Question everything

In one of our discussions, Stretch provided an excellent graph illustrating that the ISP competition seems to reduce prices almost linearly and asked me in a later comment to justify the inverse relation between subscription charges and consumer choice.

You might consider this debate to be purely between Stretch and myself, but it’s an interesting example of what you might need to do in daily your job. If you want to be a great networking engineer, you have to be prepared to question everything, including common wisdoms, “well-known truths”, “common practices” and facts that look too good to be true. Ready? Let’s go …

Start with the source. The graph came from ars technica article US broadband report: more popular, more expensive, which used the data from Pew Internet’s Home Broadband Adoption 2009 report. I had a “somewhat” biased opinion (I’m playing the devil’s advocate here), but this clearly falls into “looks too good to be true” category. Anyhow, if you want to understand what’s being reported, you have to go as far back to the original data as possible; in our case, read the Pew Internet’s report, not the summary provided by ars technica (although they did a good job).

The obvious conclusion. Looking at the graph, the message is simple: whenever there’s competition, the prices go down. Are you happy with this explanation? Do you question its validity? Do you understand what the underlying unknown variables might be?

Find the alternatives. Whenever something looks too good, consider the alternatives. How about this: it’s easy to have low prices in urban environment with very high population density (Manhattan, London, Paris, Vienna or Singapore). It’s also obvious that there’s a lot of competition there (everyone tries to cherry-pick these customers). The rural places have low population density, therefore higher prices (assuming the prices somewhat reflect the actual cost of providing services) and almost no competition. The low population density would thus automatically correlate higher prices (assuming they are related to higher costs) with low competition. Are there any other alternative explanations? Give them in the comments.

Evaluate the alternatives. Do you find the second scenario viable? I’m not claiming it’s correct, I’m trying to nudge you to think. Can you find enough information in the graph, the article or the report to decide which explanation is valid? Document you findings in the comments.

Why does this matter? Why should you as a network engineer invest your time in this process? You’re constantly bombarded with “facts” nicely laid out in media articles, design guides, performance charts or whitepapers. Do you accept the “fact” that a router should belong to at most three OSPF areas and that an area should have at most 50 routers … or are you trying to understand what the actual limitations are? Are you happy to be mediocre and use simple recipes (without knowing where they come from) or will you try to become innovative by understanding the in-depth details and using them to your advantage? Are you willing to digest whatever media throws at you or do you want to form your own opinion? The choice is yours.


  1. Ivan,

    You are correct: there could be hidden variables that're affecting the outcome, such as population density. The pew paper brings up another - they did not control for the service type, so maybe the expensive service is 3G cell data.

    Unfortunately, we don't have enough data to settle this.

    However, the idea that competition reduces prices is inline with current economic theory.

    Here's my observation:

    I live in an urban part of a major metro area.
    I get both TV and Internet from Comcast.

    The TV price goes up every year, with small to no improvement in service. The ISP price hardly changes - I don't know if it even matches inflation - and they've doubled or tripled the speed.

    I have exactly 1 choice for cable TV, and 2 choices for broadband Internet.

    At the same time, if the ISP speed tripled, the PC speed and storage size improved by a factor of 8 or 10.
  2. I am involved in the purchase of hundreds of thousands of dollars per month of network services, mostly high speed circuits from local and regional providers, from DS3 to 10Gig. (Commercial, not consumer).

    In our service area, we see:

    High population density, high cost
    High population density, low cost
    Low population density, high cost
    Low population density, low cost

    From my experience the difference between high and low cost is that we have low cost where there is both competition AND excess capacity. I don't see an automatic correlation between population density and price, nor do I see an automatic correlation between competition and price.

    To give you an idea - I have a quote for 10Mpbs ethernet from a provider in a wealthy, high density community that has no competition, and I have a quote for for 10Gbps in a low population density community a few miles away where there is both competition and excess capacity. The quotes 10Mbps is within 10% of the quote for 10Gbps.

    Unfortunately my data is anecdotal & specific to a particular region and therefor of limited value.
Add comment