i-tao

The Power of Faceted Search

Search Engine Technology as a BI Tool

The traditional approach

Traditionally a data warehouse is deployed to manipulate your data to feed your BI requirements. This usually involves building out denormalised tables and organising the data into cubes.

This is an inherently expensive approach, as it requires expertise resource as well as high end hardware and huge software licences.

The biggest downside though is the constraint of database schemas. Sometimes the search and analysis requirements cannot easily fit into highly structured schemas.

Why is search engine technology applicable

Search engine technology has largely grown from solving the problem faced by consumers when they are searching for products.

Believe it or not, when you go online to shop for a product you have a BI problem on your hands. For many commodity products, a good e-commerce site is your tool to narrow down the vast amount of choices available to you.

As well as being able to correct spelling and make alternative suggestions, one of the most important functionalities that modern e-commerce sites deploy is the Faceted Search.

Consider this scenario, you're in the market for a HDTV. You go to your favourite shopping site and hit the TV section. Immediately you are hit with a choice of over a hundred options. But on the side you have the various facets. You may decide to just go for only LCDs, this will reduce the choices a little. You then may decide to narrow it down to only 42". At this point you may be down to 20 or so choices, you narrow down some more by picking only the brands you've had good experience with. Then you narrow it down more by budget. By now you'll be down to just a few choices.

Every time you make a choice the facets dynamically change. For example, you may set a budget first and immediately you will see if 42" is available at that budget and if so how many. You may not have known that LED is available and at cost that was within your budget. This is the power of faceted search, not only does it help you to quickly narrow down but at the same time help you make discoveries too.

Enterprise Search

An example of how we've applied the same search functionalities for one of our customers.

We have collected 4 years worth of invoice and expenses information for 30,000 commercial vehicles that the customer had for rental during that period. Over 9 million records in total and growing.

The facets available to them include customer, vehicle group, depot, age, invoice type, expense type and period. They can quickly narrow down using the facets, for example, one customer, one vehicle group and specific age. They can also do a full text search on any of the fields or do a combination of free search plus faceting.

As they drill down they can see clearly what reduced set of facets are available and how many records behind each. Again helping to discover whilst searching.

Once they have narrowed down to a required subset they can export to Excel for further analysis.

To see your data in this powerful interface, talk to us now about our Proof Of Concept.