My friend Dale wrote this terrific essay about the way news stories are often made, using the example of the inevitable "Black Friday" holiday shopping story that every newspaper and radio and TV news program runs after Thanksgiving.

How News is Made, by Dale Dougherty

There should be a book titled "How News Is Made," a book that could be for journalism what "The Jungle" was to the meatpacking industry. My version would offer no conspiracy theory, but I'd point out the preponderance of sloppiness and lazy thinking coupled with a herd mentality, most especially in business journalism. I found a great example to illustrate what I've been thinking about, tipped off by an article written by Carl Bialik in the Wall Street Journal.

First, most of what we call "news" today starts out as a press release, which then becomes a headline, a sound-bite, and eventually a story. In a parallel to the way government operates, in which special interest groups lobby to create or defeat legislation, most of our news stories come as a result of PR efforts paid for by special interest groups (businesses) who have a stake in what becomes "news." (I'd love to come up with a taxonomy of stories by type just to show how few types there really are but that's a different point.)

Second, reporters like to ask good questions for which there may not be good answers. However, they'll force an answer because you can't say "nobody knows."

The third is that everybody loves numbers, regardless of where they come from, and these are the best kind of answers, regardless of whether the numbers are true.

Bialik's article mentions a press release from the National Retail Federation, which concerns the holiday shopping over the Thanksgiving Weekend.

(The NRF actually called it a "News Release" and it was released on Sunday, November 27th, even before the Thanksgiving weekend ended.)

It begins:

Blockbuster Black Friday Weekend
Sees Sales Near $28 Billion
145 Million Shoppers Hit Stores and Internet, Up From 133 Million in '04

Washington, DC, November 27, 2005--The ceremonial kickoff to the holiday season began with a great deal of fanfare as 145 million shoppers flooded stores and the Internet hunting for popular electronics, clothing, and books. An NRF survey conducted by BIGresearch found that the average shopper spent $302.81 this weekend, bringing total weekend spending to $27.8 billion, an incredible 21.9 percent increase over last year's $22.8 billion.

"As expected, retailers offered substantial discounts and savings on Black Friday to bring people into their stores and consumers held up their end of the bargain by shopping," said NRF President and CEO Tracy Mullin. "Even though many retailers saw strong sales this past weekend, companies will not be basking in their success. Stores are already warming up for the next four weeks because the holiday season is far from over."

More than 60 million shoppers headed to the stores on Black Friday, an increase of 7.9 percent over last year. Another 52.8 million shopped on Saturday, a rise of 13.3 percent over 2004. The number of shoppers out today is expected to be close to last year, with about 22 million people shopping.

Note that this story is built with two pieces of information: it has numbers and it offers an explanation for those numbers. It's really the perfect story, regardless of whether it's true or not. More importantly, the information attempts to provide an answer to a reasonable question: "How busy was Thanksgiving weekend for retailers?" and one can *not* leave the question unanswered. The possible answers are "up, down or flat." The answer "We don't really know" is not acceptable; it's not news. So the press release provides an answer that Thanksgiving Weekend sales were up significantly and an answer that the NRF people like. That answer is also believable because a national industry trade group had real data to back up its claim.

One might also want to point out that there is no real opposing trade group here to offer a counter-claim. Those who don't shop over the weekend aren't represented by anyone with a commercial interest in this question. Also, I should mention that the "big" story the day before Thanksgiving is how many people are traveling for the holidays, and how crowded the roads and airports are. That story doesn't seem to have any impact on the post-Thanksgiving story, which says everyone was shopping.

The NRF "news release" thus becomes news, variously massaged and distributed by news services. You might think of it as a kind of journalistic mash-up.

Read the beginning of this story from Bloomberg news service:

Weekend Sales Jump 22% to $27.8 Billion, NRF Says (Update1) Nov. 27 (Bloomberg) -- U.S. retail sales jumped 22 percent to $27.8 billion during the post-Thanksgiving weekend, as shoppers flocked to stores to buy electronics, clothing and books, the National Retail Federation said. Shoppers spent an average $302.81, the Washington-based trade group said today in a statement. It said 145 million shoppers went to stores and the Internet, as retailers offered substantial discounts.

Fortunately, someone looked deeper into the numbers behind the news story. In Wall Street Journal's The Numbers Guy column, Carl Bialik does a nice job of tracking down where this data comes in his article "Holiday Sales Numbers Don't Add Up."

He begins by noting: "The day-to-day behavior of American shoppers has become harder for investors and economists to measure" suggesting that the right answer to the question is "we don't really know, at least not now."

Then he looks into the research provided by the National Retail Federation, which was done by BIGResearch, LLC. He learned that the research company had 4,209 respondents to an online poll and they answered the question: "How much did you spend on holiday shopping?" On that basis of the survey, they extrapolated to the rest of the US population and come up with a 22% increase. Bialik also mentions other market research groups that are trying to measure sales, and examines their methods. ShopperTrak uses video surveillance to track shopping activity and it reported sales were about flat.

Like the Nielsen Ratings for TV, we have numbers that are treated as factual but they based on a fairly small sample of people and questionable methodology. The 4,200 respondents who said they'd spend roughly $300 on holiday shopping are themselves only responsible for $1,2M in sales but that number can be multiplied to produce a much more impressive number of $27B in sales. 4,200 people suddenly speak for 250,000,000 or so. The bigger the number, the more we seem to know!

It's like the person who tells you "everybody knows this or that", and you press the person how they know this and it comes down to a friend or relative who made an offhand comment. Hardly a significant sampling. Besides this sampling methodology just doesn't seem logical. How many truly representative people do you know? In other words, do you know anybody you'd say could speak for 60,000 other people in all matters of any importance?

Would the more truthful headline be considered news? 4,200 people say they spent on average $300 shopping over Thanksgiving Weekend.

Next I went searching for other stories that used the "News Release" as the basis for a story.

I found this item, as reported on NewsHound, which quotes Brit Hume on Fox saying:

"US retailers recorded sales of $27.8 billion over the Thanksgiving weekend & the National Retail Federation says the industry is on track for its second biggest season since 1999. Weekend sales were up 22% over a year ago and discount chains and online retailers reported record sales. So how does The New York Times report the good news? On the front page of its business section, The Times notes that shoppers avoided mall-based specialty stores in favor of the discount retailers, reporting that Friday's mall sales dropped nearly a full percentage point from last year."

Hume treats the NRF release as fact and its source is not really clear from his sentence. He takes the NY Times to task for not reporting it as "good news." He doesn't address the question

In fact the Times story, titled "Mall Stores See Trouble in Sales Data", referenced the NRF report but it certainly did more to explain how the conclusion was arrived at:

"In a survey of more than 4,000 consumers over the weekend, the federation found that 61 percent made purchases at discount retailers, 47 percent at department stores and 41 percent at specialty stores. Over all, it estimated that the weekend's spending would rise 22 percent, to $27.8 billion."

The Times' article cites more sources and does it more even-handedly. Hume apparently thinks reporters have an obligation to trumpet holiday sales, as though it were a belief made true by repetition.

In another example, the Knight Ridder newspaper in Lincoln, Nebraska, The Journal Star, ran a story that opens:

Black Friday sales best in history, retailers say; others not so sure By SUZETTE PARMLEY/Knight Ridder Newspapers Deep discounts, especially on consumer electronic goods, and longer hours caused a 22 percent increase in sales over last year during the Black Friday weekend -- making it the best weekend ever for retailers, according to the National Retail Federation.

This story echoes not only the data but also the explanation from the NRF press release. And yes, it's careful to cite the source. It also hedges its bet, answering the original question with: Way Up; Maybe Not.

We get the same kind of story from the SF Chronicle:

Holiday sales look promising Weekend yields mixed results, but retailers are hopeful
Carolyn Said, Chronicle Staff Writer
Tuesday, November 29, 2005

Steep discounts drew hordes of bargain hunters to stores this weekend but early results were mixed for the traditional kick-off to the holiday shopping season. The National Retail Federation called it Blockbuster Black Friday Weekend, saying that weekend spending hit $27.8 billion, up 21.9 percent from last year. A survey of 4,209 shoppers conducted by BIG-research for the trade group found that the average shopper spent $302.81 during the weekend.

And more of the same from the Washington Post:

For Retailers, It Looks Like Holiday Season Is Off to Strong Start

By Ylan Q. Mui
Washington Post Staff Writer
Monday, November 28, 2005; Page A10

Shoppers did not hold back this holiday weekend, spending a whopping $27.8 billion and setting the tone for what retailers hope will be a very merry Christmas, according to the National Retail Federation. The industry's largest trade group said yesterday that sales from the day after Thanksgiving -- nicknamed Black Friday -- through Sunday jumped nearly 22 percent over last year. Holiday sales traditionally account for about one-fifth of retailers' annual business.

It really is as though reporters attended the same lecture and took down the same notes.

Journalists were good students. I was told the answer so I just have to repeat it using my own words because if we all gave exactly the same answer in the same words, that would be copying.

Google shows hundreds of stories citing the NRF release. Not all of the stories conclude that sales for the kickoff to the holiday shopping season were up from the previous year but it was the basis for those that did report that sales were up. If that's the story you wanted to write, the NRF could help you write it.

It's also interesting how the lead paragraphs are so similar at conveying without qualification that sales were up, which is exactly what the NRF wants to read and hear. The assumption of the retail industry is that people will spend more if they think others are doing the same thing. This is the "good news." Similarly, "bad news" such as saying that sales are down can be used to alert people that they aren't doing their best to contribute to holidays sales.

It's as though the holiday season is a fundraising telethon, and we've got to get everyone to do their part. NRF is not alone in spreading holiday disinformation. Companies like to send out self-serving research directly themselves. The folks at BoingBoing shared this recent email sent to them, from an analyst at who wanted them to write about

"I am a marketing research analyst with and after looking at your blog, I thought some of our popular holiday shopping data might be of interest to you and your readers. It's clear from our site statistics that online shopping will have its best year ever as we are seeing impressive growth so far. On Black Friday we saw a 77% increase in shoppers compared to last year, and Cyber Monday had an 88% increase in the number of referrals we sent to merchants. Also, based on an earlier holiday survey we conducted we found that more than 50% of online shoppers surveyed said they would start their holiday shopping before Thanksgiving, and interestingly, despite that headstart, 75% of all respondents said that no matter how they plan in advance, they still expect to be shopping within 5 days of the holiday."

I am no math whiz but what's clear is there no real data, just pretend holiday data, however popular they think it might be. A percentage increase from last year, which could have been really, really low. (And who agreed that we now call the day after Thanksgiving, Black Friday; I've only cared whether or not it was a work day. And what about the Monday after Thanksgiving weekend Cyber Monday; shouldn't it be called Black Monday because it's hard to get back to work if you were lucky enough not to work retail and have a four-day weekend.) Is it really news that some people shop before Thanksgiving and won't finish until just before Christmas?

Then by the end of Tuesday, November 29, we get a quite different kind of report about Thanksgiving sales, from Reuters:

"Expectations that U.S. retail sales activity following the Thanksgiving holiday may not be as strong as some analysts had predicted knocked Google down $19.94 to $403.54 in brisk turnover of 21.4 million."

This Reuters report appears in EWeek, the New York Times and CNN:

Are they reading from a different hymnal? I don't know but this is a typical Wall Street story where the challenge of a story is to fabricate a reason for a change in the market. You'd think from the story that a bunch of people did, in fact, know that holiday sales were down, despite what the NRF said. But maybe they didn't really know anything either. There just has to be a reason Google's stock price fell and the "holiday sales were down" reason was there for the taking.

Given another day, the story changes and we have a different explanation (again from Reuters):

Shares in Web-search leader Google Inc. (GOOG.O: Quote, Profile, Research) slid 4.7 percent on Tuesday, their biggest decline in a year, as analysts expressed concern that its stock was overpriced after soaring 40 percent in two months.

Merrill Lynch analyst Lauren Rich Fine said in a research note on Monday that Google needs to develop new sources of revenue beyond its booming search advertising business to justify the current valuation on the stock.

Different analysts, different day, and you need a new story. Back to Thanksgiving sales, which were either way up or dramatically down. (Interestingly, some parts of the story from the previous day, including other quotes, were carried forward but not the part that connected Google to a slump in Holiday sales.)

On the 30th, we have a story from by Nat Worden, saying that we might know on Thursday the answer to the big question. Nonetheless, he sees "all signs pointing to...a great turnout" while also saying we won't really know until later.

Retail Reports Could Hold Holiday Clues

All signs point to a healthy November when chain stores report to Wall Street on Thursday. The drama for investors is what they say about the holiday season.

The current batch of sales data will include results from Black Friday and the following Saturday, when shoppers flooded into discount stores and electronics chains in search of holiday markdowns. Early returns on the post-Thanksgiving rush indicate a great turnout, although aggressive pricing and promotions have stoked concern about whether the spending will continue into December.

He also cites the NRF research later on in this article and he does note their methodology. He even quotes another analyst who says that the NRF numbers look high. Nonetheless, he seems very supportive of their numbers. But the point of his article should be to disregard early reports and wait for better information, if you even care.

Don't you just wish that journalists wouldn't make up answers to questions they can't answer? Don't you wish they could sift through self-serving research and not repeat numbers that are at the least, meaningless, if not wholly bogus?

The funny thing is that this same news is made every year in the same way as reliably as the turkey at Thanksgiving. The Internet allows us to see how news is made, as though we were walking through a factory tour, and we can compare the very similar results of a mass production system. Turns out the news can be as fake as a department store Santa.