Debunking Myths in Location Data: Sourcing and Collecting

    1024 576 Gravy
    Debunking Myths in Location Data: Sourcing and Collecting

    Every day, Gravy Analytics processes billions of anonymous location signals from consumer mobile devices to build our industry-leading location intelligence solutions. We aim to separate truth from fiction everywhere in our business because – contrary to what some would claim – there is great value in location data. When properly aggregated and processed, location data and the information it provides gives unprecedented insight into the offline consumer journey. This intelligence can be used by businesses across a variety of industries to acquire new customers, learn more about their existing customers, and stay one step ahead of the competition.

    In this 3 part blog series, we break down the myths of location data from sourcing and collecting, to fraud and privacy, and lay out what sets Gravy apart.

    Myth: There are problems with how location data is sourced today.

    To start, consumers need to first grant permission to use their location information, and only a fraction of publishers enable location sharing.

    What’s another piece of information we think consumers should know about? How location data is collected. More on that below:

    Truth: When consumers first install and run location-enabled apps on their mobile phones, such as those providing maps or driving directions, they are asked to opt-in to location sharing. This means that users must explicitly grant permission for an app to access their location information, which ultimately helps to protect consumers by ensuring that they have been informed that 1) their location data is being collected, and 2) the appropriate rights are in place for data use.

    Methods of Location Data Collection

    An SDK is a code within an app that has access to the location services provided by the phone’s operating system. An app may use location services to improve the app experience, for example, to find a store, gas station, or event near you. There are two modes in which SDK location data is collected:

    1. SDK

    • Continuously: Data is continuously collected, even when an app is not open (in iOS, this is called “Always”). The data collected is then relatively dense, with many distinct locations each day.
    • Intermittently: Data is collected only when an app is open and in use (in iOS, this is called “While Using this App”.) This results in location signals that are more sporadic, or sparse, but still valid.

    2. Bid Stream

    Bid stream data comes through bid requests to place an ad within a mobile app. Bid requests are only sent when the app is running and may contain location information to support real-time location-based advertising. If an app has access to a device’s location, it will use that location. If not, the location is estimated. 

    When considering foot traffic analysis, it’s important to note that web data, while available, is too low in accuracy.

    Myth: It’s unrealistic to scale location data given how hard it is to source it.

    Even if someone is using apps with location sharing enabled, very few will walk through a store or restaurant with an app open.

    Truth: Our data shows that anywhere between 0.5 percent and 5 percent of users open apps while visiting places of commercial interest. While it’s true that location data is sample data, the data volume is still massive and substantially larger than any other sample data set – including survey data and counter data.

    To dive a bit deeper here, there are two modes in which location data is collected:

    1. Dense – When a user allows data to be continuously collected by an app, most, if not all, store visits will be recorded. The user does not necessarily need to have the app open for the data to be collected. 
    2. Sparse – When a user allows data to be collected only when the app is in use, the app must be open when a user visits a store or restaurant to record a store visit. While most users don’t walk in with an app open, a few do.  

    The combination of these two collection methods results in a statistically significant sample that can be used for analysis. At Gravy, we see signals from about 250M devices on a monthly basis, and – after deduping and removing spoofed signals – about 2B deterministic visits to commercial places of interest in the U.S.

    Ready to get started with Gravy’s data-as-a-service? Learn more here or contact us today!

    %d bloggers like this: