Comments on: We Predicted Right, Could comScore Be Wrong?

By: Leo Portnoy

Leo Portnoy — Sat, 02 Feb 2008 13:23:50 +0000

Porterhouse, you’re correct that representivity of a sample is a critical issue, but if you believe that people who are willing to be monitored are inherently unrepresentative, then you’re essentially saying that all market research is flawed. The fact is that there are millions of “representative” people who are willing to express their opinions by participating in market research, allowing their behavior, attitudes and opinions to be monitored. Decades of market research have proven that to be a fact. I’m also impressed by the quarterly data that comScore has published comparing their e-commerce numbers to those of the Commerce Dept. For six years, the two data sets haven’t varied from each other by more than a few percent. To me, that seems to be very powerful confirmation of the validity of their data

By: PorterHouse

PorterHouse — Sat, 02 Feb 2008 00:12:14 +0000

Leo,

While comScore certainly looks at a more diverse cross section of the marketplace I have trouble thinking that their data is not inherently flawed. comScore gets the majority of their data from panels and people who opt in for measurement of their internet usuage. It is most definetely not a random sample of the internet population. It is completely plausible that an individual who desires thier internet usuage to be monitored may not behave like the general interent population.

In fact our data may actually be a better representaiton of the actual behavior of the preHoliday internet user, especially since the results have repeated themselves year over year. The type of product was varied (perfume, clothing, cigars, electronics) and the data robust. I’m going to stake our claim to this date as the true peak of holiday online shopping.

By: Angie

Angie — Fri, 01 Feb 2008 15:57:38 +0000

You’re right Leo, what I meant and should have written was “the second Monday before Christmas” or “two Mondays before Christmas,” not “the Monday two weeks before Christmas.” And I also realized that our data is from a very small subset of retailers compared to comScore; however, it is interesting that this is the second year in a row we have found the second Monday before Christmas to be the highest sales day. Our client base represents a variety of retail categories and some of those categories differ from last year as well. I believe our data is noteworthy, although I know it does not discredit comScore’s.

By: Leo Portnoy

Leo Portnoy — Fri, 01 Feb 2008 12:26:42 +0000

The difference is almost certainly explained by the fact that you’re not looking at a representative cross section of the entire e-commerce market — as does comScore — but rather just a subset of certain retailers. I’d bet a lot of money that comScore has it correct. You might also want to recheck your time period definitions. The term “the Monday two weeks before Christmas” can easily be interpreted to be Monday December 10, which has two weeks between it and Christmas. Monday December 17, on the other hand has only one week between it and Christmas.