Create Your Own Market Boundaries

I love maps. When I look at a map, I don’t think too much about the borders I’m looking at. It’s easy to just assume the smartest people with the best intentions thought it through completely. If you go down the deep rabbit hole of trying to understand the boundaries of our world, you realize most are just lines drawn on a map by people serving their own agendas. Take Africa, for example: most of the current African state borders were drawn by European colonial powers with little knowledge of the local geography or people. The point is, it’s easy and somewhat arbitrary to draw lines on a map, and most were drawn without thinking about the long-term implications.  

In spatial analytics, these arbitrary lines become the foundation of your work if you choose to use them. They influence every decision you make, from where to invest to how to target customers. Get the boundaries wrong, and your analysis suffers.

There are official boundaries, like those of states, counties, or cities backed by governments and law. Official boundaries are useful in many ways but have little use in spatial analytics other than to occasionally display data. Site selection is a world of subjective and unofficial boundaries, filled with trade areas, service areas, drive times, markets, and regions. My favorite unofficial boundary is the Midwest. What exactly is the Midwest? Seemingly everyone has an opinion. Here’s a fun activity at a party: stare at a US map and debate whether Pittsburgh, Louisville, and Rapid City are part of the Midwest or not. The Census Bureau has an official definition, but you can ignore it, their definition is wrong just like your friends at the party. Sorry, as a geographer, my idea of fun might be different than yours.

Maybe the most useful unofficial boundaries for business purposes are market boundaries. Suppose you want to compare Kansas City to Oklahoma City for your next marketing campaign or your next target market. The first question you have to ask yourself is what geography am I going to use? You could use official city or county boundaries, but that rarely makes sense. These are metro areas with adjacent suburbs, unincorporated areas, and neighboring towns. These invisible lines that cut through metro areas have almost no bearing on people’s daily lives. If humans don’t naturally interact and behave based on invisible lines, there’s not much logic in performing analytics based on those lines. What truly defines a market is the people in it and their economic activity.

So what boundary should you use to capture this essence of a market? Since the definition is subjective, there are no official boundaries, but there must be some really good unofficial market boundaries…right? Let’s take a look at the off-the-shelf options. This isn’t an exhaustive list, but these are the most common and readily available:

• DMAs (Designated Market Areas) – 210 media markets created by Nielsen. The markets are conterminous, so the entire US is covered by at least one of them. Some of them are enormous and stretch hundreds of miles past the core metro area. Also, since there’s only 210, many of the mid to smaller sized markets get bound to larger metro areas or bound to other mid to small markets. DMA borders are based on county borders.

  
  

• CBSAs (Core Based Statistical Areas) – 935 markets created by the US Census. Generally, CBSAs do an OK job of capturing most large markets but not great with mid and small markets. Their borders are also based on county borders. Non-conterminous.

  
  

• UAs (Urban Areas) – 2,644 markets created by the US Census. Unfortunately, they are only updated every decennial census, so these quickly become outdated. The boundaries include only populated areas, but almost too well, not leaving much of a buffer around markets. Small markets are their own market. Really inconsistent in how they treat some suburban areas. Some are their own market, while other suburban areas are included in the main metro. Non-conterminous.

As you may have noticed, the top 2 are based on county borders. That would be fine if county boundaries were consistently and rationally drawn. Unfortunately, most county boundaries were based on the whims of state politicians when the state was created. Not to get political, but we don’t know their motivations, and I doubt these politicians were thinking of the analytical ramifications hundreds of years into the future. Each of these can serve a purpose, but they each have fundamental flaws.

Which leads to the best solution…create your own! In a world of people drawing lines on a map to serve their own needs, it’s time for you to give it a shot. There is no right or wrong answer and don’t let perfection be the enemy of good. This is more about improving upon the less-than-ideal options that already exist. You could literally take an hour to freehand draw a polygon around each of the top 30 markets in the US, and it would be better than the off-the-shelf market boundaries.

Better yet, create an algorithm to do it for you for ultimate consistency. All you need is a grid file, a base layer, and your favorite LLM to create a script for you. The base layer could be street data, buildings, land use, population, business points or any number of things that signal humans living in proximity to each other. Here’s an example:

Just like most things in life, there’s an easy way and there’s a correct way. Take the time to think through all your geographies, when and how they were created, and if they really suit your needs. Know that creating your own is easier than you think and likely the best solution. Contact us to learn more about how Retail Gravity can help you.