MoneyWatch Tech

Black Friday, Cyber Monday, and Other Imaginary Sales Figures

By Erik Sherman

Updated on: November 29, 2010 / 4:26 PM EST / MoneyWatch

Hey, hey! Black Friday sales were really up! Oh, wait, make that only slightly. Unless you count inflation, in which case they are probably effectively lower. But Cyber Monday will be great -- except now it takes place on Thanksgiving.

The flurry of retail numbers that hit the business press at this time of year is prodigious. Unfortunately, much of the data is little more than hooey that either ignores major economic factors or that appear thanks to questionable sampling. Any business executive trying to make end-of-year adjustments has to learn to pick though the thorns to find the occasional useful bloom -- and to remember all the questions that the instant studies don't answer.

The rain of predictions is historically faulty, as my colleague Constantine von Hoffman recently pointed out on his blog. Look at last year's results of the Thanksgiving weekend purchase frenzy and the year-over-year increase was only 0.5 percent.

And yet last year, the National Retail Federation, in its annual survey, said that sales were up 9 percent over the previous year. Oops. Maybe someone misplaced the calculator.

It makes you wonder about this year's survey. The NRF says that shoppers spent $45 billion while last year the total was $41.2 billion. That should be 9.2 percent growth. Yeah, right. According to ShopperTrak, which physically samples mall sales, the increase over 2009 this year was 0.3 percent.

So what's going on? The NRF surveys consumers, not stores. Even though it claims a 1.5 percent margin of error, that only means the results would be close 95 times out of 100 that you administered the survey to similar samplings. But go to the site of BIGResearch, the firm that actually carries out the survey, and you find out that consumers opt-in to surveys by email. In other words, the company likely has a self-selecting sample that may not actually reflect the general shopping populace. Such an approach could easily explain the huge discrepancies. Furthermore, the consumers are reporting what they spent, and although they should know, they may make significant errors. (Try an experiment: keep all your receipts over a weekend, estimate what you spent on Monday, and then total the receipts and see the real number. You may be surprised.)

Could the difference owe to online shopping? You could ask comScore, which claims that Black Friday 2010 online sales were up 9 percent over 2009. I emailed the company and learned the following:

It gets consumers to agree to let the company monitor their behavior and the actual price they pay for goods online.
All the data comes from online and the people responding may not be representative of the populace as a whole.
The company recruits users "using a broad range of incentives," which could further affect the results.

And, similarly to the NRF survey and unlike ShopperTrak, the numbers don't come from the sellers themselves, causing even more potential error. Furthermore, there are major questions that the information doesn't even address:

How much did retailers discount products to get customers in the door?
How much did the retailers have to spend in overtime and additional operating costs to open stores early on Friday -- or late on Thursday?
Did shoppers spend broadly, or did they concentrate so heavily on the marked-down items that profit was decimated?
Did sales see growth? Or, after inflation, was it actually a net loss?

In short, there is little in most of these numbers that lend themselves to intelligent decision making by executives.

Related:

Image: RGBStock.com user lusi, site standard license.