GenuineVC David Beisel's Perspective on Digital Change

November 10, 2005

I love numbers. When someone or a company has a rich set of data that’s meaningful, I like to really dive in. And so I am sometimes disappointed when the only figures that are presented are “averages.” Usually that translates into people presenting or writing about an arithmetic mean. But I often crave more, as an average figure tells you nothing about the distribution of the dataset.

Wikipedia explains further,

“[Mean] is used for many purposes and may be abused by using it to describe skewed distributions, with highly misleading results. A classic example is average income. The arithmetic mean may be used to imply that most people’s incomes are higher than is in fact the case. When presented with an “average” one may be led to believe that most people’s incomes are near this number. This “average” (arithmetic mean) income is higher than most people’s incomes, because high income outliers skew the result higher (in contrast, the median income “resists” such skew). However, this “average” says nothing about the number of people near the median income (nor does it say anything about the modal income that most people are near). Nevertheless, because one might carelessly relate “average” and “most people” one might incorrectly assume that most people’s incomes would be higher (nearer this inflated “average”) than they are.”

Examples like the one cited above are just the beginning. I often see statistics quoted in the popular media using “averages” which are incomplete at best, and misleading at worst. To me, averages are like snapshots as compared to a whole movie; they reveal a moment in time, but they don’t tell the whole story.

  • Barnaby James

    I also hate the proliferation of “junk statistics” to give the appearance of analysis. Tools like Excel make it easy to compute stats like the variance for the normal distribution so ergo everthing must be normally distributed because it’s easy to calculate!

  • Daniel Nerezov

    Ebay’s announcement:

    There’d be a lot of interesting data there.

  • Ryan Hudson

    I also agree. For my business, we face seasonality on two fronts: bookings, and delivery. Using just the means would yield useless information particularly when analyzing the timing of cash flows. To account for this we have carefully modeled the seasonality of bookings and delivery. The peak season for bookings is Dec-Feb while the peak season for delivery is Jun-Aug. An order placed in May is clearly more likely to be one month away than an order placed in January. Using an average of four months for all bookings would yield a dangerously inaccurate picture of seasonal demand spikes.

About Me

  • avatar
  • I am a cofounder and Partner at NextView Ventures, a dedicated seed-stage venture capital firm making investments in internet-enabled startups. Read More »



Rob Cho Go

NextView Twitter Stream

  • Rob Go
     - 13 hours ago
    @gabor Atlas is an awesome name :)
  • Rob Go
     - 8 hours ago
    RT @TomWheelerFCC: Historic day here at the @FCC. Finally adopted strong, sustainable, enforceable rules to protect #OpenInternet. Thanks f…
  • Rob Go
     - 14 hours ago
    @hunterwalk @homebrew uncapped note for the first company of your next unborn child
  • Rob Go
     - 14 hours ago
    @hunterwalk @schlaf any post with the word kabuki is a good one