A few things you can count on every year around this time:
- Sales data for Black Friday will be touted by biased interest groups. They are invariably have an upside bias;
- Headline writers will get it wrong
- Survey data will be taken as the equivalent of actual sales;
- Strong forecasts will be subsequently proven wrong;
Such is the current situation with the Black Friday sales data, with reports still trickling in from around the country.
The WSJ goes for a hat trick of errors, starting with this article’s headline:
What’s wrong with this? First, as opposed to actual sales data, they rely on a “survey of 3,370 shoppers, the National Retail Federation estimated shoppers spent an average of $372.57 over the weekend, up 7.2% over last year’s $347.55.” The National Retail Federation is hardly the objective group you want crafting (or hiring 3rd parties to create) survey questions; 2nd, we know that humans are terrible at forecasting their own behaviors. Historically, their projections have had little correlation with their actual spending patterns. And third, the headline is belied by the details contained in the article. (MarketWatch was no better)
The NYT did no better. They wrote (Deep Discounts Draw Shoppers, but Not Profits):
“The National Retail Federation, adding up sales Thursday through Saturday and projected sales for Sunday, said that each shopper spent about 7 percent more this year than last year. Shoppers spent an average of $372.57 Friday though Sunday, according to the federation, a trade group”
No, the NRF did not “add up sales.” They added up self-reported answers to survey questions. As we have seen over the past few years, there is a world of difference between the two.
One of the better data sources of what the retail season might look like comes from Comscore, which tracks actual online retail sales.
“ComScore on Sunday reported that online, nontravel retail sales on the Friday after Thanksgiving, traditionally a big day for consumer spending, reached $534 million. That’s up from the same day a year ago, but just barely–online retail sales rose just 1 percent, from $531 million…
But for the four weeks of November through Friday the 28th, retail e-commerce dropped to $10.4 billion, down 4 percent from $10.8 billion for the same period in 2007, according to ComScore. For the full holiday season, even Friday’s slight gain may look good. ComScore predicted that for November and December, online sales will be flat compared with 2007, coming in again at $29.2 billion.”
Bloomberg also got it wrong, but at least provided some context:
“U.S. shoppers, lured by discounts of as much as 70 percent, came out in greater numbers than expected and spent more this holiday weekend than last year, though those figures may not foreshadow a stronger holiday season.
The National Retail Federation, AN INTEREST GROUP FOR RETAILERS AND SHOPPING MALLS, said 172 million shoppers went to stores and Web sites, a 17 percent increase from a year ago and more than a forecast of 128 million. SHOPPERS SAID THAT they spent an average of $372.57, up 7.2 percent from last year, according to an e-mailed survey conducted for the NRF by BIGresearch, a Worthington, Ohio-based polling firm. SHOPPER SURVEYS HAVE A TRACK RECORD OF BEING UNRELIABLE FORECASTORS OF BEHAVIOR.
The early turnout may not be enough to overcome consumers’ reluctance to spend as they cope with climbing unemployment and housing costs and shrinking values in their stock holdings. The NRF said its forecast remains that holiday sales will rise 2.2 percent to $470.4 billion, the slowest growth in six years.” (My edits are ALL CAPS)
In an ideal media world, Bloomberg would strike nearly all of the middle paragraph. Alternatively, they could add information explaining that the NRF is a PR group for retailers, and that their surveys have not been historically accurate forecasts of actual spending.
Later this week, we should get Master Card’s SpendingPulse data. It is the actual measurement of what people spent (via their MC Credit Card) plus checks and cash. It has proven to be more reliable than either foot traffic, surveys and other biased sources.
To get a better view of this year’s holiday spending, back out the fuel purchases (whivhmay appear depressed due to less usage and lower prices). You also need to adjust for some of the gains in online sales, which are likely to be up 20% as part of a secular trend towards online shopping, not increased retail sales.
How Good Were Holiday Sales Really? (January 2008)
More Bad Data from the NRF? (November 2006)
Repeat After Me: Spending Surveys Are Meaningless (October 2007)
Holiday Shopping Off to Strong Start
Unprecedented ‘Door-Buster’ Deals Send Sales Up Significantly Over Last Year, but Crowds Thin on Saturday
RACHEL DODES, ANN ZIMMERMAN and JENNIFER SARANOW
WSJ, DECEMBER 1, 2008
Deep Discounts Draw Shoppers, but Not Profits
NYT, November 30, 2008
ComScore: Black Friday e-commerce hits $534 million
CNET, November 30, 2008
U.S. Holiday Sales Rise in First Weekend, Group Says
Bloomberg November 30, 2008
What's been said:Discussions found on the web: