Akamai Diversity

The Akamai Blog

Black Friday weekend 2016: What we've learned so far about retail traffic and user behavior

At SOASTA, we collect a massive amount of real user data for many top retailers. This past long weekend, not surprisingly, saw an enormous surge in the amount of user experiences our customers monitored and measured. On Black Friday alone, we collected almost 2 billion beacons' worth of real user data.* By the end of Cyber Monday, we estimate that we will have collected between 5-7 billions beacons of data.

Here are a few highlights of what we've learned so far:

1. The Black Friday shopping peak lasted about five hours - from 8am till 1pm (Pacific Time)

During those peak hours, we collected almost 2 million beacons' worth of data per minute. Breaking down all the data by retailer, we saw highly individualized patterns emerge. For some sites, traffic remained high and steady for sustained periods of time, while for others traffic was more spiky, possibly due to how those sites ran time-specific email campaigns and other promotions.

Read > How to prepare your site for the holidays: 12 valuable lessons from Nordstrom.com

2. Many shoppers got the jump on the long weekend and started shopping on Thanksgiving Day

Some retailers had higher beacon counts on Thanksgiving than on the following Saturday or Sunday. For some online shops, Thanksgiving Day was the second-highest traffic day after Black Friday, beating out both Saturday and Sunday.

3. For some retailers, traffic spiked on the Wednesday night beforeThanksgiving

For those sites, we observed peaks in traffic volume at around 10pm ET / 7pm PT. This could be due to a number of factors, ranging from early online promotional campaigns to consumers wishing to get a head start on their shopping before heading into the holiday fray.

Click to check out our Ecommerce Guide to Seasonal Readiness

4. Takeaway: Consumer behavior is unpredictable and variable from site to site

You might be gearing your site - and your war room - for Thanksgiving or Black Friday, but if your shoppers decide to descend on your site en masse before that, you need to be ready for them.

5. Website slowdowns are finally making the news

Website outages have always grabbed headlines, especially during the first shopping weekend of the holidays. No surprise there. Prolonged outages are dramatic. They conjure up images of 3am alerts and frantic war rooms.

But for years, many of us in the performance space have been arguing the case that, while site slowdowns may not be as exciting, they're every bit as damaging -- and perhaps even more so -- as outages. What's interesting to me is that this year marks the first holiday long weekend where slowdowns have actually gotten widespread media attention.

There are so many ways that slowness affects the customer experience and, ultimately, the business - in both the short term and the long term. According to our research into users' holiday shopping behaviour:

  • The load time "sweet spot" for optimal conversions is less than 3 seconds.

  • Just a 1-second slowdown correlates to an 8% increase in bounce rate.

  • What's potentially even worse is the very real risk losing repeat visitors - the bread and butter of any retailer - as 28% of online shoppers say they won't return to a slow site.

* The SOASTA RUM (real user monitoring) beacon is a type of web beacon - an HTTP(S) request that site owners embed in their web pages. Our beacon collects data for each page, so 2 billion beacons equates to 2 billion pageviews (in other words, hundreds of millions of site visits). The beacon collects data about 93 different metrics for each user and visit, including session length, OS/device/browser type, and conversion and bounce rates. The SOASTA beacon is the most sophisticated beacon in the RUM industry. It's part of our mPulse solution, which also includes an industry-leading analytics engine that analyzes the data, correlating page metrics (e.g., load time, start render time) with user experience and business metrics (e.g., bounce rate, session length, pageviews, conversion rate).