As a simple example let’s say you have customers with their information is stored in your Customer Relationship Management (CRM) system. All the standard information you want to keep track of fits into it nice and neatly: Name, address, preferences, purchase history, contact information, etc. All is fine, until you want to start knowing and storing more complex information about your customer.
You want to know what their interests are via Facebook or Pinterest. Now your structured CRM has a problem. You want to store very specific information about their hobbies, likes, dislikes, associations, previous preferences on other topics, news interest, restaurants they like, etc. There are no simple ways to add that information into a traditionally relational databases. How can you add a field for hobbies when what you want is highly specific information about their unique hobby? Your customer likes:
Outdoors / Outdoor Sports / Climbing / Ice wall climbing / In Montana / Only competitive event / For seniors
Then add the fact that major sources of critical customer information, like social network information, can sprout up overnight and are unplanned. Who could have built a CRM four years ago to handle a customer’s Pinterest data – Pinterest did not even exist then! And how do you collect, store and manage complex data, like:
Unstructured data: emails, blogs, documents, text messages, photo/video tags and meta data
Social Network data: Key influencers, who is a friend of a friend
Geographic data: Where are purchases being made
Temporal information: Time, date, or date range information, photo and video tags, etc
In this simple example, ideally what you need is a way to add their interests, with as much or little specificity as you can gather, group this information for similarity and hierarchy, track and uncover all your customer likes but also remain flexible as the information grows and the categories change.
You also need to be able to FIND that information about them when you want to, for example, do a special mailing to all customers who like skateboards, Italian food, competitive ice wall climbing and Beats music gear, in or around the Bay Area, and has a network of friends greater than 100. This requires you to FIND information when it is less exact or with conceptual groupings. For example, skateboards can be traditional or more recently they can be motorized. Italian food could be a categorized on its own, but could also be part of Mediterranean food, ethnic food, groupings.
Also with 80% of all data now unstructured (blogs, documents, tweets, comments not just rows and columns) you need a semantic graph database that can help put that unstructured data into structured context to get actionable insights.
The AllegroGraph semantic graph database has been optimized specifically to deal with these complex, ever evolving types of data and make accessing them actionable and insightful. It also designed to find the hidden connections of information that would otherwise go unnoticed.