In recent discussions with industry leaders and as outlined in our recent thought leadership piece “Evolution of Real Estate Network Planning in the Age of AI”, a key barrier to any organization investing in AI is the absence of a focused data strategy. The practical application of AI to Real Estate planning requires the spatial data that fuels location intelligence discussions and applications. It's evident that harnessing and enriching spatial data - either existing or new - form the cornerstone for any meaningful and valued location intelligence applications powered by AI.
In a recent Forbes article, “2024 Is The Year of AI, But Data Will Steal The Show” by Afif Khoury, CEO of SOCi, Mr Khoury expects the top trends to be focused on data:
“While the AI revolution is real, promising significant changes in various industries, the transformation will unfold in stages—much like the digital, social and mobile eras. Given your use of AI is only as good as the breadth, quality and security of your data, the initial impact of AI will be most evident in the first wave of technologies affecting how organizations collect, ingest, analyze and secure data.”
In our research, we heard elements of this as well. Real Estate executives and GIS experts agreed that in the absence of a holistic data strategy, the organization suffered from three typical roadblocks to leveraging AI effectively:
1) data silos within the organization - non-integrated spatial and other data sources,
2) missing data - incomplete consumer or real estate supply data,
3) acquiring unique sources of new data for competitive advantage.
Data Silos
For those with a stated AI strategy in place - Centers of Excellence established - developing and integrating internal data “silos” from across the organization was the first major initiative. Most of these executives have employed cloud solutions to bring together disconnected data stores from across the organization - spatial (location) data, sales data, customer data, competitive intelligence, and the vast amount of other organizational data. Those who have not done this continue to struggle with siloed data marts reducing their ability to leverage AI technologies fully and creating promised value. Furthermore, siloed data marts impede collaboration with other Data Science (DS) teams, leading to operational bottlenecks, inefficient Real Estate processes, and perpetuating the traditional role of the GIS team as a spatial data “servant” to the rest of the organization.
In fact, in a number of interviews, our GIS professionals worked in current and previous organizations where the spatial data marts were managed and maintained by GIS analysts because other data scientists in the company were not trained to use this kind of data. Accessing and using spatial data was therefore controlled or restricted. In some cases, GIS analysts entertain requests for spatial data to be appended to loyalty data for marketing analytics professionals to use for other programs. The lack of integration and collaboration between GIS and Data Science teams means the cross-fertilization of methodologies cannot be realized and the data will remain in silos!
Missing Data
Another critical obstacle that emerges in Real Estate is what some have called “Missing Data”, especially on the listing supply side. As one of our interviewees put it, “approximately 30% of market intelligence remains unreported, residing within the realm of brokers, landlords, and developers - sometimes called pocket listings.” As such, AI/ML models as well as the resulting decision are biased towards visible data.
In addition, the official statistics from government bodies are typically delayed and do not incorporate recent or upcoming developments (eg. immigration policy that alters the fabric of neighborhoods, housing policy changes, etc.) and are therefore unaccounted for in official data sets that significantly impact local demographics. Again, boots-on-the-ground knowledge through agents, brokers and developers who do their research and homework are necessary for retail estate planning. A lot of the time, this type of data does not make it into databases in a structured way and as such, AI/ML models are missing that intelligence and will be less efficient, even biased.
On the consumer or customer side, our Retail and QSR executives say that the analysis of customer feedback and purchase behavior - enabled by loyalty programs, online ordering, and customer-centric business models (professional services) - in “real-time” can form diverse datasets to help predict trends and uncover untapped potentials in the market. However, some retail organizations still do not have customer behavior data afforded by loyalty programs or customer identification programs. Market research can fill these gaps somewhat, but historically, the data is being captured using time-boxed surveys akin to looking through the rearview mirror. The promise is that AI-generated surveys, continuous survey methodologies and Computer Vision technologies for tracking customer patterns within store locations are expected to bring more real-time customer behavior and preference data, helping uncover subtle shifts and identify untapped or emerging consumer needs to marketers and real estate analysts.
Capturing New Unique Data Sources
Finally, most of our executives believed that having “unique to them” sources of data for use in modeling and prediction was the holy grail!
For example, in certain retail sectors, health and wellness and tax and financial planning (retail professional services), the importance of understanding the depth of the client-professional relationship is critical in creating demand. In the tax and financial planning area, harnessing detailed customer and prospect information - beyond basic demographics - is a rich source of data. If gathered strategically through proprietary market research or consistently collected customer feedback, data science professionals can provide new custom attributes that are spatially distinguished to feed into market share prediction models and drive local marketing efforts.
In the Health and Wellness sector, AI becomes a strategic ally in positioning new emerging services to align with the evolving health needs of a changing population. As demographics change due to an aging boomer population and with increased immigration to western countries, predicting the dominant health needs will be of huge value to help leadership plan how they can prepare for attending to those needs. Finding data sets unique to this sub-sector would enhance the modeling and prediction capabilities of AI applications to inform the strategic discussions on what future health services to deliver in which markets!
The PiinPoint Way
At PiinPoint, we invested heavily in our GeoAI capabilities and have an ongoing appetite for finding new sources of data through data partnerships, or proprietary methods. We recognize that data is everything. Over the years, we have built our platform to leverage all kinds of unique and spatially attenuated data. We think of data in three categories:
- Foundational Data: The foundation data of our location Intelligence platform is:
- Census Information (Demographics, socio-economic),
- household expenditure,
- point-of-interest (all major retail brands), and
- municipal traffic counts.
- Third Party Data: PiinPoint invests in data partnerships that provide our clients with interesting sources of data that provide rich detail about customers and consumers in a trade area, for example:
- Mobile GPS Data to estimate pedestrian and vehicular traffic as well as visitors to a geofenced location;
- Geosocial segmentation that gets at location-based consumer behavior and brand affinity,
- Construction/Development data to capture location-based real estate supply changes.
- Client Specialized Data: Our platform enables users to bring in any spatial data they are interested in visualizing on a map or used in our AI/ML modeling applications. Things our clients look at, but not limited to:
- individual store sales and visitor data,
- customer purchase basket data,
- ecommerce sales data,
- competitive intelligence (localized pricing, assortment, customer traffic),
- location-based restaurant spending data,
- store-level customer satisfaction data, and much more!
Those that have not started the data discovery and integration journey, are typically using traditional GIS analyses with basic demographics and household spending to inform Real Estate teams. Those that have started and have a data strategy as one of their core pillars of the AI strategy, will win the day!
Check out PiinPoint and our full suite of data sources. We exist to support our clients' need to increase their speed to market and close deals faster!
