Method 1: Read the LogsWhile a glimpse into the practices of some of the largest and longest-running digital media networks is certainly interesting, much of their success can be attributed to simply using common sense, and picking the best tools available given their circumstances. As for the actual techniques and practices, many of them were covered in our five-part series on calculating digital signage ROI, which might be useful if you haven't checked it out before:
On most modern digital signage systems, playback log data is stored centrally, which gives us a complete history of every content segment ever played on every screen in every store that we're tracking. This data allows us to create complete historical models of which brands or products are represented in different parts of a store during different times of day, days of the week, etc. With access to a retailer's traffic data or register receipts, we can also calculate the specific effect that a content segment has on close-in sales, along with making projections about how the messaging impacts longer-term buying patterns.
While the amount of data can make this a daunting task at first, patterns may emerge over time that allow you to take some shortcuts, while still reaching accurate conclusions. For example, while testing a network of self-service kiosks in a chain of 310 retail stores, one of our clients was able to make statistically significant predictions by monitoring a set of only ten representative stores. By doing so, they significantly cut down their research time and costs without lessening their decision-making abilities.
Method 2: Ask the Viewers
Not all digital media networks are built alike, though, and those network managers that lack the ability to generate comprehensive playback reports will have to turn to alternate methods. John Kyle, owner of Kyle Private Networks, has been battling this problem since 2003, when he acquired PharmaSee TV and BEVision from in-store TV network operator RMS. His two networks comprise more than 1,000 screens deployed in retail locations across the country, and are powered by one-way satellite connections that can't provide any information about playback frequency or impressions delivered. To compensate for this, Kyle asks his stores to provide playback affidavits, and has developed a series of questionnaires that are mailed or faxed to his locations every few months to glean additional information about how his networks are perceived by store patrons and employees alike. While a decidedly low-tech solution, the data gathered through this channel has been invaluable to Kyle's content development team. For example, after one round of surveys, 95% of respondents cited employee education as a key big benefit of the systems, especially in stores that sold multiple brands of products that are largely undifferentiated to the untrained eye.
After learning this, Kyle's team started to produce segments that focused more on educating employees about key products, and a re-evaluation several months later indicated that their reactive technique was the right one: on average, their new content produced a 30-35% lift in sales of the advertised products, based on signed affidavits from the owners and managers of 56 stores representative of the entire chain. Sales data for both the test and control group stores was compared against the previous year's data to help identify and eliminate any natural fluctuation in sales of the tested products. Additionally, over ten separate product lines were tested this way in an attempt to get more accurate results about the impact of in-store media in general, instead of its impact on the sales of one particular type of product.
Method 3: Call in the Cavalry
Without a networked digital signage package that allows for playback logging, tracking the performance of large networks can be difficult. However, some companies have found success by turning to firms like Arbitron and ACNielsen, who are normally faced with the even larger challenge of tracking nationwide TV and radio networks. Take Accent Health, for example. This firm provides a health-oriented narrowcast network that reaches some 10,000 doctors' offices and health clinics. Relying on a "sneaker net" of couriers manually distributing DVDs to each location every month, Accent Health currently has no way to directly measure the number of impressions it delivers. Instead, they have turned to Nielsen New Media, a group dedicated to finding creative solutions for measuring the impact of non-traditional media. Since it would be logistically (and financially) challenging to directly measure all 10,000 viewing destinations to monitor things like dwell time, impact and recall, Nielsen instead compiles a statistical average based on observations from a much smaller group of representative locations. Depending on the exact nature of the network being monitored, Nielsen might deploy secret shoppers, exit pollsters or even electronic surveillance equipment to gauge how target audiences perceive and react to the content being displayed.
1: Calculating Digital Signage ROI: The Ground Rules
2: Calculating Digital Signage ROI: Understanding the Limits of Your Data
3: Calculating Digital Signage ROI: 3 Metrics that Matter
4: Calculating Digital Signage ROI: Methods to Gather your Data
5: Calculating Digital Signage ROI: Managing Expectations
As tracking technology becomes more sophisticated and useful, I'm sure that it will continue to find its way into the ever-growing list of at-retail media networks and other narrowcasting applications. At some point, though, privacy may become a real concern. Given the sophistication of today's private label credit and loyalty card systems, many retailers already know a lot about their customers' shopping habits, from the average number of trips per month to the items most likely to be purchased. But when combined with that extra level of detail - customers' exact motions within the venue and the precise amount of time spent in each aisle or looking at each product - these systems may get too close for many people's comfort. I'm very curious to see whether people will perceive these emerging monitoring techniques as unauthorized surveillance, or if they will be willing to give up some privacy in return for a tangible benefit -- such as lower prices, more targeted discounts, or simply a "better" shopping experience.