Demystifying and Maximising Location Data for Marketers

There’s considerable buzz around the potential for using “location data” in marketing circles, but how can we maximise its potential?

 

As portable devices have evolved and their use has become ubiquitous, the technologies that support and enhance the mobile device ecosystem and experience have expanded to a similar extent.

 

One of the most notable of these, in terms of its influence on consumer end users, is ‘location services’ technology – an operating system service layer which simplifies the process of determining the location of a device through a number of techniques, including triangulation of cell towers, Wi-fi and GPS.

 

The deployment of “location services” has been well received by app developers, producing exponential organic growth in the number of “anonymous location signals” generated by digital devices. This growth has enabled a wide range of new marketing possibilities through analysis of the historical and real-time signals that are broadcast by devices.

 

By applying clever algorithmic processing on huge quantities of data-points, patterns start to evolve which provide valuable insight into consumer ‘geo-behaviors. When these observations are overlaid with information about localised factors such as the following examples, the combined data provides a priceless degree of understanding into real world consumer behavior:

 

• “What’s the average weekend spend in this area of the high-street”
• “What type of businesses are located in this office building?”
• “What’s the average house price in this neighborhood?”

 

When observations like these are considered across different contexts and time scales the results provide marketers with an unprecedented ability to measure real-world behaviors at scale, through consumer generated data - rather than limited panel based analysis or shortsighted online-behavior analysis powered by the cookie.

 

 

Location data is driving new possibilities in real-world audience targeting, analytics and in-store attribution – not only in digital but out-of-home too. Whilst this is clearly positive, a degree of skepticism exists around the accuracy of location based technologies - and rightly so, because the challenges with using location data cover four categories – and marketers who address these will be ideally positioned to maximise the return on a location data investment:

 

 

Device location accuracy relies heavily on the technique used, (IP Address Geo Location, cell tower and Wi-fi triangulation or GPS). Each approach works best in different circumstances. For example, using cell-triangulated data is not ideal for footfall attribution, but using it to detect commuting patterns works well.

 

It’s therefore vital to determine the optimum technique to use in establishing the location, in order to maximise use of the data. You also need to know how the data was established, to understand what type of analysis you can make with that data.

 

This can be achieved by using clever cross-referenceable third party databases, enabling you to determine the accuracy of a signal, and purpose it adequately for any inferences made.

 

It’s also worth remembering that raw location data is only as useful as the level of analysis you apply to it, however.

 

 

With rapidly growing Programmatic mobile inventory revenues, demand for location inventory has also increased, and driven up its price. Publishers are now aware of the potential revenue streams and are devising clever ways of slipping in coordinates into their inventory.

 

Practices include adding blatantly randomized values and many publishers, SSP’s and DSP’s have struggled to counter the problem, with publishers implementing clever algorithms to avoid detection.

 

We’re a long way from a solution to effectively counter real-time location fraud. In some cases it’s necessary to remove up to 90% of location signals to ensure accurate data remains. However, by studying historical data with the right algorithms it’s possible to locate and ‘clean’ location signals.

 

 

When using location data, you must strike the perfect balance between respecting consumer privacy and extracting the best possible value from the available data. Best practice involves working with data aggregates whilst applying statistical methods in determining an audience characteristic.

 

Carefully crafting a solution that never identifies a specific individual, yet accurately describes their traits is challenging. One approach to striking this balance is to analyze many different locations frequented by devices, and look at specifics around the patterns observed, rather than making assumptions about a single observation.

 

The final challenge in location based data solutions is scalability – brought about by location data ‘accuracy’, ‘availability’ and ‘fraud’. The way to approach this challenge is to treat each valid location record as highly valuable and nurture it to gauge as much insight about as many consumers as possible.

 

Applying computer assisted learning to larger swathes of clean location signals, enables you to detect common factors between devices. These can then be used to spread information gained about a particular device, to a second, then a third, in a similar way to the splitting of DNA cells.