The IT challenge of facing holiday shopping with a foolproof system built to withstand the blitz of shoppers is not a new one. Each online retailer is well aware that uptime is king and downtime is a killer.
Uptrends recently came out with their website uptime report of 928 online retailers through Cyber Monday 2016, showing who came out on top and who crashed and burned. They report that 509 retailers experienced 100% uptime with 181 shops clocking in at 99.9%. While those numbers are great, it means that roughly 25% of online retailers haven’t figured out yet how to reach at or near 100% uptime. Twenty five percent! And that’s just the challenge of uptime.
Some have stated that 90% of all data is from the last three years alone. Cloudtweaks says that “experts now predict that 40 zettabytes of data will be in existence by 2020.” What’s a zettabyte, you ask? That’s one trillion gigabytes. Yes, you read that right. How in the world are retailers going to handle that kind of volume if they are still battling against uptime?
The ever-elusive “real-time” data
As data grows, uptime is going to be the least of their worries. It should be a given – a no brainer – a truth that they can rely on while they fight newer fights.
Now heads of IT at retail shops lie awake at night thinking about the ever-elusive “real-time” data. How will they keep up with the increasing demand for faster page load times, personalized recommendations, relevant offers based on customer preferences, omni-channel shopping experiences for tech savvy shoppers who search on mobile, purchase on their desktop and pick up in store?
They certainly don’t stand a chance if they are still tackling the uptime issue in 2017. If retailers aren’t looking into data science and machine learning, the promise of Bigger Data will only exacerbate IT challenges down the line.
In a post on TotalRetail, Alexander Gray and Eric Thorsen discuss the role of machine learning in online retailers’ ability to predict consumer trends. They explain that retailers are building solutions to “make sense of the data in real time and provide insights that translate into tangible results.”
Automating the analysis of real-time data, creating systems to predict shopping patterns and make real-time recommendations, providing immediate information on stock and inventory, effectively tracking shipments and making real-time decisions about route changes and optimal distribution centers – these are going to become the capabilities that will differentiate the winners as big data gets bigger.
One method for tackling the real-time issue is to process data in memory at the RAM level, rather than relying on a database as the main system of record. This configuration is especially handy when up against massive upticks in data load and throughput, like on Black Friday and Cyber Monday. The pressure on the database is eased with transactions processing in RAM, pushing down to the database in batches when the line frees up. The method of in-memory computing naturally lends itself to analytics systems such as Apache Spark, itself an in-memory analytics engine for the processing of fast data in real-time. Such systems of insight will one day serve as the backbone of all online retailers who wish to stay in the game against indomitable Amazon.
Tackling the next step of fast data analytics
In my role as head of marketing for the in-memory computing unit at GigaSpaces, I take great pride in watching our retail customers shine each holiday season with increased year-over-year sales and steady upticks in their rate of customer satisfaction.
We made the heads of IT at our retail customers into Cyber Monday heroes for the 7th year running with zero downtime through the holidays. As they look ahead to what the future holds, with increasing customer expectations and growing data sets, their real-time and fast data processing capabilities will be what keeps them ahead of the competition.
This should be the end-game for all online retailers as world data grows. Seventy five percent of them are ready to tackle the next step of faster analytics and more intelligent business decisions for their growing data. The remaining twenty five percent should spend the winter and spring head down on upping their uptime game if they don’t want to get left behind.
Interested in learning about a possible IMC solution for your needs? Get in touch with us.