MAC Randomization and Location Analytics
Certain data can be collected passively from Bluetooth devices, without any active engagement with the users. However, passive analytics has limitations. Read further to understand the issues and how active analytics can help.
Available Data from Bluetooth Devices
A Bluetooth device, like any wireless device, announces itself to the world by sending advertisement packets. The most common payloads are:
- Local Name
- Power Level
- Manufacturer-Specific Data (including payloads for iBeacon, AltBeacon, and Eddystone)
However, BLE devices do not advertise constantly. For example, a phone might not advertise when ringing or when receiving a text message.
MAC Randomization
While advertising, the device can also indicate if it is using its physical MAC address or a random MAC address.
Generally, MAC randomization does not occur with BLE beacon tags or accessories such as wireless headsets. However, most mobile devices randomize their MAC address. For example, each time that an iOS or MacOS device wakes up, it uses a new, random MAC address.
Impact of MAC Randomization
MAC randomization benefits users by protecting their privacy and ensuring that the device is not easily trackable. However, if you're relying on passive analytics, MAC randomization poses some challenges.
-
Hidden MAC addresses—If you try to search the data by using a device's physical MAC address, you won't find it. Only a random MAC address is advertised, so the physical MAC address is not detectable.
-
Multiple MAC addresses—Because certain devices advertise new MAC addresses whenever they wake up, you might see multiple addresses from one device.
Active Analytics
Passive analytics has challenges due to the limited data in the advertisement packets, the fact that devices do not transmit constantly, and MAC randomization.
Active analytics can be a much more valuable approach. By developing Juniper Mist™ SDK-enabled applications, you can drive user engagement and collect more information. For example, you can collect information when users log in, click buttons, and follow links. Because the users identify themselves when they log in, the data that you collect is de-anonymized. You're counting and tracking the actual users of your application, without the issues that can result from MAC randomization.