A Deeper Look at Fiber Consumption
According to the Fiber-to-the-Home (FTTH) Council, FTTH networks are now available to more than 15 percent of homes in the United States. More than 9 million households across North America are now connected directly into high-speed, high-bandwidth fiber optic networks.
In this third and final look at fiber adoption in the United States, we investigated fiber availability and adoption by income levels. Going into this study, our hypothesis was that consumers living in higher income geographic areas would be more likely to adopt a fiber broadband connection as it is typically higher priced than a standard broadband connection such as DSL or Cable. While we certainly saw that in general this was true, we also observed some interesting interactions when considering population density.
The BroadBand Scout Database* is comprised of over a billion internet transactions that link consumers’ physical addresses to their specific broadband provider. The database is freshened on a frequent and periodic basis, allowing for the tracking of provider-level usage, technology adoption, and carrier market share shift.
Before we jump into the relationships between fiber and income levels, let’s first step back and quickly review what we saw in our last blog. In that blog, we demonstrated that the more population dense an area was, the more likely consumers were to adopt a fiber broadband connection. This can be seen below in Chart 1.
As population density increases, fiber broadband adoption increases. However, it is interesting to note that when we compare the areas with the second highest population density (i.e. Decile 9) with the areas with the highest population density, fiber adoption actually drops from 38 percent to 28 percent. This logically makes sense when you consider that the most urban areas in the United States have different socio-economic characteristics. Therefore, for the purposes of this blog, we looked at a very important socio-economic attribute–median household income.
To look at fiber broadband adoption by median income, we appended county level median income obtained from the latest Census estimates. ** We then ranked all of the counties in the United States by median household income and placed them into ten equal-sized groups (i.e., deciles). Decile 1 represents the 10 percent of counties that have the lowest median household income while Decile 10 represents the 10 percent of counties with the highest median household income.
As is shown in Chart 2, there is a strong and positive relationship (R2=.51) between median income and fiber adoption—that is, the adoption rate of fiber increases as the median income goes up. However, there are significant deviations from the overall trend. Do you see the very large dips in median income when compared to their “neighbors” for Deciles 2, 6, and 8?
When examined independently, we saw that both population density and median income are positively correlated with fiber adoption. But we also noticed some possible anomalies. As such, we dug a little deeper to see if there were any interactions between population density and median income. In other words, were there areas where the combination of population density and median income deciles does not follow the general trend? To answer this question, we created a heat map to help visualize fiber adoption rates when looking at income and population density in combination. Chart 3 below reports fiber adoption percentage for the combinations of population density decile and median income decile. The blue areas represent geographic areas with lower adoption rates and the red areas represent areas with higher adoption rates. The darker the blue, the lower the adoption rate; the darker the red the higher the adoption rate. [Note: blank or white cells with no numbers present represent cells where we do not have enough information to accurately measure adoption percentages].
Chart 3: Fiber Adoption Heat Map
Inspecting the heat map provides us with a different perspective. Overall, we can still see the same general trends, but now we begin to visualize the interactions. The olive-colored oval shows that for the middle income areas (Median Income Decile 5) but lower population density areas (Pop Density Deciles 2 and 3) there is a much higher level of fiber adoption as indicted by the dark red shading The adoption rate for both cells is approximately 50 percent. The patterns and relationships that pop out in the heat map made us explore even more deeply. We found that over 65 percent of our sample within these cells (i.e., inside the olive-oval) resided in California which has a very high adoption rate. Note: overall only about 10 percent of our total sample is from California.
Similarly, we looked at the purple and blue ovals. For the purple oval, we again found that nearly 100 percent of these came from the State of California. For the blue oval, we saw the same general trend but here we found that 60 percent of these came from the State of Maryland which also has a high adoption rate.
These outliers point out to us that when it comes to broadband adoption, other factors can have a significant influence on broadband adoption. This can include pricing, geography, and even cultural differences.
Summary and Conclusions
Through this series of blog posts we have seen that factors such as geographic areas of the country, population density and income can and do have an impact on broadband adoption. In addition, we understand that other factors such as pricing and cultural climate can impact adoption.
There are hundreds of municipal fiber broadband projects going on all across the country. To have a successful project, it is critical to understand all the components of the project area including the existing infrastructure and pricing, the opportunity size, the demographics and more.
* We took a random sample from the Broadband Scout Database for transactions occurring between March 2013 and June 2014.
** For Census Median Income data, we used the 2013 Census American Community Survey results at a county level.
*** For fiber availability, we used the Dec 2013 National Broadband Map data from the NTIA’s State Broadband Initiative Program.