You’ve spent a lot of time and money setting up your data warehouse. It’s where large amounts of your company’s data are stored, typically in a relational database ready to be searched by SQL processes.
It brings together data from multiple sources and relieves the burden from the operational databases – enabling you to gain essential insights into your business. These insights can grow sales, increase customer engagement, save money, reduce product warranty costs, and improve departmental collaboration.
If you know what to search for, you can usually find it with some work. Your data warehouse provides data. It’s up to the user to connect the dots and gain insights.
But you’re just scratching the surface of what’s possible.
Your warehouse is not intentionally hiding insights from you. The very nature of its organization obscures the relationships between the data points. Even if you can figure it out, a relational database requires a JOIN operation for every single step of a path.
This task is time-consuming and can be very expensive. Key insights are evasive and non-trivial to retrieve.
Revealing relationship secrets
Enterprise knowledge graphs, however, open up the data in ways that can readily reveal its relationship secrets. Instead of SQL, with graph databases, data is stored and retrieved in terms of relationships and paths, usually using a specialized language (e.g., Gremlin).
This process provides graph functions to analyze paths and patterns in the data. The data tells you its stories by actively enabling analytics, recommendations, and machine learning.
Graph databases provide a more natural way to represent data visually. This includes life cycles, journeys, activities, and anything tied to a path. It enables richer analytics and recommendations.
Graph databases can either augment yourdata warehouse or be an alternative—no more relationship secrets.
How a graph database works
Imagine a large manufacturer with an extensive distributor network who wants to answer these questions:
“How does our training impact sales and customer engagement?”
“How does our sales process affect our warranty expanse?”
Many potential data points reveal the relationship between training, sales, and product warranty claims.
For this example, look at this process in terms of journeys connecting data from CRM, LMS, product IoT, and warranty data.
First, the sales journey from CRM:
Next, the product journey from IoT and warranty claims:
Then the training journey from a Learning Management System (LMS):
Connecting the knowledge dots – this graph emerges:
Now we can see how any failure along these separate journeys impacts specific product sales and warranty claims. An insufficiently trained sales rep can sell the wrong product to a customer resulting in dissatisfaction and a returned product.
Further, a well-trained sales rep should have a higher success rate in the sales process and higher customer satisfaction. You can now optimize sales and operational training to focus on specific areas of improvement in the sales process and customer satisfaction.
Traversing a graph like this in a traditional relational database is expensive. Every question requires a lot of work has to go in to get insights. The added value of a knowledge graph is that you can traverse such paths easily and look for behavior patterns.
Further, in this example,training ROIis clearly demonstrated.
Imagine adding marketing and customer support data to the analysis. Even greater insights would emerge.
Now for the recommendations
Once insights are revealed, you can use a knowledge graph to drive two kinds of recommendations:
- Rule-based recommendations would be a matter of looking up or summarizingrelationships in the knowledge graph.
- Recommendations are obtained using machine learning and observed patterns of behavior.
In the above example,sales reps would be reminded to take the necessary training when the customer is interested in a specific product. Sales performance issues can be identified earlier, and notifications can be sent to sales management for corrective action.
Further, product IoT data can alert sales of potential issues before they reach the critical stage. And sales, technical, and customer support can quickly collaborate on action plans to help customers.
Wrapping up
Quoting @SethGodin, “Insightful data is called information.”By adding a knowledge graph to your data warehouse, you gain additional insights faster, giving your company a competitive edge.
Growing sales, increasing customer engagement, saving money, and increasing collaboration between groups are all good things you can achieve with this approach.
Check out the other videos and articles on our Resources Page
