steep an angle to be able to scan the codes, which cannot rotate so quickly as to be missed entirely.
Bluetooth beacons are another tool that can be helpful. These tiny radio transmitters search for nearby cell phones properly equipped and can, when paired with a phone application, trigger powerful options for interacting with the screen and its content. But even without that, a bluetooth beacon can give you a sense of how many people are near your screen.
The web application for your digital signage system isn't just for submitting and distributing content. It can be an indirect way to find which content users are engaging with. When a user sees an event of interest on a screen, they may well go on the web to find the exact date and time later. Good web analytics correlated with which screens content displayed on can tell you which content is so compelling that users go in search of it after seeing it on a screen.
Right message, right people, right place
Before taking on your screen’s effectiveness with the full panoply of computer vision and web analytics tools, it's worth considering if you've gotten the fundamentals right. Without rehashing the blog post on making digital signage work in higher education (https://www.inniogroup.com/blog_posts/making-digital-signage-work-in-higher-education), it's important that a display have a diversity of content, an audience that finds it valuable and a location that integrates the screen with its environment.
Understanding what sort of content in which location will inform, engage, or entertain is far more valuable than have your analytics tell you after the fact that your screen full of ads is located in a narrow corridor that everyone hustles past each day.
A little market research in the vicinity of the display can help determine if you've targeted the audience you’re hoping for. Is your content optimized for passers-by in a prominent place or is it meant to be studied at some length by people milling around? Are you mostly showing HR announcements in the place where a queue of support tickets might be more appropriate? Just as with a conventional marketing campaign, identifying answers to key questions about audience, location, and message are crucial. A market study has to be done once. A mounted screen, computer, and supporting contractor work are your commitment to that research.
So measuring is good, right?
If your digital signage usage is primarily in brand-building or ad-revenue realms, the analytics systems mentioned here and their wealth of information may well be for you. But if your signage tends more towards providing information, event advertising, or way-finding, it may be better to do your research the old-fashioned way and look to some indirect means to gauge user interest. But keep in mind that screens, computers, and bandwidth are cheap. Any reasonably thoughtful screen deployment with relevant content behind it is likely to offer some value. Knowing the who and how many of that value proposition can be valuable in advertising scenarios, but good research ahead of time and the use of simple but indirect metrics afterward can be just as good.