Holiday Predictions Based on 1 Trillion Website Visits

The continuous release of holiday ecommerce forecasts and findings is just getting started, but not all of them are created equal in terms of objectivity and breadth.

The "Holiday Prediction" report from Adobe Digital Insights (ADI), however, is based on one of the most robust datasets. The forecast is based on an analysis of 55 million product SKUs and aggregated and anonymous data of more than 1 trillion visits to 4,500 retail websites. What's more, $7.50 out of every $10 spent online with the top 500 U.S. retailers go through Adobe Marketing Cloud, and Adobe is able to track 80 percent of all transactions from the top 100 retailers in the U.S. That said, here are highlights of Adobe's predictions

+ With an extended holiday season (Nov-Dec), Adobe predicts 57 total days will exceed $1 billion in sales, 53 of which will be consecutive days (that's a 71 percent year-over-year increase)

+ The holiday season will grow 11 percent year-over-year (YoY), which is 1.7 percent less growth than 2015 but is still $9.1 billion more than 2015 for a total of $91.6 billion overall

+ Large retailers will account for the bulk of online sales growth (16.6 percent) while small retailers will take 7 percent of sales

+ $8.4 billion in sales will be generated from just three days: Thanksgiving ($2 billion for a 15.6 percent YoY growth), Black Friday ($3.05 billion for a 11.3 percent YoY growth) and Cyber Monday (an all-time record of $3.36 billion, a 9.4 percent YoY growth)

+ For the first time, mobile will exceed desktop shopping visits for the full holiday season (53 percent)

+ Despite mobile traffic hitting its tipping point, mobile will only contribute to 34 percent of online sales (a 19 percent gap between visits)

+ Thirty-one percent of U.S. consumers will start shopping before November (a 5 percent YoY increase)

For those interested in more consumer-focused insights like when to get the best deals, what will be the hottest products, etc., check out the infographic below: