CEO and cofounder of Unacast, a human mobility insights platform.
The world has an abundant supply of raw location data. It flows from every smartphone, wearable and connected device on the planet. But in its natural state, location data is of limited commercial value. It requires sophisticated refinement for use in the market, and it is not privacy friendly.
That refining process includes filtering for false or duplicate data, strict anonymization and security protocols, supply correction, the aggregation of datasets for application in various use cases, such as real estate investment, and understanding retail store performance.
Typically, these processes require advanced technical teams. However, due to the natural noise in location data and its sophisticated refining processes, the acceleration of location intelligence and analytics offerings leaves room for inefficiencies.
In the age of AI, utilizing machine learning algorithms can overcome all mentioned limitations and build future-proof business insights. Machine learning collects various data sources, such as location data, demographics and weather data, and builds a relationship among all these sources toward a target source, like revenue data.
When enough training data is available, the machine can automatically learn these relationships and thus estimate how variations in the data sources impact the target source. These machine learning models allow businesses to make more informed decisions sooner and with greater capital efficiency.
Machine learning-powered foot traffic data can be relied upon to be more accurate and faster, which is something we are consistently told is key to our clients so that they can invest, divest, build, go to market or pivot sooner.
In fact, it is now standard among leading multinational enterprises to use machine learning to inform their businesses. To illustrate, here are a few snapshots of how machine learning-powered location data is helping to disrupt three different industries.
Real Estate Investment
In the context of real estate investment, machine learning algorithms can be used to study a given neighborhood or zip code to get an understanding of who is there and how that is changing. Machine learning models can add additional value here by predicting or even forecasting those insights based on the collection and combination of data.
The output value for the investor includes holistic insights that are intuitive to use and based on all relevant information and the ability to create a better story around an investment opportunity. With machine learning, the investor can go from researching properties to writing deals faster.
The insurance world is rife with use cases for machine learning and foot traffic. Perhaps one of the more pressing has to do with devising underwriting strategies for business interruption coverage. The technique is to leverage mobility data to understand the real-world impact of some catastrophic event on a business’s foot traffic and, therefore, revenue.
A new way of doing this includes incorporating location intelligence in parametric applications, meaning insurance products that are triggered immediately by an actual event rather than an extended claim process. This helps insurers become more efficient and confident in their underwriting process.
Retailers are most often concerned with using machine learning-powered foot traffic data sets to extrapolate mobility patterns and measure visitation trends to their physical stores. These insights help retailers understand store performance, perform trade area analysis and determine the best sites for their next physical locations.
The same models can be used to benchmark performance and understand how individual stores are performing relative to the industry category, competitors and other brands in a market. This helps retailers measure market share and regional dominance over time, in addition to staying ahead of the competition in identifying new markets and opportunities.
To help you get started, here are a few “must-do” best practices for using machine learning-powered location data in your organization.
• Ensure Data Quality: Quality of data is paramount when it comes to machine learning applications. This is especially true for location data. Before diving into machine learning, organizations must ensure that the location data they are using is accurate, up-to-date and relevant to their specific use case.
• Choose The Right Tools: The choice of machine learning algorithms and tools can greatly influence the success of location data intelligence projects. For instance, some machine learning algorithms are suited to handling large volumes of data, while others are better at dealing with high-dimensional data.
• Address Privacy Concerns: Location data can be sensitive, and its use must comply with privacy regulations. A robust data governance policy that respects privacy while enabling meaningful analysis is a must.
• Build Or Buy: Building a solution in-house requires considerable expertise and can be time-consuming, but it also provides greater control. Buying a solution can be quicker and requires less technical expertise, but it may not be as customizable to specific needs.
• Preparing For Implementation Challenges: Like any other technology, implementing a machine learning solution comes with its share of challenges. These could range from technological hurdles, like integrating the solution with existing systems, to managing change and upskilling staff.
• Securing Buy-In: To secure buy-in from stakeholders, focus on communicating the benefits of location data intelligence. Use case studies, success stories and data-driven arguments to make your case. Highlight how investing in this technology can lead to better decision making, improved operational efficiency and new insights that can drive business growth.
Machine learning-powered location data intelligence is a powerful tool that can offer transformative benefits. However, like any tool, it needs to be used wisely and responsibly. The best practices outlined here provide a starting point, but each organization will have its unique journey.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Read the full article here