Smart Clustering: The Future of Defining Market Boundaries

Before determining the feasibility of an opportunity or value of an asset, real estate analysts and investors must complete a competitive market analysis.

Many real estate professionals rely on their market knowledge to conduct a competitive analysis. However, for an analyst who does not have decades of experience under their belt or someone venturing into a new market, the market analysis process can be difficult and time consuming. Furthermore, if done incorrectly, it can lead to poor investment decisions and results.

How Real Estate Professionals Define Markets Today

To establish accurate market boundaries and comparable properties one must either (1) have detailed and up-to-date knowledge of the market, or (2) make speculative assumptions based on previous experience.

Four common factors used to define the boundaries of a market are:

1. Neighborhoods — Neighborhoods may be formally delineated by streets or geographical features, but oftentimes are informal, and understood best by local residents and real estate professionals.
2. Local Government Defined Boundaries — Aldermanic wards or congressional districts are often referenced to define boundaries.
3. Zip Codes — Zip Codes are a well-defined geographical boundary within which real estate data can be aggregated.
4. Radial Distance — Drawing a radial distance around a particular asset (e.g. 1, 3, and 5 mile radii) is another technique to define a market. Demographic and economic data sets are often queried using this technique.

For analysts and investors without detailed knowledge of an area, it is common to define a market using geometric areas like a circular radius. On a map, this may look like a good representation of a market, but in reality, different sides of a major geographic boundary like a river or highway within that boundary may have substantially different demographic and economic characteristics.

ZIP Codes and government defined boundaries suffer the same drawbacks. Government boundaries are shaped in myriad ways to suit various administrative needs, but aren’t crafted for the purpose of classifying a homogeneous populations of residents or housing types.

An analyst or investor must typically be familiar with the most granular of these geographic boundaries: Neighborhoods. As outlined above, the extent of these geographies can be ambiguous — on occasion the official government boundaries for neighborhoods do not accurately reflect a true demographic or economic community.

Census Tracts — The Market Boundaries That Matter

All things considered, the process of defining a market leaves a lot to speculation. When Enodo looked for a better way to identify comparable properties, our team saw two areas that could greatly improve the analyst’s workflow: sourcing more precise market information and leveraging machine learning to programmatically delineate markets.

By aggregating and analyzing data from properties and markets across the country, Enodo found that utilizing Census Tracts allows for the most precise measurement of a market. Census Tracts are designed to be relatively homogeneous units with respect to population characteristics, economic status, and living conditions.

Each tract typically represents approximately 4,000 inhabitants, making their geographical footprint proportional to the population density in a market. This means that dense urban areas will have more Census Tracts and therefore greater granularity of data from one block to the next, while rural areas will have fewer Census Tracts and more homogeneous characteristics over large geographic areas.

Because Census Tracts revolve around homogeneous populations, the risk of overgeneralization of market characteristics is reduced by using these shapes. However, individual Census Tracts are often too small to contain a large enough sample of property transactions that can be deemed comparable.

Smart Clustering — The New Approach to Defining Your Market

For this reason, Enodo aggregates and analyzes economic, demographic and real estate market data among clusters of Census Tracts. Using a subject property as the basis for a market analysis, Enodo computes the statistical similarity of adjacent Census Tracts based on the characteristics found within the tract to converge on the optimal “market” for the subject comparison. Machine learning models built into the Enodo platform enable analysts to calculate accurate market rents, identify the incremental amenity impact for a subject property, and surface statistically similar comparable properties.

By sourcing more precise market information and leveraging programmatically delineated markets, real estate professionals can quickly assess any market and extract new insights previously left to speculation.