Real-World Analytics

Personalization, Visitor Tracking, Visitor Profile, Traffic Source, Visit Duration, Bounce Rate, Targeted Ad, Funnel Tracking, Impressions, Clickthrough, Conversion, A/B Test…

If you were told that this article directly relates to the terms above, you would be forgiven for thinking it was about the Internet. It’s not. It’s about going to the mall, a restaurant or a gas station. It’s about driving down the freeway, going to a movie or listening to the radio.

A collection of new technologies has brought us to a point where real-world behavior can be captured, aggregated and analyzed in the same way that online behavior can. All of the terms above can now apply to activities that occur offline.

Take a minute to let that sink in.

With most technological advances, our understanding of how to best take advantage of that technology evolves as the technology itself matures and becomes more available. Because real-world tracking is so analogous to online tracking, the situation here is different. For example, we already have a commonly understood set of terminology related to web analytics that can now directly apply to real-world activity. For example, targeted ads online are frequently billed as “pay per impression” or “pay per click.” With real-world tracking, those business models are now possible for traditional, real-world advertisements. That business model doesn’t have to evolve with the capabilities, because it already “evolved” online. The new tracking technologies enable businesses to apply all that they have learned about online tracking and personalization to their physical businesses.

An Example Scenario: The Mall

Consider the following: A person goes to a mall and visits a clothing store targeted at preteen girls and makes a purchase. They then visit Victoria’s Secret. After leaving Victoria’s Secret, they notice a digital sign promoting a sale at a women’s shoe store in the mall. They then visit the shoe store and make a purchase.

Image of an advertisement board inside an airport
Two types of tracking technologies are being used in this example: passive device tracking (which tracks your cellphone location via Wi-Fi) and optical tracking using a single camera in the digital sign. Both of these technologies are currently available and in use. For more information related to the various technologies available, see The Coming Revolution of Real-World Analytics infographic below. Now, let’s examine that seemingly simple scenario while keeping those established web visitor tracking terms in mind.

Subject Enters Mall and Visits Stores
The entrance to the mall begins the visitor tracking. Just as with web analytics, as they go from store to store, we can track where they came from (traffic source), how long they stay in each store (visit duration) and purchase activity (conversion). From this information, a visitor profile can be built. Based on the scenario, the system would surmise that the visitor is a female mother of a preteen girl.

Subject Enters the Vicinity of Digital Sign
In the scenario, the digital sign has been developed with the ability to use the visitor profile to display a targeted ad. This is a form of content personalization. In online ad unit terms, when the shoe store ad is displayed, that is counted as an impression.

Subject Notices Ad
In online tracking, there is no way to tell if a visitor actually notices an ad unless they interact with it in some way. However, in the physical world, the fact that the ad has been noticed can be tracked with a fairly high degree of accuracy. This is accomplished through the use of camera-based tracking embedded in the sign. In web analytic terms, this is probably the point at which funnel tracking begins.

Subject Visits the Advertised Store
Going to the store itself would be tracked as clickthrough.

Subject Makes a Purchase at the Advertised Store
This would be tracked as a conversion, and complete the funnel from the perspective of the shoe store.

With real-world customer tracking, we aren’t cautiously maturing with the technology. We are being born into adulthood. Retailers have seen and experienced the huge benefit in online-targeted advertising. They are already intimately familiar with the cost-per-impression/cost-per-click business model. There are no technical barriers that prevent that same business model from being deployed into the physical world today.

There are, however, ethical and legal issues that need to be worked out. It’s important to note these are “issues” and not “barriers.” Currently, there are no laws specifically prohibiting the type of tracking being discussed here. However, there have already been a number of forays into real-world visitor tracking that have led to consumer backlash. While these ethical issues are more than worthy of discussion, it will be a far more effective discussion if it takes place with an unfettered understanding of what is possible.

So…what is possible?

A Deeper Look at Our Example

As stated before, in our example scenario there were two technologies at work: passive device tracking using Wi-Fi and optical tracking used at the digital sign. By having these two technologies tied to the same tracking system, a number of useful real-world tracking features, over and above those demonstrated in the example, are made possible.

Repeat Visitor Tracking
Every device that can be connected to a network has a unique ID called a media access control (MAC) address. When Wi-Fi is used for tracking, that MAC address is recorded. This unique “signature” gives the tracking system the ability to track customers across multiple visits. The obvious benefit of this capability is the ability to build a better, more targetable profile over the course of many visits. The less obvious use involves building a system that watches for, and records, relationships between multiple devices. That approach would enable the system to know a number of interesting, and potentially useful, things:

  • Is the visitor currently shopping alone or with others?
  • If with others, are those others consistently the same shopping companions?
  • What’s in the profiles of the others? Are there common interests?

What if, through repeat visitor tracking, the system established a relationship between a mother and her teenage daughter. It then observes that on Saturday they visit a specific store together two times without a purchase. Then, on Sunday, the daughter returns to the mall with other “known profiles” (her friends) and visits the same store, still without making a purchase. On Tuesday, the mother goes to the mall by herself. What if the digital sign presented her with an ad for the store in which her daughter was interested? That ad could go a little further and entice her with a call to action to “Scan this QR code and give us your email address for 25% off your next purchase.”

Bounce Rate
Online, we frequently judge a page’s success by looking at its bounce rate. A bounce is recorded for each visit to that page where the visitor ends that visit by clicking the back button or closing the browser window/tab. In our example scenario, if the customer had visited the shoe store and left almost immediately, then that tells us something similar. The ad they saw on the digital sign was appealing enough to get them to the store. So, what was it about the store that didn’t fulfill the expectations the ad set up?

A/B Testing
Digital signs equipped with cameras that are used to trigger content, and even measure that content’s effectiveness, are becoming more and more common. However, if we consider the possibilities enabled by tying that sign to a larger Wi-Fi tracking system, as in our scenario, the possibilities multiply. Now the system can A/B test different advertising options to different profiles and immediately capture their effectiveness. With just the camera, the sign is limited to recording only when it is successful in getting someone’s attention. Tied to the Wi-Fi tracking system, it can now also measure if those that pay attention to it are visiting the store (clickthrough) and making a purchase (conversion). That same sign could also change how it displays ads based on overall traffic pattern, the average age of the shoppers passing by or the number of individuals within viewing distance.

But That’s Only Scratching the Surface

As impressive as our scenario is, it is only a very small part of the possible real-world analytics ecosystem. Let’s look at a few other brief scenarios.

Tracking Billboard Effectiveness
Cars are commonly tracked a number of different ways: RFID (toll tag), license plate (optical) and Wi-Fi (passive device tracking). Those same technologies can be used to “tag” vehicles that are within viewing distance of a digital billboard when a specific ad is displayed. This idea has been around for some time now. However, now that “wide-area tracking” can be tied to the visitor tracking from our mall scenario. This gives us the ability to not only track impressions (who saw the sign), but also clickthrough (who entered the store) and conversion (did they purchase). Apply the same metrics to two or more ads for the same promotion, and you have A/B testing.

The ability to track such things will impact how retailers pay for billboard ad space, possibly moving billboards from an impression-based fee to a pay-per-click model.

Passive Experience Tracking
“Second screen” apps, those that use the audio from a mobile device’s microphone to deliver synchronized content related to a show someone is watching, are becoming more common. With that in mind, consider that newer Android phones are touting the device’s ability to always be listening for a verbal command so the device can respond to those commands “hands free.” The next logical step is to give apps the ability to “register” for specific audio cues so they could better target content based on hits against their “audio cue list.” This type of passive, in-the-background, always-on experience tracking will become more common in the future.

This type of advancement would allow business owners to know that 25 minutes ago, the person buying something in their store had been in their car listening to the radio when one of their spots played. Just like the billboard example, radio and TV spots could potentially move from an impression-based model to a pay-per-click model.

Linking Online with Offline
Remember the “Scan this QR code” idea? The customer in the scenario was not scanning a generic QR code. They were scanning one that was created on the fly. This would allow for an individualized special offer. However, it also would allow the retailer to connect that customer’s Wi-Fi tracking profile with that customer’s online profile. Then both online ads and real-world ads could be further refined and work in concert. Has the customer been looking at a particular item repeatedly online? Display an ad for that item on the digital sign. After they make a purchase at the store, display personalized ads for relevant accessories the next time they visit your website.

Here is a real-world example: GameStop currently has a program where customers can track their personal inventory of games using their GameStop account. GameStop then encourages them to trade those games for newer games as they come out. What if, when a member walked into a GameStop store, digital signs in that store displayed ads based on what they currently own? Those signs could avoid ads for games the customer already owns, but choose to advertise others in a similar genre.

Infographic detailing The Coming Revolution of Real-World Analytics

Where to Go from Here

Here are three takeaways for brands to consider:

  1. The debate about proper use of these technologies is coming. Don’t leave that debate to others. Discuss the issues internally, and engage in the debate externally.
  2. Effectiveness in that debate will greatly depend on understanding the technologies involved. Continue learning about them, and watch the media for new developments.
  3. Acknowledge that these techniques will be used, and that, if the early adopters can avoid a backlash, the organization that uses them first will have an advantage. If you decide to be an early adopter, plan ahead for how to deal with any potential fallout.

Most of these tracking techniques are already in use, but we are reaching a point where the convergence of cost-effective hardware, the ubiquity of mobile technology and the software to tie them all together will accelerate the adoption of more and more capable tracking implementations. With careful consideration of the public perception issues, these technologies can be utilized to provide many more options for brands to find and interact with their audience.


  • Private Eyes: Are Retailers Watching Our Every Move?

  • E-ZPasses Get Read All Over New York (Not Just at Toll Booths)

  • Nordstrom Using Smart Phones to Track Customers Movements

Welcome to the Post-Mobile World