Current Census Demographics by DMA

Once in a blue moon, I get a custom data request for current population and demographic data for Neilsen’s designated market areas or DMAs.

Getting current Census demographics for DMAs is a bit tricky but definitely do-able. Unlike “normal” geographies (counties, zip codes/ZCTAs, etc.), the US Census Bureau doesn’t provide demographic data for DMAs. For example, you can get demographic data for New York city or for New York County from the US Census Bureau. But you can’t get demographic data for the New York DMA from the US Census Bureau.

To get current demographic estimates for DMAs using Census data, you need to:

  1. identify Census geographies like counties or zips/ZCTAs that make up DMAs, and
  2. sum up the demographics for those geographies.

That sounds reasonable, right? Let’s dig a little deeper.

First, you need to decide which Census geography type you’ll be using. Your options and the positives & negatives of those options are as follows:

  1. Counties
    a. Negative: Higher Potential for Geographic Errors
    b. Positive: You Can Be More Confident in the Demographics (sometimes)
  2. Zip code/ZCTAs
    a. Negative: Higher Potential for Demographic Errors
    b. Positive: You Can Be More Confident in the Geographic Boundaries (sometimes)
  3. A combination of counties and zip codes/ZCTAs

Option 1. DMAs And County Demographics

According to Neilsen, “a DMA region is a group of counties that form an exclusive geographic area in which the home market television stations hold a dominance of total hours viewed.”

Quick & Dirty Boundary Check

Let’s look at a simple example like San Diego DMA. When you compare the boundaries for the San Diego DMA on a 2012 Neilsen map with the boundaries of San Diego County on a US Census Bureau map, the boundaries of the county and the DMA look similar.


The above is a quick & dirty check to see if the county boundaries align with the DMA boundaries. In a perfect world, the next step would be to overlay a Neilsen DMA boundary shapefile with a US Census Bureau county shapefile to verify that the boundary of the San Diego DMA precisely matches the boundary of San Diego County. But alas…Nielsen has never called me back despite my voicemails requesting to purchase said shapefile from their DMA department.

2nd Source Boundary Check

So in our imperfect world, you can then check the county to DMA boundaries using SRDS maps or simzou/nielsen-dma data. An enterprising individual (simzou) has generously created shapefiles of the Nielsen DMAs. But I have no idea the source for either the SRDS maps or the nielsen-dma data, so I can’t be sure that the DMA boundaries in either of these sources are accurate. But personally, I feel more confident after checking a couple of different data sources. Note: if these checks are too imprecise for your project skip to Option 2. 1. An Easy, Accurate & Costly (?) Way to get Zip to DMA Lists below.


Pull The Demographics & Sum Them

If you’ve checked a couple of sources & you’re comfortable assuming that the boundaries of the DMA roughly match the boundaries of the county/counties, you can then pull the demographic data for the counties from the US Census Bureau. In the San Diego DMA example, you can pull the demographics for San Diego County from the US Census Bureau, and you’d have the DMA demographics. But most DMAs are a collection of counties – so you’d need to add the demographics for county A + the demographics for county B + the demographics for county C…to get the DMA demographics.

Since we’re talking about summing demographics for multiple geographies (and actually since we’re talking about the most current Census demographics, which as of today are in the 2013 American Community Survey), I have to write a scary phrase. Brace yourself.

Margins of error.

WAIT! Come back. Don’t tune me out just yet. The least you need to know is that county demographics typically have smaller margins of error than zip/ZCTA demographics. And smaller margins of error are typically better given the questions that my clients ask of this data. You can get more details about margins of error and Census American Community Survey data here.

Do the above steps sound painful or time-consuming? You’re not alone in this conclusion. Skip to the part where I do this work for you.

Option 2. DMAs And Zip Code/ZCTA Demographics

Or What Happens If The Quick & Dirty Boundary Check Fails

So using county demographics to estimate DMA demographics works in cases where the DMA boundaries align with the county boundaries. But in your boundary research, you might find DMAs that don’t align with county boundaries like Los Angeles DMA or Palm Springs DMA. For example, the Los Angeles DMA only includes part of Kern County. And the Palm Springs DMA only includes part of the Riverside DMA. Using counties boundaries doesn’t work for these DMAs. Instead, you can use zip code/ZCTA boundaries for DMAs.


2 Ways to Get Zip to DMA Lists

1. An Easy, Accurate & Costly (?) Way to get Zip to DMA Lists. You can buy a list of zips by DMAs from Neilsen. I don’t know how much this costs, but sometimes when I’m doing a custom data request, my clients provide me with a list of zips by DMAs. And I’ve always assumed that this data comes from Nielsen, but I’ve never thought to ask.

I’ve run into two problems with zip to DMA lists in the past.

  1. I’ve found data errors in these lists. For example, maybe we’re pulling population data for the Austin DMA in central Texas. But I’ll find a random zip code on the zip to DMA list that’s way out in north Texas. There’s no way that this zip code should be included in a central Texas zip to DMA list.
  2. The zip codes are accurate, but the US Census Bureau doesn’t have a ZCTA that corresponds to this zip code. Read more about zip codes and ZCTAs.

2. Alternative ZCTA to DMA Way. Another option is to produce your own ZCTA by DMA list by using your favorite GIS to calculate the intersections between this simzou/nielsen-dma data and the US Census Bureau’s ZCTA boundaries. This intersection would be straightforward, but I would only proceed with this option if you were comfortable with the accuracy of the DMA boundaries – for which I haven’t done the necessary due diligence.

Another potential problem with using zip codes/ZCTA demographics is that they can have large margin of errors (see previous margin of error reference). Sure, there’s documentation about Census data margin of errors. But 99% of the time, folks don’t want to talk about margins of error. They just want to answer a business question – like where do I spend my advertising dollars to get the largest ROI. And margins of error seem to complicate that discussion – which I completely understand. When I’m making a business decision for Cubit, I don’t want to see 500 +/- 25. I just want to see 500 even if 500 +/- 25 is accurate.

Pull The Demographics & Sum Them

Once you have your zip codes/ZCTAs to DMA lists, you can pull the ZCTA demographics for your DMAs from the US Census Bureau’s website and sum them up to produce DMA demographic estimates.

Do the above steps sound painful or time-consuming? You’re not alone in this conclusion. Skip to the part where I do this work for you.

Option 3. Hybrid Approach – Counties AND Zip Code/ZCTA

If I had to do this project for myself, I would use BOTH counties AND zip code/ZCTAs.

So in the case of a San Diego DMA where the DMA boundaries appears to align nicely with the county boundary, I would use county data. But in the case of a Los Angeles DMA where the DMA boundaries don’t align nicely with the county boundaries, I’d use zip/ZCTA data.

That said, as of today, I’ve never had a client choose to use a hybrid approach. Maybe I haven’t done a good job of explaining it. Or perhaps because it’d be more expensive to verify the right approach to take for each DMA. But more likely, folks look at the positives and negatives of counties versus zip codes/ZCTAs and pretty quickly self-identify the approach that works best for their project.

Do The Above Steps Sound Painful & Time-consuming? You’re Not Alone In Thinking This.

Sure, you could do the above steps yourself — or you could hire Cubit to do this work for you for likely significantly less than your hourly rate. Our other clients typically like DMA demographics in the following formats:

  • $39 Custom Starter Reports. These reports contain basic demographics (i.e. population, race, income, etc.). They are print-ready PDF documents with colorful graphs that you can easily include in your presentations and final reports. This option is a typically good fit if you need demographics for a handful of DMAs. Check out a sample Starter Report for the San Diego DMA below.
  • Spreadsheet Custom Data Requests starting at $199. These Excel ™ spreadsheets allow you to analyze the demographics for many DMAs. Plus, you can customize these spreadsheets to get only the demographics that you need or to get more detailed data than basic demographics. This option is typically a good fit if you’re doing an analysis of many DMAs. Typically, demographic data projects for DMAs are heavily customized to answer your specific question. So the best starting point for custom data work is to either fill out the Custom Data Request form or call me, Kristen, at 1.800.939.2130 to discuss your project.

No TweetBacks yet. (Be the first to Tweet this post)
Comments are closed.

Powered by WordPress. Designed by WooThemes