Did Twitter predict Black Friday and Cyber Monday sales results?

There’s been lots of talk lately about whether tweets can predict social and market trends. Recent data have demonstrated links between what’s posted on Twitter and the stock market, flu rates, even election results. So, what can Twitter tell us about shopping and holiday spending?

Early reports about this year’s Black Friday and Cyber Monday are looking good. Sales are up from last year – ShopperTrak says brick-and-mortar retail sales on Black Friday were $10.69 billion, comScore data indicate Black Friday online sales of $648 million and Cyber Monday sales of $1.03 billion. On the other hand, there have been recent discussions about the hype of Cyber Monday and how it’s just that – hype. These discussions focus on the fact that Cyber Monday is not the biggest online shopping day of the year at all, and its status as a shopping “holiday” is overstated.

So, we thought Twitter could help us understand more about the state of holiday shopping in 2010. In the days leading up to Black Friday and Cyber Monday, we used our new TweetReach Tracker to monitor and measure tweets about both days. We kept the Trackers pretty simple. For the Black Friday Tracker (BF), we gathered all public tweets that included the terms “Black Friday” or #blackfriday, and for the Cyber Monday Tracker (CM), we monitored the terms “Cyber Monday” and #cybermonday.

So, just how active were these Trackers? Well, they were the most active Trackers we’ve ever run (and that includes tweets about all kinds of celebrities, elections, news events, even TSA). In just four days, more people mentioned Black Friday than during the entire Brazilian presidential election in October. And how many people were talking about and how many saw tweets about these two shopping holidays? What does that tell us about Black Friday and Cyber Monday overall? Read on for answers. (FYI, in the following graphics, Black Friday is represented in blue, Cyber Monday in green.)

Who tweeted?

First, let’s look at how many people tweeted and who those people were.

Nearly 300,000 more people tweeted about BF than CM. As you’ll see throughout, Black Friday was more active in every way than Cyber Monday. It’s particularly pronounced here though, as 7 times more people tweeted about BF than CM. If we just used contributors as a predictor for BF and CM sales, then we’d expect to see about 7 times more sales on BF than CM. But, just glancing through the tweets themselves, we see as many tweets about how stupid Black Friday is (like this one) as tweets about people actually planning to shop.

Who was talking about BF and CM? We sorted the top 10 contributors by number of impressions generated. These are mostly news outlets and celebrities, with only a handful of retailers appearing even in the top 35 (including @amazongames at #11, @WalmartSpecials at #17, and @BestBuy at #31).

And then we sorted the top 10 contributors by the number of retweets they received. Celebrities, humorists and media outlets drove the most retweets (that Justin Bieber sure seems popular). The @amazonmp3 account tweeted about sales frequently in the days leading up to BF and CM, and each tweet received a solid amount of retweets. No other retailers got even close to the number of retweets Amazon did. We’ve seen this for a long time though; people like to get bargains, but they don’t necessarily spend much time reposting those bargains. Also, the number of news accounts represented in the CM chart could lend support to those recent arguments that Cyber Monday is mostly just media-driven hype.

How did they tweet?

Next, on to the tweets themselves. All those contributors posted hundreds of thousands of tweets. Take this graph of tweets per day for both BF and CM.

You can see that in the days before Black Friday, the numbers of tweets per day increased rapidly, culminating in 290,762 tweets on the Thursday before BF. Cyber Monday also saw increases each day leading up the day itself (ending with 68,976 tweets on Sunday before CM). Black Friday saw a great deal more conversation overall, though. In the four days leading up to BF, 641,233 total tweets were posted. In the four days leading up to CM, 120,888 total tweets were posted. That’s a ratio of roughly 5:1.

Most tweets about both BF and CM were standard tweets – around 3/4 of all tweets were regular tweets. Both CM and BF had about 17% retweets and a smaller percentage of replies.

What was the impact of these tweets?

The most impressive numbers are the reach and exposure metrics. Reach is the number of unique Twitter accounts that received tweets about a search term and exposure is the total number of impressions generated. [Note: When we say impressions, we mean the numbers of times a tweet was delivered to a Twitter feed, since there's currently no way to know if someone actually read a given tweet. There's more about how TweetReach calculates those numbers here.]

More than 43 million unique Twitter accounts received at least one tweet about Black Friday (and more than 20 million received tweets about Cyber Monday). These numbers aren’t as far apart as some of the other metrics, which suggests something about saturation. For example, we know from the Tracker that 93% of all contributors to the BF Tracker tweeted three or fewer times; lots of people tweeted a little bit. And only 2% of people tweeted more than 10 times, which means followers of some accounts got bombarded with BF messages (like anyone who still follows the 239 accounts that tweeted more than 50 times in 4 days about BF).

Cyber Monday had 1/5 as many tweets as Black Friday, but those CM tweets still reached 20 million different Twitter accounts, which is almost 1/2 of BF’s reach.

Black Friday generated far more exposure than Cyber Monday (728MM and 197MM, respectively). Again, that suggests that BF tweets were saturating Twitter – most people probably saw multiple BF tweets from multiple accounts. Both had huge exposure, though. They each generated hundreds of millions of impressions and reached tens of millions of Twitter users in just four days.

So, can tweets predict sales results?

Assuming the above-cited numbers are accurate (BF generated some $10-11 billion in sales, while CM was just over $1 billion), AND if we believe Twitter to be related to these trends in some way, then we’d expect to see significantly higher tweet stats for Black Friday compared to Cyber Monday. We certainly see that – some of these BF numbers are many times bigger than the CM ones. Marketers are going to have to work a lot harder if they want Cyber Monday to be as successful as Black Friday.

While it would be statistically irresponsible to call these data a prediction of sales results, the number of contributors seems like a useful metric to watch for future events, as it gives us an indication of participation. More than 7 times more people tweeted about Black Friday than about Cyber Monday, and that’s meaningful.