Holiday e-commerce driven by real-time analytics

(Image credit: Image credit: Bruce Mars / Pexels)

E-commerce retailers need to pamper their customers more than ever this holiday season.  If a page takes too long to load, online shoppers will abandon their shopping carts. Actually, 40% of consumers will leave a page that takes longer than three seconds to load, and 79% of shoppers are less likely to return. And if they can’t see the expected delivery data for each item they order, or pick and choose the post convenient pick up point, they will browse elsewhere. Even before customers make their selections, they are expecting offers and services that are customized to their needs based on past purchases and seasonality. To keep customers engaged, online retailers need to offer a highly personalized shopping experience that requires ingesting, storing and analysing huge amounts of data in microseconds.

Tier based architectures can collapse under the heavy loads of holiday traffic due to multiple clusters and network hops.  By ingesting, analyzing and processing the hot data in memory, in-memory computing systems can eliminate disk and file I/O to provide faster data access while also automatically expanding to provide capacity on-demand.  This helps not only for rapid data retrieval but also to fuel real-time machine learning algorithms used to recommend products, predict inventory levels and product delivery dates, as well dynamically price items.   

Here are some examples of how retailers can ensure superior customer digital experiences.  

Fast clicks and page loads 

A quick page load time is vital. One of the worst nightmares for retailers is a poor user experience that sabotages online orders. It’s important to predict with a high level of accuracy the amount of processing power that’s needed to manage the highest volume of transactions during the busiest time of the year. It’s highly advisable to optimize images and compress your website, and make sure you have a data infrastructure that can scale up quickly as needed to accommodate bursts of traffic without affecting performance and without leading to down-time. 

Window shopping on the small screen

Smartphones have become the main touchpoint with shoppers with this Black Friday reporting a rise to 33.5% of online sales from 29.1% last year.  With the space limitations of smartphones online retailers need to provide fold out product information with photos that can be enlarged and viewed from different angles. Customized product recommendations can appear on the bottom on the screen including items that are especially “hot” this season and more importantly leverage the buyer’s past buying history and sentiment in addition to his current actions.  Mobile shoppers are even more demanding when it comes to quickly check out their items with hassle free automatic forms.

Pricing analysis

In recent years, artificial intelligence has enabled pricing solutions to track buying trends and determine more competitive product prices based on external factors and individual buying habits.  During the holiday season it’s important to track competitors’ price drops and promotions and analyze social and news feeds to better understand market sentiment in real time to stay competitive and assure that clicks convert into sales.


Today customers want all the information they need at their fingertips, so they can make the most informed shopping decision without having to consult with a call center agent.   This includes all available colors, similar styles, where the product was produced and from which materials, whether or not the product is in stock, forecasted delivery date, recommended related products, when the product was packed, shipped, and when it arrived locally.  It’s not enough to forecast the delivery date for the entire order, customers want to know the delivery date for each separate item in the shopping cart, so if one item is delayed the order can be eliminated.  When it comes to delivery information customers want the option to have the gift sent directly to the recipient, sent to the buyer’s home or office or to have the product delivered to the store that is the most convenient for pick up. The more details shared for complete transparency the better.

Unified in store and online experience 

Many customers window shop on the mobile phone but complete the purchase at the store. The website should also include real time reliable information about in-store product availability.  This should include not only the closest store based on the shopper’s location but also other stores within a certain geographic area if the product is unavailable or low in stock.  By using location based marketing, when customers enter the store, retailers can provide a reminder of items viewed online - offer a special price on those items, and provide information on which aisle they can find it and more. 

Personalized shopping experience

This holiday season online shoppers are more demanding than ever and the competition is only a click away.  In-memory computing is architected to support low latency ingestion from various data sources so that relevant data is available in the moment.  Its ability to process data transactions and scale up fast to accommodate huge peaks in holiday traffic while retaining performance is key to operational excellence and real-time machine learning results. Investing in a fast, personalized shopping experience with the necessary underlying data infrastructure is the best way to profit from the busiest shopping time of the year. 

Karen Krivaa, VP of Marketing  at GigaSpaces

Karen Krivaa

Karen Krivaa is the VP of Marketing at GigaSpaces. She has a vast experience with product definition, market analysis, and management of release process.