Check out other #Datacember posts
Want to ensure you don’t miss any of the #Datacember posts? Preregister now to receive the #Datacember e-book, a compilation of all of our data-driven posts from this month. Plus some bonus features!
I’m guilty of getting wrapped up in the holiday shopping horse race: Is Black Friday spending up or down? What about Cyber Monday? What percentage of shoppers are buying on mobile? How is consumer behavior shifting? Who wins? Who loses?
It’s a holiday tradition that’s nearly as time-honored at turkey and stuffing. After all, the numbers charting Thanksgiving weekend shopping come at us from all angles and a myriad of methodologies. But it turns out that when the numbers are not handled with care, they can lead to some false conclusions. In fact, the FiveThirtyEight blog, which digs into data, statistics and how they are used and misused, points out that many of the Thanksgiving weekend shopping stories, and the prognostications contained in them, turn out to be wrong.
The blog post talks about the problems with in-store traffic reporting based on anecdotal evidence, a genre this CNBC report strays into. And it says that even stories backed by data can go wrong when they attempt to project what will happen based on what has happened. Blog author Ben Casselman notes that last year about this time the National Retail Federation reported that Thanksgiving weekend sales were off 11 percent. Stories of coming holiday gloom followed. But in the end, the U.S. Commerce Department found that retail sales for November and December 2014, actually rose 4 percent year-over-year.
In some ways, the difficulty with characterising the success of the annual holiday season is a reflection on the difficulty of success metrics in general. The key is to know what you want to measure and to understand the limitations of your methodology.
The NRF, for instance, with Prosper Insights & Analytics, surveys nearly 4,300 consumers during the Thanksgiving weekend to come up with its numbers. Given the need for speed, the method makes sense. Journalists and even the NRF’s retail members want a quick answer to the question, “How is the season going?” The organization can’t take the time to count actual retail receipts or to even go through the deep two-month analysis that the commerce department conducts.
The NRF found that more people said they were shopping online (103 million) than said they were shopping in stores (102 million), figures that would be relatively useless to key constituents had they come out weeks from now.
But using such Thanksgiving weekend figures to forecast the entire shopping season has become incredibly perilous. David Morin, a retail analytics expert, says the traditional heavy holiday shopping activity around Thanksgiving weekend is definitely spreading out beyond the stretch of Thanksgiving to Cyber Monday.
“The holiday shopping period is changing so much, I would definitely say there is probably a little bit of danger in using historical shopping metrics to project the future, because of the way we’ve seen both consumers’ behavior change, as well as the way that retailers are trying to engage their shoppers for longer periods of time in different, maybe more creative, ways,” says Morin, who works for Prism, a firm that helps retailers collect and use in-store analytics.
Retailers are starting Black-Friday-like promotions earlier, both online and in-store, and will likely keep them in place longer.
How and when consumers shop is going through an accelerating change that is reducing the significance of retail holidays like Black Friday and Cyber Monday. Consumers now shop when they want to shop, using the device — or more likely a combination of devices — that they want to use, often before during and after visiting a physical store.
“Historically, the Thanksgiving weekend has been the dominant shopping period for the the holiday season,” Morin says. “I think especially, as recently as this year, we’ve seen that window open a whole lot wider, especially on the e-commerce side. But I think in-store, that promotional period is also extending. We’re seeing, especially with our clients, and just in the general retail space, a lot more focus on in-store experiences — and that doesn’t have to be on Black Friday — to drive sales and traffic.”
The NRF’s numbers are a prime example of where those who analyze weekend shopping numbers can go wrong. While one could point to the figures as a sign that online shopping is eclipsing in-store sales, the contest isn’t even close. First consider that less than 10 percent of retail spending typically happens online. Second, consider The Financial News USA report that says that Adobe was estimating that consumers spent $4.45 billion online on Thursday and Friday, while ShopperTrack estimated consumers spent $12.1 billion in stores on those days.
Finally, the NRF reported that their survey indicated that a total of 151 million people shopped online or in-stores over the weekend, a clear indication that the separate online and in-store figures (103 million + 102 million) include people who fell into both categories.
All this just to underscore how tricky numbers can be and how important it is to understand both what you want to measure and the legitimate ways available to do that. That said, some figures of note from Monday:
Adobe reported that online sales were up 12 percent year-over-year for Cyber Monday, with 28 percent of sales happening on mobile devices. For Black Friday, Adobe’s numbers showed 15-percent revenue growth and 35 percent of sales coming from mobile devices. That was generally in line with a Monday report from Custora, a predictive marketing company in New York, that said online revenue was up 15.8 percent on Cyber Monday and 16.1 percent on Black Friday, with 36.1 percent of that revenue coming from mobile devices. The Black Friday mobile figure was up from 30.3 a year ago. The firm did not cite a figure for the percentage of Cyber Monday sales from mobile devices.
Mike Cassidy is BloomReach’s storyteller. Contact him at email@example.com; follow him on Twitter at @mikecassidy.