The promise of measuring attribution across any advertising medium is within reach.
The solution the mobile advertising industry historically focused on used your phone’s lat/long coordinates. There are two big flaws with this approach.
Accuracy – According to quarterly studies by Thinknear, roughly 35% of lat/long coordinates are accurate to within 100 meters. That’s the size of a football field, and not very accurate.
The remaining 65% are so inaccurate they can’t be placed within a 100 meters of their actual location. If you’re trying to demonstrate attribution, this lack of accuracy should shatter any confidence you have in the results.
Sniffing the Data – Most companies attempting attribution off of lat/long source their data not directly from the apps themselves, but rather by sitting on the ad exchanges. They build enormous systems to sniff the trillions of data points flowing through the exchanges, filtering out the fraudulent and inaccurate lat/long points to then plot on a map. As a testament to their engineering departments, they’ve solved the engineering challenge. But when you start with inaccurate data, you get bad results.
But now, there’s better data available. It will soon become the most requested and most valuable location data: beacon data. We believe that it’s not a device’s location that is the next “cookie”. Rather the Bluetooth beacon becomes the next cookie, or more appropriately, the real-life equivalent of a digital tracking pixel.
In the last month, we’ve seen announcements from RiteAid and Walgreens that they’re deploying beacons by the thousands. Macy’s installed over 4,000 beacons in the U.S, with other retailers moving from test phases to full-scale deployment. Business Insider forecasts that retailers and companies will deploy 4,500,000 beacons by 2018.
When you examine this growing network of beacons outside of the context of “we must use beacons to send people push notifications”, a significant opportunity emerges. Placing beacons inside millions of locations across the globe becomes the equivalent of installing tracking pixels on website pages. As smartphones bump into these beacons, we see the same behavioral trail form in real-life that we see when people navigate from website to website.
Measuring beacon bumps as foot traffic has two advantages. First, beacon detection operates passively in the background without the need for an app to be opened. Second, it brings location accuracy to within meters. This creates bigger attribution audiences and shows lift, regardless of the advertising medium. Campaigns on television, radio, print, website, and mobile create more store visits, easily measurable based upon smartphones bumping into beacons. Combine these beacon bumps with lat/long coordinates derived from apps and retailers have a potent one-two punch to measure attribution.
Using beacons for attribution does face a challenge. Some beacon manufacturers, such as Gimbal, use their own proprietary Bluetooth protocol, instead of the iBeacon or Eddystone standards. This creates a headache for companies looking to maximize their return on investment from beacon deployments. No other app can detect that beacon, unless they’ve installed that specific SDK and received explicit permission from the beacon manufacturer to detect those beacons.
Retailers should seriously consider the open and secure standards of iBeacon and Eddystone for their beacon investment. Imagine the headache that retailers will face when five different beacon companies want their five proprietary beacon signals in use. Not only will they have to manage five different beacons, but also five different SDKs. Any outside app partnering with that retailer will also have to keep up with these SDKs. Need to make a switch to a different beacon provider? Brands will incur the cost to switch out the proprietary system. This isn’t a sustainable approach.
Here’s why this ultimately matters.
Retailers acknowledge that their mobile app audience alone isn’t big enough to deliver enough return on investment from their beacon deployment. A recent article by PYMNTS.com describes Walgreens CIO Abhi Dhar’s approach to their nationwide beacon deployment:
While the beacons currently only work by sending coupons and other promotions through the official Duane Reade app, Dhar isn’t a believer in a closed beacon network that ultimately hurts the user experience. “Whether they use our app or a third-party app, we don’t care as long as we have fulfilled them in our stores,” Dhar said.
Dhar understands that there’s incredible value in letting other apps detect the Walgreens beacons. This value will be in additional audience data, building retargeting audiences, and showing attribution. Walgreens will look to partner with mobile apps that have both sizable and engaged audiences. Partnership becomes a crucial element to increasing the ROI of beacon deployments by building bigger audiences for retargeting and attribution.
Brands and retailers should view their beacon investment as much more than a real-time push notification system. They have a revolutionary opportunity to leverage beacons to see attribution, and ensure their advertising dollars are well spent.
One of the great promises of digital advertising is the ability to prove your advertising campaigns actually work. When done right, you attribute which online campaigns resulted in online sales. But is it possible to measure the impact of digital advertising on driving new foot traffic to stores and increasing physical retail sales?
One of the simplest approaches to measuring online to offline attribution still works: coupons. Create a campaign, assign it a specific coupon, measure how many that customers redeem.
Usage – The vast majority of app usage doesn’t occur while walking around a store. Yes, some small percentage of people open apps while shopping. They are bored or using a shopping app. This is the exception, not the norm. This attribution model only captures a very small fraction of people exposed to the campaign.
Without sounding too harsh, we must give credit. People are actively trying to solve attribution on mobile, and using lat/long made the best use of the data available at the time.