Providing real-time product and service recommendations has become a “must-do” for online retailers. But the secret ingredient is one that you likely already have — all the customer data you’ve accumulated.
The trick is being able to act on this data. You might imagine the perfect recommendation to incorporate information about the customer’s past purchases, browsing history and favorited items. You would match that up against relevant products and promotions, and take into account practical considerations such as inventory and supply chain.
The calculation would look not only at the customer’s interests, but behavioral patterns of other similar users. And the perfect recommendation would also understand what products go together, by understanding the relationships between products and the category tree.
Advertisement
A common limitation of today’s IT systems is that they often simply aren’t able to crunch all of this data in real time. The result is compromised recommendations, often in the form of precomputed actions that are based on yesterday’s data and emphasize simplicity over richness.
However, the retailer who is able to instantly offer the most rich and relevant information to customers in real time will reap immediate dividends in increased sales and loyalty.
This need not be a missed opportunity, however. Retailers are finding graph databases to be an invaluable framework for creating value from all this super-connected data.
Compelling And Relevant Offers
A graph database is the ideal model for offering more effective recommendations by leveraging relationships between data: not just historical data but inside of the active session. The graph model differs from traditional (relational) business databases by using the patterns inside your data to your advantage, not just the raw data.
Graph databases enable intractable data operations to be done at scale with large datasets in real time. For example, a recommendation engine in a graph database can be a million times faster than a relational database: offering 1,000x performance improvements with a 1,000x increase in data volume.
And retailers really are starting to adopt graph technologies and techniques. The adidas Group recently turned to graph database technology to provide the most compelling and relevant sports content to its consumers as well as offering highly relevant product recommendations.
Whether it’s to direct a fan to a great piece of Real Madrid football merchandise or to serve up a video of a consumer’s favorite basketball team, graph technology combs through that pile of consumer data (purchase history, social media data, etc.) to deliver personalization at an unprecedented level.
Walmart is another major player using graph database technology to take recommendations to a new level. Their solution uses customer purchase information to integrate physical in-store purchases and online browsing into their recommendations.
We’re also starting to see real-time personalization with graphs in industries like travel, health care and media.
Originally, such advanced data architectures were limited to in-house and proprietary solutions, developed by corporate web giants (Google, Facebook, LinkedIn) for their own use. But today, graph database tools are readily available to retailers off the shelf.
Meeting The Challenge Of Personalization
The bottom line is that any business today can leverage its data for maximum insights by using graph technology. It’s little wonder then that the graph database market is growing exponentially.
According to a recent report by industry watcher DB-Engines, graph databases are gaining traction more than any other database category — growing at a staggering 300% since January 2013.
Analysts are also watching the graph explosion with sharp eyes.
Gartner predicts that for data-driven operations and decisions, graph analysis is now “possibly the single most effective competitive differentiator” next to sourcing the data. Furthermore, Forrester Research estimates that one in four enterprises will be using the technology by 2017.
The attraction of graph databases lies in their sheer power, as they can provide a 360-degree view of a customer in real time.
In order to meet the challenge of personalizing customer recommendations, you need a graph database that makes the most of the connections in your existing data that are just waiting to be tapped.
Emil Eifrem is CEO of Neo Technology and co-founder of Neo4j, the world’s leading graph database. Before founding Neo, he was the CTO of Windh AB, where he headed the development of highly complex information architectures for Enterprise Content Management Systems. Committed to sustainable open source, he guides Neo along a balanced path between free availability and commercial reliability.