The simple sales-vs.-time plot has long been a staple of the corporate boardroom (or Google analytics dashboard). These familiar graphs are intended to show patterns over time, and can be useful for generating forecasts of where sales or revenue are going in the coming periods.
While these projections are useful to members of a finance team, they do little to help the online retailer hone a marketing strategy. Additionally, this approach washes over a large amount of the information generated when a customer makes a purchase, obscuring the millions of records of transaction data that many online retailers collect in the normal course of business
In the age of big data and big data analytics, the way towards better utilizing transaction data lies in narrowing the lens from a period of years and days, down to seconds. If we know when people tend to make purchases or how much they are spending at a given time of day, than we can better coordinate email promotions to reach more potential buyers with higher value.
A well-timed email can encourage a customer to spend more, convince a customer teetering on the edge to click “buy,” or convince someone to buy from you, rather than a competitor.
Every time a customer makes a purchase, a time stamped record is generated indicating the exact moment in time the transaction was processed, as well as a range of other information that can include item-price, location, brand, and product category. As one can imagine, over the course of several years, this data can add up, even for retailers of modest volume.
There are many ways to make use of this wealth of information, but as it pertains to time, the most valuable insights can be found in how shopper behavior varies over time of day.
In order to see how this might look, I have produced a case study using roughly 1.5 million records of data collected over a 7-year period from an online-clothing retailer. Looking at the pattern of sales by time of day, we see identifiable spikes just before noon, and again after dinner.
The average customer is more likely to make purchases after they have settled in at work, answered their emails, and had a productive morning. This level of activity remains constant throughout the workday, before dropping off during dinner, between 6 p.m. and 8 p.m. Activity picks up again and reaches an absolute peak in the hours after dinner and before going to bed. Based on this information, you should send marketing content around mid-morning or sometime in the evening in order to reach the maximum number of likely purchasers.
Price reaches a peak in the morning and trends downward over the course of the day. Perhaps some early-morning shoppers decided to sleep on a big purchase before committing, or maybe there is something about that morning coffee that inspires big-ticket purchases. Whatever the case may be, this information is actionable, and suggests you should send email marketing content early, in the hopes of capturing high impact buyers.
This data can be broken out even further, by conducting the same set of analysis for each day of the week. Looking at volume by time shows some intuitive patterns. Volume is generally lower in the evening on Friday (hard to get serious shopping done while out on the town) and lower in the morning on Sunday (perhaps still recovering from Friday).
Price, on the other hand does not vary so clearly by day of the week, with the general pattern of higher prices in the morning that trail off towards the end of the day relatively constant across days of the week.
Knowing that volume and price vary in a systematic way across time of day can serve to fundamentally improve your marketing strategy. Emails sent before the midday or evening spike in volume may prove more effective than those sent at other times.
Emails sent in the early morning may reach customers intending on purchasing high-priced items and marketing content should be geared towards high-priced items. Since customers across time zones behave similarly, emails should be staggered to capture time-dependent peaks in volume.
However, these intuitions eventually shape your marketing strategy, the ability to collect and analyze large amounts of time stamped transaction data offers exciting new possibilities for online retailers to improve their marketing campaigns.
Ryan Sloan is a market analyst at customer analytics firm Custora.