We all have known quite intuitively that people, who shop online, usually shop during lunch hours, and at work.
We decided to validate this with actual data, and not surveys with samples so small that they don’t make any sense. We took real social data of over 3M users and ran our algorithms through their shopping behavior. Below is what our algorithms came up with.
The figure below shows the mean rank of a given hour over all the days of data. For every day’s worth of data, we counted the number of requests each hour and rank the hours from the 1st to the 24th by traffic for that day. So if hour 4 has the most traffic on Day 1 and the 2nd most traffic on Day 2, hour 4 has a mean rank of (1+2)/2 = 1.5
So shorter bars are higher traffic hours!
Clearly, lunch hour (from 11am – 2pm) has the most traffic, thus validating the fact that most working people shop online in their lunch hours. As the evening approaches, the traffic starts tapering off. Morning hours are when retailers have the least traffic.
This has several implications.
For retailers trying to get the best deals/offers to their users, this might be a good indicator of when to roll out these offers. It also helps to know at what times in the day, could you expect traffic on your site. It can also give you a sense of when your marketing emails will be opened up and acted upon by your users.
For the techies at the retailers, morning time is probably the best time to roll out new features and test out site changes on the live site, without affecting too many users.
The other inference from this is that people usually use their office Internet to do their online shopping, as the network might be the fastest there. They might not have really good connections at home, as might be inferred from the tapering off in the evenings.
We have all known this intuitively, but this is now backed by real data, from across different networks.
Infinite Analytics has always been at the forefront of cutting edge predictive analytics and semantic technologies. Our Natural Language Processing (NLP), Machine Learning and Semantic Technologies help us establish relationships between users, brands, stars who endorse those brands, movies, affiliation to social causes & prediction of voter inclination.
Watch this space for more analyses on consumers, brands, stars, movies, causes and election analytics.